US20100246588A1 - System architecture for very fast ethernet blade - Google Patents

System architecture for very fast ethernet blade Download PDF

Info

Publication number
US20100246588A1
US20100246588A1 US12/702,031 US70203110A US2010246588A1 US 20100246588 A1 US20100246588 A1 US 20100246588A1 US 70203110 A US70203110 A US 70203110A US 2010246588 A1 US2010246588 A1 US 2010246588A1
Authority
US
United States
Prior art keywords
packet
packets
data
receive
backplane
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
US12/702,031
Inventor
Ian Edward Davis
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Foundry Networks LLC
Original Assignee
Foundry Networks LLC
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Foundry Networks LLC filed Critical Foundry Networks LLC
Priority to US12/702,031 priority Critical patent/US20100246588A1/en
Assigned to FOUNDRY NETWORKS, LLC reassignment FOUNDRY NETWORKS, LLC CHANGE OF NAME (SEE DOCUMENT FOR DETAILS). Assignors: FOUNDRY NETWORKS, INC.
Publication of US20100246588A1 publication Critical patent/US20100246588A1/en
Abandoned legal-status Critical Current

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/35Switches specially adapted for specific applications
    • H04L49/351Switches specially adapted for specific applications for local area network [LAN], e.g. Ethernet switches
    • H04L49/352Gigabit ethernet switching [GBPS]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/20Support for services
    • H04L49/201Multicast operation; Broadcast operation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/30Peripheral units, e.g. input or output ports
    • H04L49/3009Header conversion, routing tables or routing tags
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/30Peripheral units, e.g. input or output ports
    • H04L49/3063Pipelined operation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/35Switches specially adapted for specific applications
    • H04L49/354Switches specially adapted for specific applications for supporting virtual local area networks [VLAN]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/50Overload detection or protection within a single switching element

Definitions

  • the invention described herein relates to computer networking and, in particular, to improved methods, systems, and software for routing data at high speeds through a switch or other network routing device.
  • Ethernet one of the earliest networking protocols, is today the most widely used method to transport data in computer networks. Robert Metcalf and David Boggs developed Ethernet as an experiment at the XEROX Palo Alto Research Center in 1973. At Ethernet's inception, the struggle to accommodate users needs for bandwidth had not yet started. As network traffic demands at this time were quite low, Ethernet initially had a data transmission rate of 2.94 megabits per second (Mbps).
  • Metcalf recognized the potential for rapid network growth and posited a theorem now known as “Metcalf's Law” which states that the value of a network expands exponentially as the number of users increase.
  • Gordon Moore an expert in the field of semi-conductor development, posited another theorem known as Moore's Law which states that the power of microprocessors will double every 18 months and their price will be reduced by half.
  • Moore's Law which states that the power of microprocessors will double every 18 months and their price will be reduced by half.
  • Ethernet The evolution of Ethernet technologies has followed theory. The first commercial release of Ethernet occurred in 1979 with a transmission rate of 10 Mbps—more than a three-fold increase over the experimental system created just five years earlier. Ethernet went through a variety of standardizations during the 1980s and line rates remained constant at 10 Mbps while the technology matured. In 1995, however, Ethernet became available at 100 Mbps. In 1998, bandwidth jumped again to 1 gigabit per second (Gbps). Most recently, a new standard was adopted for Ethernet transmission rates at 10 Gbps representing a 100-fold increase in seven years.
  • the present invention provides a switch or router for providing data transmission speeds up to 10 gigabits per second between one or more source devices and one or more destination devices.
  • the switch includes a blade or board having several discrete integrated circuits embedded thereon, each performing one or more discrete functions required to meet the speed required for the switch.
  • the blade includes a media access control interface (MAC) to facilitate receipt and transmission of packets over a physical interface.
  • the blade further includes four field programmable gate arrays.
  • a first field programmable gate array is coupled to the MAC array and operative to receive packets from the MAC interface and configured to perform initial processing of packets.
  • the first field programmable gate array is further operative to dispatch packets to a first memory, such as a dualport memory.
  • a second field programmable gate array is operative to retrieve packets from the first memory and configured to compute an appropriate destination and to dispatch packets to a backplane.
  • a third field programmable gate array is operative to receive packets from the backplane and configured to organize the packets for transmission and to dispatch packets to a second memory.
  • a fourth field programmable gate array is coupled to the MAC interface and operative to retrieve packets from the second memory and to schedule the transmission of packets to the MAC interface for transmission to one or more destination devices.
  • the invention comprises a switch or router for providing data transmission speeds up to 10 gigabits per second between one or more source devices and one or more destination devices through the use of two sets of one or more field programmable gate arrays.
  • a first set of one or more field programmable gate arrays is coupled to a media access control (MAC) interface and a memory structure, the MAC interface used to facilitate the receipt and transmission of packets over a physical interface.
  • the first field programmable gate array set is operative to receive and transmit packets from and to the MAC interface.
  • the first field programmable gate array set is configured to perform initial processing of received packets and to schedule the transmission of packets to the MAC interface for transmission to one or more destination devices, in addition to dispatching and retrieving packets to and from the memory structure.
  • This embodiment of the invention also comprises a second set of one or more field programmable gate arrays coupled to the memory structure and a backplane.
  • the second field programmable gate array set is operative to retrieve packets from and dispatch packets to the memory structure, and configured to compute an appropriate destination and organize packets for transmission.
  • the second field programmable gate array set is further operative to receive and dispatch packets from and to the backplane.
  • FIG. 1 is a block diagram of a system architecture for an Ethernet blade in accordance with one embodiment of the present invention
  • FIG. 1A is a block diagram of a system architecture for an Ethernet blade in accordance with a second embodiment of the present invention
  • FIG. 2 is a high level flow diagram of a connection of a packet processor component of the present invention to an outside network, in accordance with one embodiment of the present invention
  • FIG. 3 is a block diagram of receive and transmit packet processors of one embodiment of the present invention.
  • FIG. 4 is a block diagram of a receive packet processor in accordance with one embodiment of the present invention.
  • FIG. 5 is a flow diagram showing the data flow in the receive packet processor of FIG. 4 in accordance with one embodiment of the present invention
  • FIG. 6 is a block diagram of a backplane manager in accordance with one embodiment of the present invention.
  • FIG. 7 is a flow diagram showing the data flow in a transmission accumulator in accordance with one embodiment of the present invention.
  • FIG. 8 is a block diagram of a transmit packet processor component in accordance with one embodiment of the present invention.
  • FIG. 1 a block diagram is presented depicting a high-level schematic of the components of one possible embodiment of the invention to allow data transfer speeds at or in excess of 10 gigabits per second.
  • the invention comprises a printed circuit board (“PCB”) 10 used to house and provide interconnections for a media access controller (“MAC”) 12 , a packet processor (“PP”) 14 , one or more content addressable memory (“CAM”) controllers 16 , one or more controllers for random access memories containing parameter information (“PRAM”) processors 18 , a receive dual-port memory buffer 20 , a transmit dual-port memory buffer 22 , a transmission manager 24 , and a backplane interface 26 .
  • PCB printed circuit board
  • the PCB 10 provides a surface on which to place other components of the invention.
  • the PCB 10 also known as a “blade” or “module”, can be inserted into a slot on the chassis of a network traffic management device such as a switch or a router.
  • a network traffic management device such as a switch or a router.
  • BigIron® switch produced by Foundry Networks, Inc. of San Jose, Calif.
  • the BigIron switch chassis consists of multiple distributed switching modules each of which contain a high-bandwidth memory system for scalable chassis bandwidth.
  • the local switching fabric of the BigIron switch houses the forwarding engines, provides packet-level examination and classification based on Layer 2/3/4 information, and performs IP subnet look-ups and packet modifications of IP and IPX packets.
  • the MAC 12 is the interface by which data is received and transmitted to and from the network.
  • network data comprises Ethernet packets.
  • the MAC 12 forwards received packets to the PP 14 for further processing and also receives packets for transmission to the network from the PP 14 .
  • the MAC 12 performs any data conversions required for network data to be processed by the PP 14 for routing within the device chassis and for data processed by PP 14 to be transmitted to the network.
  • the MAC 12 performs data conversions because network data comprises 32 bit double data rate (“DDR”) data, while the PP 14 processes only 64 bit single data rate (“SRD”) data.
  • DDR 32 bit double data rate
  • SRD 64 bit single data rate
  • the MAC is typically responsible for data validity checking, as well as data gathering.
  • the PP 14 is a processor chip responsible for receiving packets from the MAC 12 and processing them for forwarding through the device chassis, as well as for processing packets received from the device chassis intended for transmission over the network. These two functions, while performed on the same chip, are preferably performed simultaneously and in parallel. There are thus, in a sense, two pipelines in the PP 14 : a receive pipeline for processing network packets intended for transmission within the chassis and a transmit pipeline for processing internally routed packets intended for transmission over the network.
  • the packet processor is a field programmable gate array (“FPGA”), which is an integrated circuit that can be programmed in the field after manufacture.
  • FPGA field programmable gate array
  • An advantage of using FPGAs with the invention is that an FPGA provides significant flexibility over an application specific integrated circuit (“ASIC”) and is also much less expensive to prototype and implement.
  • the receive pipeline of the PP 14 is responsible for packet classification, performing CAM and PRAM lookups, generating packet headers for forwarding packets through a chassis, and preparing packet modifications.
  • Network packets are received by the PP 14 from the MAC 12 in multi-byte bursts based on scheduling priorities determined at the MAC 12 .
  • the PP 14 examines packets and extracts packet forwarding information from the packets such as the destination address (“DA”) of the packet and the source address (“SA”) of the packet.
  • the PP 14 extracts the type of service (“TOS”), whether the packet has a virtual local area network (“VLAN”) tag, session related data such as in the case of IPv4 or IPX data, and other additional Layer 3 and Layer 4 information useful in routing the packet through the chassis.
  • the PP 14 passes this forwarding information extracted from the packet header to a CAM processor 16 for further processing.
  • the CAM controller or processor 16 takes information forwarded by the PP 14 and performs a lookup comparing this information to data stored in a local memory of the CAM processor 16 . If the information matches information stored in the local memory of the CAM processor 16 , additional forwarding information regarding disposition of the packet is available in the local memory of the PRAM processor 18 and can be retrieved for future incorporation into the packet header.
  • the PRAM processor 18 retrieves additional forwarding information from its local memory for incorporation into the header of the packet.
  • the packet is reformatted with a new internal hardware header for routing the packet within the chassis and stored in the receive dual-port memory buffer 20 for processing by the transmission manager.
  • This internal hardware header is also sometimes referred to as a chassis header.
  • Pipelining is an advanced technique used by processors, wherein a processor begins executing a subsequent instruction before a prior instruction has finished executing. Accordingly, a processor can have multiple instructions processing in its “pipeline” simultaneously with each instruction at a different processing stage.
  • the pipeline is divided into processing segments, with each segment executing its operation concurrently with the other segments.
  • a segment completes its operation, it passes the result to the next segment in the pipeline and fetches data for processing from the preceding segment.
  • temporary memory buffers are used to hold data values between segments, which allows operations to complete faster since each segment no longer waits for the other segment to finish processing prior to handing off data. The final results of the process emerge at the end of the pipeline in rapid succession.
  • the receive dual-port memory 20 acts as a pipeline buffer in the embodiment of the invention depicted in FIG. 1 .
  • the receive dual-port memory 20 enables the PP 14 to store processed data and continue processing the next packet without having to wait for the transmission manager 24 to become available, thereby expediting operations of both the PP 14 and the transmission manager 24 .
  • Other buffers are used throughout the invention and in its various components to achieve pipelining and faster packet processing in an analogous manner.
  • the transmit pipeline of the PP 14 retrieves data from the transmit dual-port memory 22 according to a programmable priority scheme.
  • the PP 14 extracts network destinations from the dual-port data and reassembles packet header forwarding information by removing any packet header modifications that take place in order to route the packet through the switch chassis.
  • the PP 14 performs sanity checks on packet data to ensure that only those packets intended for transmission are passed on to the MAC 12 .
  • packets routed through the chassis carry header information pertaining to forwarding within the chassis, this information must be removed and replaced with header forwarding information appropriate for routing over the network. After the proper network header forwarding information is reassembled and the chassis header information is removed, the PP 14 forwards the data to the MAC 12 for eventual transmission over the network to the intended address.
  • the transmission manager 24 handles traffic flow to and from the backplane interface 114 .
  • the transmission manager 24 is a processor chip that implements a dual pipeline architecture: a receive pipeline for network data to be internally routed within the device chassis and a transmit pipeline for internally routed data intended for network transmission. These two functions, while performed on the same chip, are preferably performed in parallel according to one embodiment of the invention.
  • the transmission manager 24 is an FPGA, although use of other processor types is within the scope of the invention.
  • the transmission manager 24 fetches network data intended for routing through the device chassis from the receive dual-port memory 20 and stores internally routed data intended for network transmission in the transmit dual-port memory 22 .
  • the receive pipeline of the transmission manager 24 retrieves data from the receive dual-port memory 20 according to instructions issued to the transmission manager 24 by the PP 14 .
  • the transmission manager 24 determines data transmission priority for the data retrieved and schedules transmissions to the backplane 26 according to this priority scheme. In one embodiment of the invention, there are four different priority levels assigned to data.
  • the transmission manager 24 extracts backplane destinations from data, and sends data to those destinations according to predetermined priority algorithms.
  • Backplane destinations may comprise other blades in the chassis or, in some cases, may comprise the blade of the transmission manager 24 itself, which is called “one-armed routing.”
  • the transmit pipeline of the transmission manager 24 handles internally routed packets received from the backplane interface 26 and intended for transmission over the network.
  • the transmission manager 24 collects packets from the backplane interface 26 and organizes them into per-source, per-priority transmit queues stored in the transmit dual-port memory 22 .
  • the transmission manager 24 notifies the PP 14 when a packet is stored in the transmit dual-port memory 22 and available for processing.
  • FIG. 1 a presents a block diagram depicting a high-level schematic of the components of an alternative embodiment of the invention.
  • the invention comprises a printed circuit board 100 , a media access controller 102 , a receive packet processor 104 (“RXPP”), one or more CAM processors 106 , one or more PRAM memory processors 108 , a receive dual-port memory buffer 110 , a backplane manager 112 , a backplane interface 114 , a transmission accumulator (“TX accumulator”) 116 , a transmit dual-port memory buffer 118 , and a transmit packet processor (“TXPP”) 120 .
  • RXPP receive packet processor
  • TX accumulator transmission accumulator
  • TX accumulator transmit dual-port memory buffer
  • TXPP transmit packet processor
  • the PCB 100 provides a surface on which to place many of the other components of the invention.
  • the PCB 100 also known as a “blade” or “module”, can be inserted into one of a plurality of slots on the chassis of a network traffic management device such as a switch or a router.
  • a network traffic management device such as a switch or a router.
  • the MAC 102 is the interface by which a blade receives and transmits data to and from the network.
  • network data comprises Ethernet packets.
  • the MAC 102 forwards received packets to the RXPP 104 for further processing and receives packets for transmission to the network from the TXPP 120 .
  • the MAC 102 also performs any data conversions required for network data to be processed by the RXPP 104 or for data processed by TXPP 120 to be transmitted to the network. For example, the MAC 102 may perform data timing conversions where network data comprises 32 bit DDR data while the RXPP 104 and the TXPP 120 process only 64 bit SDR data.
  • the receive packet processor 104 is responsible for packet classification, performing CAM and PRAM lookups, generating packet headers for forwarding packets through a chassis, and preparing packet modifications.
  • the receive packet processor 104 is an FPGA.
  • the RXPP 104 is an ASIC. Packets are received by the RXPP 104 from the MAC 102 in multi-byte bursts based on scheduling priorities determined at the MAC 102 .
  • the RXPP 104 examines packets and extracts packet forwarding information from a packet, such as the destination address of the packet and the source address of the packet.
  • the RXPP 104 extracts the TOS, any defined VLAN tags, session related data such as in the case of Ipv4 or IPX data, and other additional Layer 3 and Layer 4 information useful in routing the packet through the chassis.
  • the RXPP 104 passes this forwarding information to one of the CAM processors 106 for further examination.
  • the CAM processor 106 takes information forwarded by the RXPP 104 and performs a lookup, comparing received information to data stored in local memory of the CAM processor 106 . If the comparison returns a match, additional forwarding information regarding disposition of the packet is stored in local memory of one of the PRAM processors 108 and can be retrieved for future incorporation into the packet header. The PRAM processor 108 retrieves additional forwarding information from its local memory for incorporation into the header of packet. The packet is then stored in the receive dual-port memory buffer 110 for processing by the backplane manager 112 . Those of skill in the art will recognize that additional processing may be performed before storage in the receive dual port memory.
  • the receive dual-port memory 110 acts as a pipeline buffer between processes.
  • the receive dual-port memory 110 enables the RXPP 104 to store processed data and continue processing the next packet without having to wait for the backplane manager 112 to become available.
  • Pipelining operation execution expedites processing of both the RXPP 104 and the backplane manager 112 .
  • Other buffers are used throughout the invention and within its various components to achieve pipelining and faster packet processing in this manner.
  • the next segment in the receive pipeline is the backplane manager 112 .
  • the backplane manager 112 is a processor designed for retrieving data from the receive dual-port memory buffer 110 and dispatching packets to the backplane interface 114 . Data is retrieved from the receive dual-port memory 110 according to instructions issued to the backplane manager 112 by the RXPP 104 .
  • the backplane manager 112 determines data transmission priority for the data retrieved and schedules transmissions to the backplane 114 according to this priority scheme. According to one embodiment of the invention, there are four different priority levels assigned to data.
  • the backplane manager 112 extracts backplane destinations from received data; the data sent to indicated destinations according to programmable priority algorithms.
  • Backplane destinations may comprise other blades in the chassis or, in the case of OAR, may comprise the blade of the backplane manager 112 that initially receives the data.
  • packets scheduled for OAR are detected, they are forwarded to the transmission accumulator 116 via the OAR data path as shown in FIG. 1 a .
  • the backplane manager 112 is an FPGA.
  • the backplane manager 112 is an ASIC.
  • the transmit accumulator 116 is a processor that receives packet data from the backplane 114 intended for transmission.
  • the transmit accumulator 116 collects packets from the backplane 114 and organizes them into per-backplane-source, per-priority transmit queues stored in the transmit dual-port memory 118 .
  • the transmit accumulator 116 notifies the TXPP 120 when data comprising a packet is stored in the transmit dual-port memory 118 and available for processing.
  • the transmit accumulator 116 is an FPGA.
  • the transmit packet processor 120 retrieves data from the transmit dual-port memory 118 according to a programmable priority scheme.
  • the TXPP 120 extracts network destinations from the data and reassembles packet header forwarding information by removing any packet header modifications that took place in order to route the packet through the device chassis.
  • the TXPP 120 performs sanity checks on packet data to ensure that only those packets intended for transmission are passed on to the MAC 102 . Since packets routed through the chassis carry header information pertaining to forwarding within the chassis, this information must be removed and replaced with header forwarding information appropriate for routing over the network. After the proper network header forwarding information is reassembled and the chassis header information is removed, the transmit packet processor 120 forwards the data to the MAC 102 for eventual transmission over the network to the intended address.
  • the transmit packet processor 120 is an FPGA. In an alternate embodiment of the invention, the transmit packet processor 120 is an ASIC.
  • FIG. 2 presents a high-level schematic of one embodiment of the invention as it connects to a network, e.g., an optical network comprising fiber optic connections.
  • the optics block 202 is the interface through which all network data traffic passes.
  • the optics block 202 contains a transmitter for generating the optical signals to the network when data is received from the transceiver 204 .
  • the transmitter might comprise a laser or a light emitting diode.
  • the optics block 202 also contains a detector for receiving optical data traffic from the network. When optical data is received, a photodetector generates an electrical current that is amplified to level useable by the transceiver 204 . The signal is then communicated to the transceiver 204 for further processing.
  • data passes to and from the transceiver 204 and the MAC 206 at a fixed speed.
  • the datapath 208 between the transceiver and the MAC 206 operates sending 4 clock signals along with 32 bit DDR data at 156.25 MHz.
  • the datapath 212 between the MAC 206 and the packet processor 210 may operate at a different speed.
  • the datapath 212 between the MAC 206 and the packet processor 210 operates sending 4 clock signals along with 64-bit SDR at 66 MHz. Multiple clock signals are sent with the data and used to minimize timing differences between groups of data signals and a clock.
  • one clock signal is included per 8 bits of DDR data and one clock signal is included per 16 bits of SDR data.
  • control signals are also sent along with data to indicate packet boundaries and possible error conditions. In one embodiment of the invention, control signals are distributed across 4 clock groups of data.
  • the packet processor 210 can also capture the data on that clock via a FIFO.
  • Data from the MAC 206 is captured inside a FIFO, which allows the packet processor to synchronize, in the presence of this data, between the source synchronous clock contained in the FIFO data and the clock the packet processor 210 is using at its core.
  • the invention uses source synchronous clocking in a symmetric manner. For example, data passing from the packet processor 210 to the MAC 206 is also captured in a FIFO to allow the MAC 206 to synchronize, in the presence of the FIFO data, between the source synchronous clock (of the packet processor 210 core) and the clock the MAC 206 is using at its core clock.
  • the invention also implements differential source synchronous clocking which is known to those skilled in the art. Differential source synchronous clocking works in much the same manner as source synchronous clocking, except that two clock signals are sent with the data instead of one clock signal. The two clock signals, a high and low signal, are used to calculate a more precise approximation of the signal value being transmitted which those skilled in the art will recognize is used to reduce noise and generate more accurate data transmissions.
  • FIG. 3 is a block diagram depicting one embodiment of the components of the MAC 102 as presented in FIGS. 1 and 1 a .
  • Components of the MAC 102 are embodied in the MAC processor chip 302 .
  • the MAC chip 302 is an FPGA.
  • the MAC chip 302 is an ASIC.
  • the MAC 102 is the interface between the network via the PHY transceiver 300 and the RXPP 104 and TXPP 120 packet processor chips.
  • the MAC 102 communicates directly to the PHY layer transceiver 300 via a DDR interface and with the packet processor chips of the RXPP 104 and the TXPP 120 via an SDR interface.
  • the receive block 304 performs a variety of tasks on data received from the receive front end 306 and is very flexible in operation.
  • the receive block 304 internally converts data received from the receive front end 306 into a format suitable for transmission to the RXPP 104 .
  • the receive block converts 32 bit DDR data into 64 bit SDR data for transmission.
  • the receive block 304 may also perform other tasks as required according to various embodiments of the invention such as verifying and extracting XGMII tokens, realigning bytes such that the start of packet (“SOP”) token is placed in a “lane zero” position, verifying SOP and EOP framing, detecting giant packets, verifying and optionally stripping packet cyclic redundancy checks, tracking the full suite of RMON statistics, and other useful operations.
  • SOP start of packet
  • EOP framing detecting giant packets
  • verifying and optionally stripping packet cyclic redundancy checks tracking the full suite of RMON statistics, and other useful operations.
  • the receive block 304 also generates flow control packets via the pause and flow control sync block 332 .
  • the receive block 304 operates off of the recovered source synchronous clocks contained in the incoming data packets received from the PHY transceiver 300 .
  • Other components of the MAC 102 including the transmit block 328 , however, are operating off of an internal core clock generated locally. Although these two clocks are nominally the same frequency, there is some variance since they are not really the same clock and therefore tend to “drift” over time. This difference between the two clocks requires periodic synchronization of the receive block 304 and the transmit block 328 for the purposes of passing flow control messages to generate pause frames and avoid network congestion.
  • the receive block 304 receives an incoming message from a remote source (to which the transmit block 328 is sending data) indicating that the remote source is becoming congested and requesting that the transmit block 328 pause transmission for a requested interval.
  • the pause and flow control sync block 332 synchronizes the receive block 304 clock with the transmit block 328 clock to permit the receive block 304 to pass the pause frame request to the transmit block 328 and reduce the network congestion.
  • the pause and flow control sync block 332 would synchronize the two clocks to permit the receive block 304 to instruct the transmit block 328 to start issuing flow control pause frames to a remote sender to reduce network congestion in the MAC 102 .
  • the receive block 304 passes processed data to the receive FIFO RAM 308 via the write port 310 of the receive FIFO RAM 308 which enables the receive block 304 to process the next packet without waiting for the receive FIFO block 314 to become available.
  • the receive FIFO RAM 308 is a two-port memory having a write port 310 that accepts incoming data from the receive block 304 and a read port 312 that transmits data stored in the receive FIFO RAM 308 to the receive FIFO block 314 .
  • the write port 310 and the read port 312 operate independently of each other thus permitting more efficient use of the receive FIFO RAM 308 by the receive block 304 and the receive FIFO block 314 .
  • the FIFO RAM 308 further permits data flow though the asynchronous boundary.
  • the receive block 304 operates at a different speed than the receive FIFO block 314 .
  • the FIFO RAM 308 acts as a bridge, allowing data flow to be synchronized between these asynchronous components.
  • the receive block 304 operates at a 156.25 MHz clock recovered from the arriving data and the FIFO block 314 operates on a locally generated 156.25 MHz clock that differs slightly and drifts in phase relationship over time.
  • the receive block 304 starts streaming data into the receive FIFO RAM 308 when the receive block detects the start of a packet and stops streaming data into the receive FIFO RAM 308 when the receive block 304 detects the end of the packet. All of the packet processing components of the invention stream data into FIFOs in this manner which greatly reduces processing time since components are not required to wait until an entire packet is finished processing to start copying the packet into a FIFO.
  • the receive FIFO block 314 reads data stored in the receive FIFO RAM 308 via the read port 312 .
  • the receive FIFO block 314 also notifies the RXPP 104 that packet data is contained in the receive FIFO RAM 308 and available for transmission. This data is transmitted to the RXPP 104 for further processing.
  • the receive block FIFO 314 transmits 64 bit SDR data to the RXPP 104 .
  • the MAC 102 In addition to the receive pipeline of the MAC 102 as set forth above, the MAC 102 also contains a transmit pipeline that operates in a similar fashion with similar flexibility.
  • the transmit FIFO block 320 is the interface of the MAC 102 that receives data from the TXPP 120 . According to one embodiment of the invention, the transmit FIFO block 320 receives 64 bit SDR data from the TXPP 120 .
  • the transmit FIFO block 320 streams received data to the transmit FIFO RAM 322 via the write port 324 of the transmit FIFO RAM 322 , enabling the transmit FIFO block 320 to process the next incoming packet without waiting for the transmit block 328 to become available.
  • the transmit FIFO RAM 322 is a two-port memory having a write port 324 that accepts incoming data from the transmit FIFO block 320 and a read port 326 that transmits data stored in the transmit FIFO RAM 322 to the transmit block 328 .
  • the write port 324 and the read port 326 of the transmit FIFO RAM 322 operate independently of each other, thus permitting pipelining and more efficient use of the transmit FIFO RAM 322 by the transmit FIFO block 320 and the transmit block 328 .
  • the transmit block 328 reads data stored in the transmit FIFO RAM 322 via the read port 326 . Similar to the receive block 304 , the transmit block 328 performs a variety of tasks and is very flexible in operation.
  • the transmit block 328 internally converts data received from TXPP 120 into a format suitable for transmission to the PHY transceiver 300 . According to one embodiment of the invention, the transmit block converts 64 bit SDR data into 32 bit DDR data for transmission.
  • the transmit FIFO RAM 322 facilitates this conversion by bridging the asynchronous boundary between the transmit block 328 and the transmit FIFO block 320 .
  • the transmit block performs other tasks as required according to embodiments of the invention, such as generating flow control packets to the PHY side sender at the request of the TXPP 120 (and in addition to internal flow control requests generated by the receive block 304 via the pause and flow control sync 332 when the receive FIFO RAM 308 is full) to avoid network congestion, calculating and optionally appending a cyclic redundancy check to a packet, determining and inserting XGMII tokens, and tracking the full suite of RMON statistics.
  • the transmit block 328 stores data in a programmable FIFO buffer used for data rate matching which allows the MAC 102 to connect to a packet processor that is receiving data slower than line rate.
  • the transmit block 328 passes data processed for to the transmit front end 330 thus enabling the transmit block 328 to begin processing the next packet.
  • the transmit front end 330 is an interface that receives data from the transmit block 328 and passes this data to the PHY transceiver 300 for transmission over the network. According to one preferred embodiment of the invention, the transmit front end 330 transmits 32 bit DDR data to the PHY transceiver 300 .
  • FIG. 4 presents a block diagram depicting one embodiment of the components of the RXPP 104 .
  • the RXPP 402 is responsible for packet classification, performing CAM and PRAM lookups, generating hardware packet headers used for internally forwarding packets within the chassis of the network device, such as between blades, and for preparing packet modifications.
  • Components of the RXPP 104 are embodied in the RXPP chip 402 .
  • the RXPP chip 402 comprises an FPGA.
  • the RXPP chip 402 comprises an ASIC.
  • the XGMAC 404 interface is responsible for requesting data for the RXPP 402 from the MAC 102 .
  • the receive lookup handler 406 is available to parse additional data and the receive data FIFO 438 is available to store additional data
  • the XGMAC 404 instructs the MAC 102 to begin streaming packet data into the RXPP 104 .
  • the XGMAC interface 404 is connected to the MAC 102 and receives data for processing by the RXPP 104 .
  • the XGMAC interface 404 also acts as an asynchronous boundary, writing source-synchronous 64-bit data from the MAC 102 in a small internal FIFO, then sending the synchronized data at 66 MHz in 256-bit chunks for subsequent processing.
  • the XGMAC interface 404 sends synchronized data as it is received from the MAC 102 to the receive data FIFO 438 , where it is held until CAM and PRAM lookups are performed.
  • the receive data FIFO 438 thus acts as a delay buffer until packet processing is completed and the packet data can start being written by the dual-port interface 440 into the receive dual-port memory 110 .
  • the XGMAC interface 404 While all data related to a packet is streamed to the receive data FIFO 438 , the XGMAC interface 404 also parses the incoming data as it is received from the MAC 102 and streams only the packet header information to the receive lookup handler 406 where it will be used to perform CAM and PRAM lookups.
  • the receive lookup handler 406 performs sanity checks on the packet data as it is received from the XGMAC interface 404 . For example, the receive lookup handler 406 identifies valid packet contexts by identifying consistent start-of-packet and end-of-packet boundaries. In this respect, the receive lookup handler 406 also monitors a bad packet control signal from the MAC 102 indicating a data fault. If a data fault is detected, the receive lookup handler 406 discards the header data from the bad packet and also flushes any associated data already stored in the receive data FIFO 438 related to the bad packet.
  • a data fault flag indicating a bad packet is stored in the receive data FIFO 438 .
  • the dual port interface 440 will later discard the packet when the data fault flag is retrieved from the receive data FIFO 438 .
  • the receive lookup handler 406 strips VLAN tags, compares the packet MAC destination address against the port MAC address, performs IPv4 TOS field lookups as required, and also checks the protocol used to encode the packet. Examples of encoding protocols include IP, IP ARP, IPv4, IPv6, 802.3, IPX RAW, IPX LLC, IPX 8137, IPX SNAP, Appletalk, Appletalk ARP, NetBios, IP SNAP, and IP ARP SNAP. This information will be used to assemble an internal hardware packet header to be appended to the packet for use in forwarding the data internally throughout the chassis of the network switch. This additional information is passed from the receive lookup handler 406 to the RX scheduler FIFO 407 . The RX scheduler FIFO 407 holds this information until the CAM and PRAM lookups are completed on the destination and source addresses extracted by the receive lookup handler 406 from the packet header.
  • the receive lookup handler 406 Based upon the information extracted, the receive lookup handler 406 forms the CAM lookups and builds part of the hardware packet header for internally forwarding the packet through the chassis of the network device. The internal state of the receive lookup handler 406 containing this information is then split into two CAM lookup FIFOs 408 and 410 , which are memory buffers that permit the receive lookup handler 406 to start processing the next packet received from the XGMAC interface 404 . Packet processing is thus pipelined, allowing the receive lookup processor 406 to continue processing packets without waiting for either the CAM1 interface 412 or the CAM2 interface 410 to become available.
  • Information relating to the destination address of the packet and other protocol fields from the header related to Layer 3 are passed to CAM1 lookup FIFO 408 .
  • Information relating to the source address of the packet and other protocol fields from the header related to Layer 4 are passed to CAM2 lookup FIFO 410 .
  • the two pipelines are merged into a single pipeline containing a single CAM interface and a single FIFO interface for lookups.
  • the CAM1 interface 412 becomes available, retrieves the data stored in the CAM1 lookup FIFO 408 , and submits requests regarding this data to the external ternary CAM1 414 memory bank that contains a data array of values against which to perform lookups.
  • the CAM1 interface 412 is also pipelined and supports dispatching lookups for multiple packets to the external ternary CAM1 414 memory bank since it takes longer than four clocks for the external CAM1 414 to respond.
  • additional forwarding information exists in the PRAM1 426 memory bank regarding the disposition of the packet. Forwarding information might include details such as the destination port of the packet, the port mirror requirement, the packet type, VLAN handling information, packet prioritization data, multicast group membership, replacement destination MAC addresses (used in network routing), and/or other similar packet data known in the art.
  • the CAM1 414 array entry also contains a link to the memory address of the additional forwarding information stored in the PRAM1 426 memory bank. This link is stored in the CAM1 result FIFO 420 until the PRAM1 interface 424 is available to perform lookups.
  • the CAM2 interface 416 retrieves source address data from the CAM2 lookup FIFO 410 , performs lookups by submitting requests to the external ternary CAM2 memory bank 418 , and stores the results of these lookups in the CAM2 result FIFO 422 until the PRAM2 interface 428 is available to perform lookups.
  • the CAM2 interface 416 operates in parallel with the CAM1 interface 412 to allow CAM lookup operations to complete faster.
  • the PRAM1 interface 424 retrieves the data associated with the successful CAM1 interface 412 lookups from the CAM1 result FIFO 420 .
  • the PRAM1 interface 424 extracts from this data the link to the memory address of the additional forwarding information stored in the PRAM1 426 memory bank.
  • PRAM1 interface 424 lookup results are stored in the PRAM1 result FIFO so work can immediately start on the next packet. According to one embodiment, PRAM lookups for a packet take 3 clocks.
  • the PRAM2 interface 428 retrieves data associated with successful CAM2 interface 416 source address lookups from the CAM2 result FIFO 422 , performs lookups to obtain additional forwarding information stored in the PRAM2 430 memory bank, and stores the results in the PRAM2 result FIFO 434 .
  • the receive packet evaluator 436 extracts the data from the PRAM1 result FIFO 432 , PRAM2 result FIFO 434 , and the RX scheduler FIFO 407 .
  • the receive packet evaluator 436 uses this information to construct the internal hardware header used to forward a packet through the chassis with the most advanced forwarding in this aspect permitting total destination address/VLAN/TOS replacement and packet header modification to support hardware packet routing.
  • the internal hardware header comprises sixteen bytes.
  • the receive packet evaluator 436 also determines the priority level of the packet according to the CAM and PRAM lookups and may optionally adjust the packet priority according to whether the packet is VLAN tagged or contains IPv4 TOS fields. The priority level is inserted into the internal hardware header of the packet.
  • the receive packet evaluator 436 notifies the dual-port interface 440 that processing is complete and passes the new internal hardware header to the dual-port interface 440 for integration with the packet data stored in the receive data FIFO 438 .
  • the dual-port interface 440 reads from the receive data FIFO 438 , applying packet modifications to incorporate the new hardware packet header and stores this packet data in the receive dual-port memory 110 .
  • the dual-port interface 440 also detects the end of packet (“EOP”) signal and issues a receive packet processing completion notification to the backplane manager 112 so the backplane manager 112 will know to retrieve the packet. If a packet is flagged as bad (for example, an invalid cyclic redundancy check) the buffer is instead immediately recycled for the next packet and the current packet is deleted.
  • EOP end of packet
  • FIG. 5 presents a block diagram depicting the operations of the RXPP 402 presented in FIG. 4 more discretely.
  • Data flow commences with the receive lookup handler 501 receiving packet data from the XGMAC interface 404 as illustrated in FIG. 4 .
  • the XGMAC interface 404 parses data received from the MAC 102 and sends only the packet header information to the receive lookup handler 501 .
  • the receive port tracker 502 examines the port information contained in the packet header to ensure that any VLAN information tags contained in the packet header will be accepted at the destination address port. If the destination address port is not configured to accept the packet header VLAN information or lack thereof, then the receive lookup handler 501 either sets an error bit in the packet header if debugging is supported or the packet is discarded. Alternatively, the receive lookup handler 501 will strip the VLAN tag from its field in the packet and store the VLAN tag in the internal hardware packet header for future use.
  • the receive lookup handler 501 checks the protocol used to encode the packet and classifies the packet accordingly in block 504 .
  • Examples of encoding protocols include IP, IP ARP, IPv4, IPv6, 802.3, IPX RAW, IPX LLC, IPX 8137, IPX SNAP, Appletalk, Appletalk ARP, NetBios, IP SNAP, and IP ARP SNAP.
  • This information is used to assemble an internal hardware packet header to be appended to the packet for use in forwarding the data internally throughout the chassis of the switch.
  • This additional information is passed from the receive lookup handler 501 to the RX scheduler FIFO 522 .
  • the RX scheduler FIFO 522 holds this information until the CAM and PRAM lookups are completed on the destination and source addresses extracted by the receive lookup handler 501 from the packet header.
  • the receive lookup handler 501 also forms the CAM lookups and builds part of the hardware packet header in block 506 .
  • the receive lookup handler 501 extracts source and destination address information from the packet header for use in the CAM lookups.
  • the internal state of the receive lookup processor 501 containing this information is then passed to the CAM lookup FIFO 508 , which is a memory buffer that permits the receive lookup processor 501 to start processing the next packet received from the XGMAC interface 404 . Packet processing is thus pipelined allowing the receive lookup processor 501 to continue efficiently processing packets without waiting for the CAM interface 509 to become available.
  • the CAM interface 509 When the CAM interface 509 becomes available, it fetches the address data stored in the CAM lookup FIFO 508 as shown in block 510 .
  • the CAM interface 509 dispatches requests regarding data in block 512 to the external ternary CAM memory 516 that contains a data array of values against which to perform lookups.
  • the CAM interface 509 is pipelined and supports cycling lookups for multiple packets to the external ternary CAM 516 memory since it takes longer than four clocks for the external CAM 516 to respond.
  • Block 514 illustrates a programmable delay incorporated into the CAM interface 509 pipeline that compensates for this delay while the CAM lookup is being performed.
  • additional forwarding information regarding disposition of the packet is available in the PRAM memory 530 .
  • Forwarding information might include details such as the destination port of the packet, the port mirror requirement, the packet type, VLAN handling information, packet prioritization data, multicast group membership, and/or other similar packet data known in the art.
  • the CAM array 516 entry also contains a link to the memory address of the additional forwarding information stored in the PRAM memory 530 . This link is returned by the CAM memory 516 as shown in block 518 and stored in the CAM result FIFO 520 until the PRAM interface 523 is available to perform lookups.
  • the PRAM interface 523 When the PRAM interface 523 becomes available, it fetches the link to the address in the PRAM memory 530 that is stored in the PRAM lookup FIFO 520 as shown in block 524 . In block 526 , the PRAM interface 523 dispatches requests to retrieve the additional forwarding information for the packet to the external PRAM memory 530 .
  • the PRAM interface 523 is pipelined and supports cycling lookups for multiple packets to the external PRAM memory 530 since it takes multiple clocks for the external PRAM memory 530 to return results from a lookup.
  • Block 528 illustrates a programmable delay incorporated into the PRAM interface 523 pipeline that compensates for this delay while the PRAM lookup is being performed.
  • the external PRAM 530 returns the additional forwarding information in block 532 and these results are stored in the PRAM result FIFO 534 until the receive packet evaluator 535 is available.
  • the receive packet evaluator 535 fetches data from the PRAM result FIFO 534 and the receive scheduler FIFO 522 .
  • the receive packet evaluator 535 evaluates this information in block 538 and uses the results to construct the internal hardware packet header in block 540 .
  • the internal hardware packet header is used to forward the packet through the chassis among other blades inserted into slots on the backplane. The most advanced forwarding in this aspect permits total destination address/VLAN/TOS replacement and packet header modification to support hardware packet routing.
  • the internal hardware header comprises sixteen bytes.
  • the receive packet evaluator 535 notifies the dual-port interface 542 that processing is complete and passes the new internal hardware header to the dual-port interface 542 for integration with the packet data stored in the receive data FIFO 438 , as illustrated in FIG. 4 .
  • the dual-port interface 542 reads from the receive data FIFO 438 applying packet modifications to incorporate the new hardware packet header for internally forwarding the packet through the chassis of the switch and stores this packet data in the receive dual-port memory 110 .
  • the receive dual-port memory is organized as four large FIFOs corresponding to four exemplary priority levels.
  • the dual-port interface 440 also detects the end of packet (“EOP”) and issues a receive packet processing completion notification to the backplane manager 112 so the backplane manager 112 will know to retrieve the packet. If a packet is flagged as bad (for example, an invalid cyclic redundancy check) the packet is deleted and the buffer is recycled for the next packet.
  • EOP end of packet
  • FIG. 6 presents a block diagram depicting the components of the backplane manager 112 as illustrated in FIG. 1 .
  • Components of the backplane manager 602 are embodied in the backplane manager chip.
  • the backplane manager chip 602 comprises an FPGA.
  • the backplane manager 602 is responsible for retrieving data from the receive dual-port memory 610 , determining backplane destinations for this data, and sending this data to those destinations.
  • the backplane manager 112 also manages four large FIFOs stored in the external dual-port memory 610 . These FIFOs store data according to priority levels by which the data is to be processed by the backplane manager 112 .
  • the receive done handler 604 receives EOP information from the receive packet processor 104 , including information regarding packet length and packet priority. This information is used to assist the receive done handler 604 in tracking receive dual-port memory 110 utilization for the four priority levels and scheduling packets for dispatch by the transmit queue dispatch 606 . If the backplane manager 602 or the receive dual-port memory FIFOs 610 are running low on resources, the receive done handler 604 sends a throttle control back to the receive packet processor 104 .
  • the transmit queue dispatch 606 is responsible for ordered packet dispatch from the four priority levels of the receive dual-port memory FIFOs 610 .
  • the transmit queue dispatch 606 receives packet length and priority information from the receive done handler 606 and uses this information to schedule packet retrieval from the dual-port RAM 610 by the dual-port interface 608 according to prioritization algorithms contained in the transmit queue dispatch 606 .
  • absolute priority is used with higher priority packets being unconditionally transmitted before any packets of lower priority. Absolute priority, however, is not always desirable.
  • some fraction of the transmission bandwidth available to the backplane manager 112 is dedicated to lower priority packet transmission regardless of whether higher priority packets are also pending because packets are often received by the invention faster than they can be transmitted. If some bandwidth were not allocated to lower priority packets in this manner, a bottleneck might be created with lower priority packets not being transmitted due to higher priority packets monopolizing all available transmission bandwidth. Packets are thus scheduled and posted for use by the transmit queue dispatch 606 .
  • the dual-port interface 608 fetches data from the receive dual-port memory 610 based on instructions received by the transmit queue dispatch 606 .
  • the dual-port interface 608 extracts a forwarding identifier (“FID”) from the packet and sends the FID to the FID lookup interface 612 .
  • the FID is an abstract chassis/system wide number used to forward packets. Each packet type has a FID to instruct the blade how to handle a given type of packet. This allows each blade in the chassis to look at the FID separately to decide how to individually forward the packet.
  • the FID lookup interface 612 translates the FID received from the dual-port interface 608 into a port mask by performing a lookup against addresses stored in the external FID RAM 614 .
  • the port mask is a multi-bit field representing a port on the blade and also other possible backplane slot destinations in the device chassis. According to one embodiment, the port mask is an 8-bit field representing a 10 Gigabit Ethernet port on the blade and seven other possible backplane slot destinations.
  • the FID lookup takes a number of clock cycles to complete during which time read data is posted to the delay FIFO 616 by the dual-port interface 608 .
  • the FID lookup by the FID lookup interface 612 into the external FID RAM 614 requires a delay of six clocks to complete in order to resume processing the data.
  • the FID lookup is completes and the results are passed from the FID lookup interface 612 to the merge port mask 618 .
  • Read data stored in the delay FIFO 616 is also passed to the merge port mask 618 .
  • the merge port mask 618 integrates the read data with the appropriate FID lookup port mask result and other port masks as set forth below to ensure that the data is transmitted to all intended destinations.
  • the merge port mask 618 takes the FID lookup port mask result and combines it with CPU and monitor information stored in configuration registers of the backplane manager. For example, a FID indicates a physical destination or possibly a list of destinations, but the receive packet processor 104 might have determined that the CPU also needs a copy of the data and therefore sets the CPU flag for combination with the FID lookup port mask by the merge port mask 618 . Alternatively, when a packet needs to be sent to a monitor port for network debugging or similar purpose, the monitor port mask is combined with the FID port mask. The merge port mask 618 thus generates a “qualified” port mask indicating all destinations for which the packet data is intended.
  • the merge port mask 618 may also apply source port suppression.
  • the blade that receives the data packet is listed as part of a FID port mask; source port suppression conditionally prevents the blade from retransmitting packets it just received. For example, this might occur in a broadcast situation where packets with unknown addresses are sent to all ports.
  • the merge port mask 618 stores the final result in the receive data FIFO 620 enabling the merge port mask 618 to process the next packet without waiting for the backplane FIFO dispatch 624 to become available.
  • the backplane FIFO dispatch 624 reads data from the receive data FIFO 620 , duplicating the data for each destination indicated in the qualified port mask.
  • the backplane FIFO dispatch 624 restructures the data into a format required by the backplane, generates backplane state and slot information, and posts the results into the backplane data FIFO 626 .
  • the backplane data FIFO 626 also acts as an asynchronous boundary between the backplane manager 602 core clock and the actual backplane clock. By posting the results in the backplane data FIFO 626 , the backplane FIFO dispatch 624 can process the next packet without waiting for the backplane dispatch 628 to become available.
  • data posted to the backplane data FIFO 626 is equivalent to two backplane transfers since the backplane manager runs at approximately one-half the clock speed of the backplane interface 114 .
  • the backplane dispatch 628 reads data from the backplane data FIFO 626 and outputs the data to the backplane via the backplane interface 114 .
  • the backplane dispatch 628 reads data from the backplane data FIFO 626 suitable for more than one transfer because the ratio of the backplane interface 114 clock speed and the clock speed of the backplane manager 602 is not identical.
  • the backplane dispatch 628 reads the number of transfers from the backplane data FIFO 626 that fully utilizes the transmission capacity of the backplane interface 114 . For example, if the clock speed of the backplane interface 114 is double that of the backplane manager 602 , then the backplane dispatch 628 will read two transfers from the backplane data FIFO.
  • the backplane dispatch 628 also monitors backplane status and directs backplane transmission rates since it is possible for a backplane slot destination to become congested or otherwise unavailable. For example, if a plurality of blades comprising a single chassis are devoting all of their transmission capacities to a single blade, then they may overload the destination blade. Such a case might occur when two blades both transmit at 8 Gbps to a single destination blade that, according to the capacity of a backplane slot, can only receive 8 Gbps it total. The two blades would have to throttle back transmissions to the destination blade to 4 Gbps to avoid congestion.
  • the transmission accumulator 116 collects packets from the backplane and organizes them into per-source, per priority transmit FIFOs stored in the transmit dual-port memory 118 .
  • Components of the transmission accumulator are embodied in the transmission accumulator chip 702 .
  • the transmission accumulator chip 702 comprises an FPGA.
  • the backplane front end 704 passes received data to the backplane slot receive accumulator 706 .
  • the backplane slot receive accumulator 706 is divided into a series of equal storage structures or memory buffers, with one buffer allocated for each slot or source on the chassis of the device. According to one embodiment of the invention, the backplane slot receive accumulator 706 is divided into eight buffers for receipt of data.
  • the backplane slot receive accumulator 706 When a particular quantity of data is received into one of the backplane slot receive accumulator 706 buffers, the backplane slot receive accumulator 706 notifies the backplane data polling logic 708 to indicate the buffer and priority of the data being stored. In one embodiment of the invention, the backplane slot receive accumulator 706 waits to notify the backplane data polling logic 708 until 32 bytes of data have been received in a bucket and transfers between the two components thus comprise 32 bytes. If the backplane slot receive accumulator 706 is full, then the transmission accumulator is congested and no longer accepts data until the congestion is relieved.
  • the backplane data polling logic 708 reads data from the backplane slot receive accumulator 706 and organizes data according to source and priority. If packets are aborted from the backplane, the backplane data polling logic 708 deletes the packet in order to avoid propagation of the packet to the TXPP 120 .
  • the backplane data polling logic 708 processes the data and the final result is stored in the backplane receive FIFO 710 , enabling the backplane data polling logic 708 to process the next packet without waiting for the dual-port interface 712 to become available.
  • the backplane receive FIFO 710 also permits dataflow through the asynchronous boundary between the backplane data polling logic block 708 and the dual-port interface 712 .
  • the dual-port interface 712 reads data from the backplane receive FIFO 710 and stores this packet data in the transmit dual-port memory 118 .
  • the dual-port interface 712 also detects valid end-of-packet (“EOP”) indications and notifies the TXPP 120 via transmission of an EOP message that a packet is available in the transmit dual-port memory 118 .
  • the transmit dual-port memory 118 also comprises a series of FIFOs similar to the receive dual-port memory 110 .
  • the transmit dual-port memory 118 has four FIFOs for each buffer of the backplane slot accumulator 706 , thereby comprising 28 FIFOs for these buffers, plus an additional four FIFOs for the OAR path, yielding a total of 32 FIFOs.
  • FIG. 8 depicts a block diagram of the components of the transmit packet processor 120 as illustrated in FIG. 1 a .
  • Components of the TXPP 120 are embodied in the TXPP chip 800 .
  • the TXPP chip 800 comprises an FPGA.
  • the TXPP 800 is responsible for retrieving data from the transmit dual-port memory 803 , determining network destinations for this data and sending data to identified destinations.
  • the TXPP 120 strips hardware header forwarding information used to route packets throughout the chassis of the switch and replaces this information with header forwarding information necessary to route packets over the network.
  • the TXPP 120 also manages the FIFOs priority queues stored in the transmit dual-port memory 803 . These FIFOs store data according to priority levels by which the data is to be processed by the TXPP 800 .
  • the transmit done handler 801 receives EOP information from the TX accumulator 116 , including information regarding packet length and packet priority. This information is used to assist the transmit done handler 801 in tracking transmit dual-port memory 803 utilization for the four priority levels and scheduling packets for dispatch in the transmit queue dispatch 802 . The transmit done handler 801 notifies the transmit queue dispatch 802 regarding packet availability and priority.
  • the transmit queue dispatch 802 is responsible for ordered packet retrieval and dispatch from the four priority levels of the transmit dual-port memory 803 FIFOs.
  • absolute priority is used with higher priority packets being unconditionally transmitted before any packets of lower priority. Absolute priority, however, is not always desirable.
  • some fraction of the transmission bandwidth available to the TXPP 120 is dedicated to lower priority packet transmission regardless, of whether higher priority packets are also pending because packets are often received by the invention faster than they can be transmitted. If some bandwidth were not allocated to lower priority packets in this manner, a bottleneck might be created with lower priority packets not being transmitted due to higher priority packets monopolizing all available transmission bandwidth. Packets are thus scheduled and posted for use by the dual-port handler 804 .
  • the dual-port handler 804 fetches the data from the transmit dual-port memory 803 according to instructions received from the transmit queue dispatch 802 . At the start-of-packet boundary, the dual-port handler 804 extracts the FID from the packet and sends the FID to the FID lookup block 808 . The dual-port handler 804 also extracts any VLAN tags from the packet and sends this information to the multicast start offset lookup block 806 .
  • the FID received from the dual-port handler 804 is used to perform a lookup against a FID table.
  • the FID lookup block 808 functions similarly to the interaction between the FID lookup interface 612 and the FID RAM 614 as presented in FIG. 6 . Accordingly, the results obtained from the FID table indicate how the packet should be handled for transmission by the receiving blade. For example, the FID might indicate that although the packet may have arrived at the blade, the packet should not be transmitted by the blade. This might occur in a broadcast situation where a packet is broadcast to all blades within a chassis.
  • the FID lookup block 808 determines that a packet has been erroneously received in this manner, the packet is deleted and no longer processed by the TXPP 120 . In this sense, the FID lookup block 808 also functions as a transmit filter to ensure that only valid packets are actually sent out over the network.
  • Results of the FID lookup are stored in the delay FIFO 810 . This permits the FID lookup block 808 to begin processing the next packet without waiting for the context track and internal header removal block 814 to become available. Pipelining processing data in this manner allows packet processing operations by the TXPP 120 to complete faster.
  • a VLAN is a local area network identifier that maps locations based on a basis other than physical location. For example, devices attached to a VLAN might be grouped according to department, division, application, etc. Devices that are part of the same VLAN behave as if they were connected to the same wire even though they may actually be physically connected to different segments of a LAN. VLANs are configured using software protocols rather than in hardware and are therefore extremely flexible with respect to implementation. For example, a computer may be moved to a different physical location on the same VLAN without any hardware reconfiguration.
  • VLAN tags placed in a header field indicate whether a packet is intended for routing over a VLAN. Additionally, the VLAN tag in the header may also indicate that a packet is intended for VLAN multicasting. VLAN multicasting occurs when a packet is sent over a VLAN to more than one destination address. Since the header of each packet must be changed to reflect each destination address during VLAN multicasting, this process can be very resource intensive when performed using software.
  • the multicast start offset lookup block 806 supports hardware VLAN multicast replication.
  • the multicast start offset lookup block 806 examines the VLAN tag extracted from the packet header and performs a lookup against a table stored in RAM in the multicast start offset lookup block 806 . If the packet VLAN tag matches an entry in the table, additional information pertaining to that VLAN is available at an address location in a memory array stored in the multicast replacement lookup block 812 .
  • multicast replacement lookup block 812 might contain information to assist with setting unique VLAN ID values, VLAN priorities, and TXA/SAS/srcport suppression behaviors for each packet transmitted over the VLAN.
  • the multicast start offset lookup block 806 takes the address to the memory array location of the multicast replacement lookup block 812 and stores this result in the delay FIFO 810 . This permits the multicast start offset lookup block 806 to begin processing the next packet without waiting for the context track and internal header removal block 814 to become available. Pipelining processing in this manner allows packet processing operations by the TXPP 120 to complete faster.
  • the delay FIFO 810 also stores values from the FID lookup block 808 and the multicast start offset lookup block 806 for retrieval by the multicast replacement lookup block 812 and the context track and internal header removal block 814 .
  • the multicast replacement lookup block 812 retrieves the results of the multicast start offset lookup block 806 calculations from the delay FIFO 810 for processing packets subject to VLAN routing.
  • the multicast replacement lookup block 812 takes the address of the memory array location contained in the multicast replacement lookup block 812 and retrieves the additional information that is stored at that location pertaining to routing over the VLAN tag referenced in the packet header. This information is passed to the context track and internal header removal block 814 for incorporation into the outgoing packet header.
  • the context track and internal header removal block 814 removes the internal hardware header from the packet and begins the process of assembling an outgoing packet header suitable for transmission over the network.
  • the context track and internal header removal block 814 passes information regarding any data offset to the header which may have occurred to the barrel shifter 816 .
  • the context track and internal header removal block 814 passes information regarding the TXA/PTYPE to the SA substitution and L3 assist block 818 .
  • the context track and internal header removal block 814 passes information regarding the packet VLAN ID and the VLAN tag status to the VLAN insertion block.
  • the barrel shifter 816 normalizes any changes to the packet header that occurred during internal routing through the chassis.
  • One function of the internal hardware header of a packet is to permit the CPU to add an encapsulation to a packet. Encapsulation is used by the CPU to complete operations more efficiently by avoiding having to copy the entire packet into CPU memory and then writing the packet back to the buffer pool. Instead, the CPU performs a small modification to the packet header. For example, this might occur when the CPU determines that a packet must be forwarded, but that the CPU must first add data to the header before forwarding can take place. Alternatively, the CPU might also remove data from the header temporarily to assist with forwarding.
  • the CPU might move data within the packet header into a non-standard format.
  • the destination address might appear at the wrong location within the packet for transmission over the network.
  • the barrel shifter 816 analyzes the composition of the packet header and shifts the data within the header to normalize it and correct for any CPU modifications that might have occurred.
  • the packet header data is then in a standard format and is passed to the SA substitution and L3 assist block 818 for further processing.
  • the SA substitution and L3 assist block 818 performs further modifications on the packet header to prepare the packet for transmission over the network.
  • the SA substitution and L3 assist block 818 replaces the MAC address that is required for routing packets.
  • each packet header contains a destination address and a source address. The source address must be changed on transmit to reflect which port the packet is being broadcast from.
  • the SA substitution and L3 assist block 818 also modifies other Layer 3 header fields as required, such as changing the IPv4/IPX time to live value or the checksum.
  • the packet is passed to the VLAN insertion block 820 for further processing.
  • VLAN tags that were removed on receipt anywhere in the chassis are stored in the internal hardware header for future use on transmission.
  • the VLAN insertion block 820 takes the internal hardware header information that is passed from the context track and internal header removal block 814 and reintroduces this information into the outgoing packet header as appropriate. This information includes the packet VLAN ID and the Tag Status.
  • the packet When the outgoing header packet is reassembled for transmission over the network, the packet is stored in the TX FIFO 822 prior to being passed to the XGMAC interface 824 .
  • the TX FIFO 822 enables the VLAN insertion block 820 to begin processing the next packet without having to wait for the XGMAC interface to become available and enables faster operation by the VLAN insertion block 820 .
  • the TX FIFO 822 permits data flow though asynchronous boundaries.
  • the TXPP 120 operates at a different speed than the MAC 102 .
  • Data flow must be synchronized between asynchronous components so the TX FIFO 822 acts as a bridge between these components.
  • the MAC 102 operates at a 156.25 MHz clock and the TXPP operates at only a 66 MHz clock.

Abstract

The system of the present invention provides data transmission speeds at or in excess of 10 gigabits per second between one or more source devices and one or more destination devices. The system comprises a media access control (MAC) interface to facilitate receipt and transmission of packets over a physical interface. A first field programmable gate array is coupled to the MAC interface and operative to receive packets from the MAC interface and configured to perform initial processing of packets, which are dispatched to a first memory. A second field programmable gate array is operative to retrieve packets from the first memory and configured to compute an appropriate destination, which is used to dispatch packets to a backplane. A third field programmable gate array is provided that is operative to receive packets from the backplane and configured to organize the packets for transmission, which are dispatched to a second memory. A fourth field programmable gate array is coupled to the MAC interface and operative to retrieve packets from the second memory and configured to schedule the transmission of packets to the MAC interface for transmission to one or more destination devices.

Description

    CROSS-REFERENCE TO RELATED APPLICATION
  • This application is a continuation of and claims priority under 35 U.S.C. §120 to U.S. patent application Ser. No. 10/139,831, entitled “SYSTEM ARCHITECTURE FOR VERY FAST ETHERNET BLADE,” filed May 6, 2002, currently pending, assigned to the same assignee as the present application, and incorporated herein by reference in its entirety.
  • COPYRIGHT NOTICE
  • A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent files or records, but otherwise reserves all copyright rights whatsoever.
  • BACKGROUND OF THE INVENTION
  • The invention described herein relates to computer networking and, in particular, to improved methods, systems, and software for routing data at high speeds through a switch or other network routing device.
  • The explosive growth of the Internet has brought more and more users online every day, and computer networks have assumed an increasingly important role in today's highly interconnected world. As users increasingly rely on the network to deliver required data, network traffic has increased exponentially. Moreover, with the adoption of new and more bandwidth-intensive applications, enormous burdens are placed on network infrastructure. Network administrators are thus constantly seeking faster and more reliable methods and equipment to transport data to accommodate these demands.
  • Ethernet, one of the earliest networking protocols, is today the most widely used method to transport data in computer networks. Robert Metcalf and David Boggs developed Ethernet as an experiment at the XEROX Palo Alto Research Center in 1973. At Ethernet's inception, the struggle to accommodate users needs for bandwidth had not yet started. As network traffic demands at this time were quite low, Ethernet initially had a data transmission rate of 2.94 megabits per second (Mbps).
  • Metcalf, however, recognized the potential for rapid network growth and posited a theorem now known as “Metcalf's Law” which states that the value of a network expands exponentially as the number of users increase. Gordon Moore, an expert in the field of semi-conductor development, posited another theorem known as Moore's Law which states that the power of microprocessors will double every 18 months and their price will be reduced by half. When taken together, these two laws predict rapid growth of networking technologies: as users join the network, more people will want to join at an exponential rate equivalent to the rise in value of the network, while processing technologies to support this growth and faster transport are constantly increasing at rapidly diminishing costs.
  • The evolution of Ethernet technologies has followed theory. The first commercial release of Ethernet occurred in 1979 with a transmission rate of 10 Mbps—more than a three-fold increase over the experimental system created just five years earlier. Ethernet went through a variety of standardizations during the 1980s and line rates remained constant at 10 Mbps while the technology matured. In 1995, however, Ethernet became available at 100 Mbps. In 1998, bandwidth jumped again to 1 gigabit per second (Gbps). Most recently, a new standard was adopted for Ethernet transmission rates at 10 Gbps representing a 100-fold increase in seven years.
  • Implementation of 10 Gbps network infrastructure requires overcoming significant hurdles not addressed by current advances in the art. For example, previous generations of Ethernet technology, although fast, had an ample number of clocks in which to perform packet analysis and retransmit data. With the rise of 10 Gbps Ethernet, however, calculations previously carried out over a given number of clocks must now be completed in a fraction of the time so that the desired bandwidth is in fact available.
  • There is thus a need for a systems and methods capable of efficiently accommodating data transfer rates over a network in excess of 10 Gbps.
  • BRIEF SUMMARY OF THE INVENTION
  • The present invention provides a switch or router for providing data transmission speeds up to 10 gigabits per second between one or more source devices and one or more destination devices. The switch includes a blade or board having several discrete integrated circuits embedded thereon, each performing one or more discrete functions required to meet the speed required for the switch. The blade includes a media access control interface (MAC) to facilitate receipt and transmission of packets over a physical interface. In one embodiment, the blade further includes four field programmable gate arrays. A first field programmable gate array is coupled to the MAC array and operative to receive packets from the MAC interface and configured to perform initial processing of packets. The first field programmable gate array is further operative to dispatch packets to a first memory, such as a dualport memory.
  • A second field programmable gate array is operative to retrieve packets from the first memory and configured to compute an appropriate destination and to dispatch packets to a backplane. A third field programmable gate array is operative to receive packets from the backplane and configured to organize the packets for transmission and to dispatch packets to a second memory. A fourth field programmable gate array is coupled to the MAC interface and operative to retrieve packets from the second memory and to schedule the transmission of packets to the MAC interface for transmission to one or more destination devices.
  • According to an alternative embodiment, the invention comprises a switch or router for providing data transmission speeds up to 10 gigabits per second between one or more source devices and one or more destination devices through the use of two sets of one or more field programmable gate arrays. A first set of one or more field programmable gate arrays is coupled to a media access control (MAC) interface and a memory structure, the MAC interface used to facilitate the receipt and transmission of packets over a physical interface. The first field programmable gate array set is operative to receive and transmit packets from and to the MAC interface. The first field programmable gate array set is configured to perform initial processing of received packets and to schedule the transmission of packets to the MAC interface for transmission to one or more destination devices, in addition to dispatching and retrieving packets to and from the memory structure.
  • This embodiment of the invention also comprises a second set of one or more field programmable gate arrays coupled to the memory structure and a backplane. The second field programmable gate array set is operative to retrieve packets from and dispatch packets to the memory structure, and configured to compute an appropriate destination and organize packets for transmission. The second field programmable gate array set is further operative to receive and dispatch packets from and to the backplane.
  • BRIEF DESCRIPTION OF THE DRAWINGS
  • The invention is illustrated in the figures of the accompanying drawings which are meant to be exemplary and not limiting, in which like references are intended to refer to like or corresponding parts, and in which:
  • FIG. 1 is a block diagram of a system architecture for an Ethernet blade in accordance with one embodiment of the present invention;
  • FIG. 1A is a block diagram of a system architecture for an Ethernet blade in accordance with a second embodiment of the present invention;
  • FIG. 2 is a high level flow diagram of a connection of a packet processor component of the present invention to an outside network, in accordance with one embodiment of the present invention;
  • FIG. 3 is a block diagram of receive and transmit packet processors of one embodiment of the present invention;
  • FIG. 4 is a block diagram of a receive packet processor in accordance with one embodiment of the present invention;
  • FIG. 5 is a flow diagram showing the data flow in the receive packet processor of FIG. 4 in accordance with one embodiment of the present invention;
  • FIG. 6 is a block diagram of a backplane manager in accordance with one embodiment of the present invention;
  • FIG. 7 is a flow diagram showing the data flow in a transmission accumulator in accordance with one embodiment of the present invention; and
  • FIG. 8 is a block diagram of a transmit packet processor component in accordance with one embodiment of the present invention
  • DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
  • Embodiments of methods and systems according to the present invention are described through reference to FIGS. 1 through 8. Turning to FIG. 1, a block diagram is presented depicting a high-level schematic of the components of one possible embodiment of the invention to allow data transfer speeds at or in excess of 10 gigabits per second. As shown, the invention comprises a printed circuit board (“PCB”) 10 used to house and provide interconnections for a media access controller (“MAC”) 12, a packet processor (“PP”) 14, one or more content addressable memory (“CAM”) controllers 16, one or more controllers for random access memories containing parameter information (“PRAM”) processors 18, a receive dual-port memory buffer 20, a transmit dual-port memory buffer 22, a transmission manager 24, and a backplane interface 26.
  • The PCB10 provides a surface on which to place other components of the invention. The PCB 10, also known as a “blade” or “module”, can be inserted into a slot on the chassis of a network traffic management device such as a switch or a router. This modular design allows for flexible configurations with different combinations of blades in the various slots of the device according to differing network topologies and switching requirements. Furthermore, additional ports for increased network connectivity may easily added by plugging additional blades into free slots located in the chassis.
  • An example of such a switch is the BigIron® switch produced by Foundry Networks, Inc. of San Jose, Calif. The BigIron switch chassis consists of multiple distributed switching modules each of which contain a high-bandwidth memory system for scalable chassis bandwidth. The local switching fabric of the BigIron switch houses the forwarding engines, provides packet-level examination and classification based on Layer 2/3/4 information, and performs IP subnet look-ups and packet modifications of IP and IPX packets.
  • The MAC 12 is the interface by which data is received and transmitted to and from the network. In one embodiment, such network data comprises Ethernet packets. The MAC 12 forwards received packets to the PP 14 for further processing and also receives packets for transmission to the network from the PP 14. The MAC 12 performs any data conversions required for network data to be processed by the PP 14 for routing within the device chassis and for data processed by PP 14 to be transmitted to the network. For example, in one embodiment of the invention, the MAC 12 performs data conversions because network data comprises 32 bit double data rate (“DDR”) data, while the PP 14 processes only 64 bit single data rate (“SRD”) data. The MAC is typically responsible for data validity checking, as well as data gathering.
  • The PP 14 is a processor chip responsible for receiving packets from the MAC 12 and processing them for forwarding through the device chassis, as well as for processing packets received from the device chassis intended for transmission over the network. These two functions, while performed on the same chip, are preferably performed simultaneously and in parallel. There are thus, in a sense, two pipelines in the PP 14: a receive pipeline for processing network packets intended for transmission within the chassis and a transmit pipeline for processing internally routed packets intended for transmission over the network.
  • In one embodiment of the invention, the packet processor is a field programmable gate array (“FPGA”), which is an integrated circuit that can be programmed in the field after manufacture. An advantage of using FPGAs with the invention is that an FPGA provides significant flexibility over an application specific integrated circuit (“ASIC”) and is also much less expensive to prototype and implement.
  • The receive pipeline of the PP 14 is responsible for packet classification, performing CAM and PRAM lookups, generating packet headers for forwarding packets through a chassis, and preparing packet modifications. Network packets are received by the PP 14 from the MAC 12 in multi-byte bursts based on scheduling priorities determined at the MAC 12. The PP 14 examines packets and extracts packet forwarding information from the packets such as the destination address (“DA”) of the packet and the source address (“SA”) of the packet. The PP 14 extracts the type of service (“TOS”), whether the packet has a virtual local area network (“VLAN”) tag, session related data such as in the case of IPv4 or IPX data, and other additional Layer 3 and Layer 4 information useful in routing the packet through the chassis. The PP 14 passes this forwarding information extracted from the packet header to a CAM processor 16 for further processing.
  • The CAM controller or processor 16 takes information forwarded by the PP 14 and performs a lookup comparing this information to data stored in a local memory of the CAM processor 16. If the information matches information stored in the local memory of the CAM processor 16, additional forwarding information regarding disposition of the packet is available in the local memory of the PRAM processor 18 and can be retrieved for future incorporation into the packet header.
  • When such successful CAM matches occur, the PRAM processor 18 retrieves additional forwarding information from its local memory for incorporation into the header of the packet. The packet is reformatted with a new internal hardware header for routing the packet within the chassis and stored in the receive dual-port memory buffer 20 for processing by the transmission manager. This internal hardware header is also sometimes referred to as a chassis header.
  • An important technique in implementing the invention is pipelining. Pipelining is an advanced technique used by processors, wherein a processor begins executing a subsequent instruction before a prior instruction has finished executing. Accordingly, a processor can have multiple instructions processing in its “pipeline” simultaneously with each instruction at a different processing stage.
  • The pipeline is divided into processing segments, with each segment executing its operation concurrently with the other segments. When a segment completes its operation, it passes the result to the next segment in the pipeline and fetches data for processing from the preceding segment. Often, temporary memory buffers are used to hold data values between segments, which allows operations to complete faster since each segment no longer waits for the other segment to finish processing prior to handing off data. The final results of the process emerge at the end of the pipeline in rapid succession.
  • The receive dual-port memory 20 (as well as its counterpart, the transmit dual-port memory 22) acts as a pipeline buffer in the embodiment of the invention depicted in FIG. 1. The receive dual-port memory 20 enables the PP 14 to store processed data and continue processing the next packet without having to wait for the transmission manager 24 to become available, thereby expediting operations of both the PP 14 and the transmission manager 24. Other buffers are used throughout the invention and in its various components to achieve pipelining and faster packet processing in an analogous manner.
  • The transmit pipeline of the PP 14 retrieves data from the transmit dual-port memory 22 according to a programmable priority scheme. The PP 14 extracts network destinations from the dual-port data and reassembles packet header forwarding information by removing any packet header modifications that take place in order to route the packet through the switch chassis. The PP 14 performs sanity checks on packet data to ensure that only those packets intended for transmission are passed on to the MAC 12.
  • Since packets routed through the chassis carry header information pertaining to forwarding within the chassis, this information must be removed and replaced with header forwarding information appropriate for routing over the network. After the proper network header forwarding information is reassembled and the chassis header information is removed, the PP 14 forwards the data to the MAC 12 for eventual transmission over the network to the intended address.
  • While the PP 14 handles traffic to and from the MAC 12 and conversions of packet headers between network packet headers and internal chassis packet headers, the transmission manager 24 handles traffic flow to and from the backplane interface 114. Like the PP 14, the transmission manager 24 is a processor chip that implements a dual pipeline architecture: a receive pipeline for network data to be internally routed within the device chassis and a transmit pipeline for internally routed data intended for network transmission. These two functions, while performed on the same chip, are preferably performed in parallel according to one embodiment of the invention. In one embodiment of the invention, the transmission manager 24 is an FPGA, although use of other processor types is within the scope of the invention.
  • The transmission manager 24 fetches network data intended for routing through the device chassis from the receive dual-port memory 20 and stores internally routed data intended for network transmission in the transmit dual-port memory 22. The receive pipeline of the transmission manager 24 retrieves data from the receive dual-port memory 20 according to instructions issued to the transmission manager 24 by the PP 14. The transmission manager 24 determines data transmission priority for the data retrieved and schedules transmissions to the backplane 26 according to this priority scheme. In one embodiment of the invention, there are four different priority levels assigned to data.
  • The transmission manager 24 extracts backplane destinations from data, and sends data to those destinations according to predetermined priority algorithms. Backplane destinations may comprise other blades in the chassis or, in some cases, may comprise the blade of the transmission manager 24 itself, which is called “one-armed routing.”
  • The transmit pipeline of the transmission manager 24 handles internally routed packets received from the backplane interface 26 and intended for transmission over the network. The transmission manager 24 collects packets from the backplane interface 26 and organizes them into per-source, per-priority transmit queues stored in the transmit dual-port memory 22. The transmission manager 24 notifies the PP 14 when a packet is stored in the transmit dual-port memory 22 and available for processing.
  • FIG. 1 a, presents a block diagram depicting a high-level schematic of the components of an alternative embodiment of the invention. As shown, the invention comprises a printed circuit board 100, a media access controller 102, a receive packet processor 104 (“RXPP”), one or more CAM processors 106, one or more PRAM memory processors 108, a receive dual-port memory buffer 110, a backplane manager 112, a backplane interface 114, a transmission accumulator (“TX accumulator”) 116, a transmit dual-port memory buffer 118, and a transmit packet processor (“TXPP”) 120.
  • The PCB 100 provides a surface on which to place many of the other components of the invention. The PCB 100, also known as a “blade” or “module”, can be inserted into one of a plurality of slots on the chassis of a network traffic management device such as a switch or a router. This modular design allows for flexible configurations with different combinations of blades in the various slots of the device according to differing network topologies and switching requirements.
  • The MAC 102 is the interface by which a blade receives and transmits data to and from the network. In one embodiment, such network data comprises Ethernet packets. The MAC 102 forwards received packets to the RXPP 104 for further processing and receives packets for transmission to the network from the TXPP 120. The MAC 102 also performs any data conversions required for network data to be processed by the RXPP 104 or for data processed by TXPP 120 to be transmitted to the network. For example, the MAC 102 may perform data timing conversions where network data comprises 32 bit DDR data while the RXPP 104 and the TXPP 120 process only 64 bit SDR data.
  • The receive packet processor 104 is responsible for packet classification, performing CAM and PRAM lookups, generating packet headers for forwarding packets through a chassis, and preparing packet modifications. In one embodiment of the invention, the receive packet processor 104 is an FPGA. In an alternate embodiment of the invention, the RXPP 104 is an ASIC. Packets are received by the RXPP 104 from the MAC 102 in multi-byte bursts based on scheduling priorities determined at the MAC 102. The RXPP 104 examines packets and extracts packet forwarding information from a packet, such as the destination address of the packet and the source address of the packet. The RXPP 104 extracts the TOS, any defined VLAN tags, session related data such as in the case of Ipv4 or IPX data, and other additional Layer 3 and Layer 4 information useful in routing the packet through the chassis. The RXPP 104 passes this forwarding information to one of the CAM processors 106 for further examination.
  • The CAM processor 106 takes information forwarded by the RXPP 104 and performs a lookup, comparing received information to data stored in local memory of the CAM processor 106. If the comparison returns a match, additional forwarding information regarding disposition of the packet is stored in local memory of one of the PRAM processors 108 and can be retrieved for future incorporation into the packet header. The PRAM processor 108 retrieves additional forwarding information from its local memory for incorporation into the header of packet. The packet is then stored in the receive dual-port memory buffer 110 for processing by the backplane manager 112. Those of skill in the art will recognize that additional processing may be performed before storage in the receive dual port memory.
  • The receive dual-port memory 110 (as well as its counterpart, the transmit dual-port memory 118) acts as a pipeline buffer between processes. The receive dual-port memory 110 enables the RXPP 104 to store processed data and continue processing the next packet without having to wait for the backplane manager 112 to become available. Pipelining operation execution expedites processing of both the RXPP 104 and the backplane manager 112. Other buffers are used throughout the invention and within its various components to achieve pipelining and faster packet processing in this manner.
  • The next segment in the receive pipeline is the backplane manager 112. The backplane manager 112 is a processor designed for retrieving data from the receive dual-port memory buffer 110 and dispatching packets to the backplane interface 114. Data is retrieved from the receive dual-port memory 110 according to instructions issued to the backplane manager 112 by the RXPP 104. The backplane manager 112 determines data transmission priority for the data retrieved and schedules transmissions to the backplane 114 according to this priority scheme. According to one embodiment of the invention, there are four different priority levels assigned to data.
  • The backplane manager 112 extracts backplane destinations from received data; the data sent to indicated destinations according to programmable priority algorithms. Backplane destinations may comprise other blades in the chassis or, in the case of OAR, may comprise the blade of the backplane manager 112 that initially receives the data. When packets scheduled for OAR are detected, they are forwarded to the transmission accumulator 116 via the OAR data path as shown in FIG. 1 a. In one embodiment of the invention, the backplane manager 112 is an FPGA. In an alternate embodiment of the invention, the backplane manager 112 is an ASIC.
  • The transmit accumulator 116 is a processor that receives packet data from the backplane 114 intended for transmission. The transmit accumulator 116 collects packets from the backplane 114 and organizes them into per-backplane-source, per-priority transmit queues stored in the transmit dual-port memory 118. The transmit accumulator 116 notifies the TXPP 120 when data comprising a packet is stored in the transmit dual-port memory 118 and available for processing. In one embodiment of the invention, the transmit accumulator 116 is an FPGA.
  • The transmit packet processor 120 retrieves data from the transmit dual-port memory 118 according to a programmable priority scheme. The TXPP 120 extracts network destinations from the data and reassembles packet header forwarding information by removing any packet header modifications that took place in order to route the packet through the device chassis. The TXPP 120 performs sanity checks on packet data to ensure that only those packets intended for transmission are passed on to the MAC 102. Since packets routed through the chassis carry header information pertaining to forwarding within the chassis, this information must be removed and replaced with header forwarding information appropriate for routing over the network. After the proper network header forwarding information is reassembled and the chassis header information is removed, the transmit packet processor 120 forwards the data to the MAC 102 for eventual transmission over the network to the intended address. In one embodiment of the invention, the transmit packet processor 120 is an FPGA. In an alternate embodiment of the invention, the transmit packet processor 120 is an ASIC.
  • FIG. 2 presents a high-level schematic of one embodiment of the invention as it connects to a network, e.g., an optical network comprising fiber optic connections. The optics block 202 is the interface through which all network data traffic passes. The optics block 202 contains a transmitter for generating the optical signals to the network when data is received from the transceiver 204. In some embodiments, the transmitter might comprise a laser or a light emitting diode. The optics block 202 also contains a detector for receiving optical data traffic from the network. When optical data is received, a photodetector generates an electrical current that is amplified to level useable by the transceiver 204. The signal is then communicated to the transceiver 204 for further processing.
  • The transceiver 204 directs the transmission and receipt of signals to and from the optics block 202. The transceiver 204 receives electrical data signals intended for transmission to the MAC 206 and instructs the transmitter in the optics block 202 to generate optical signals corresponding to the electrical data signals. Conversely, the transceiver 204 receives electrical data signals from the optics block 202 and passes these signals to the MAC 206 for processing.
  • There are many asynchronous boundaries between the various components of the invention. For example, data passes to and from the transceiver 204 and the MAC 206 at a fixed speed. In one embodiment of the invention, the datapath 208 between the transceiver and the MAC 206 operates sending 4 clock signals along with 32 bit DDR data at 156.25 MHz. The datapath 212 between the MAC 206 and the packet processor 210, however, may operate at a different speed. For example, in one embodiment of the present invention, the datapath 212 between the MAC 206 and the packet processor 210 operates sending 4 clock signals along with 64-bit SDR at 66 MHz. Multiple clock signals are sent with the data and used to minimize timing differences between groups of data signals and a clock. In one embodiment of the invention, one clock signal is included per 8 bits of DDR data and one clock signal is included per 16 bits of SDR data. In addition to clock signals, control signals are also sent along with data to indicate packet boundaries and possible error conditions. In one embodiment of the invention, control signals are distributed across 4 clock groups of data.
  • Those skilled in the art will recognize that an important technique in managing the dataflow between these asynchronous boundaries is the use of FIFO buffers that permit the dataflow to remain synchronized. Given the extremely high rate of data transfer provided by the invention, conventional techniques for clock distribution, such as those known in the art and used in the case of personal computer boards, will not allow reliable capture and transfer of data between components of the invention operating according to different clocks. The invention, therefore, implements source synchronous clocking wherein the clock is sent along with the data.
  • When the clock arrives at the packet processor 210 from the MAC 206, for example, the clock is exactly in relationship according to the MAC 206, but the packet processor 210 can also capture the data on that clock via a FIFO. Data from the MAC 206 is captured inside a FIFO, which allows the packet processor to synchronize, in the presence of this data, between the source synchronous clock contained in the FIFO data and the clock the packet processor 210 is using at its core.
  • The invention uses source synchronous clocking in a symmetric manner. For example, data passing from the packet processor 210 to the MAC 206 is also captured in a FIFO to allow the MAC 206 to synchronize, in the presence of the FIFO data, between the source synchronous clock (of the packet processor 210 core) and the clock the MAC 206 is using at its core clock. In an alternative embodiment, the invention also implements differential source synchronous clocking which is known to those skilled in the art. Differential source synchronous clocking works in much the same manner as source synchronous clocking, except that two clock signals are sent with the data instead of one clock signal. The two clock signals, a high and low signal, are used to calculate a more precise approximation of the signal value being transmitted which those skilled in the art will recognize is used to reduce noise and generate more accurate data transmissions.
  • FIG. 3 is a block diagram depicting one embodiment of the components of the MAC 102 as presented in FIGS. 1 and 1 a. Components of the MAC 102 are embodied in the MAC processor chip 302. According to one embodiment of the invention, the MAC chip 302 is an FPGA. In an alternate embodiment of the invention, the MAC chip 302 is an ASIC. The MAC 102 is the interface between the network via the PHY transceiver 300 and the RXPP 104 and TXPP 120 packet processor chips. According to one embodiment of the invention, the MAC 102 communicates directly to the PHY layer transceiver 300 via a DDR interface and with the packet processor chips of the RXPP 104 and the TXPP 120 via an SDR interface.
  • The PHY transceiver 300 is the component applying signals to the network wire and detecting signals passing through the network wire. According to one preferred embodiment of the invention, the PHY transceiver 300 is a 10 Gigabit Ethernet transceiver transmitting and receiving 32 bit DDR data at 156.25 Mhz. Data received by the PHY transceiver 300 is passed to the receive front end 306 of the MAC 102. The receive front end 306 is an interface that receives data, which is passed to the receive block 304 for further processing. According to one preferred embodiment of the invention, the receive front end 306 receives 32 bit DDR data.
  • The receive block 304 performs a variety of tasks on data received from the receive front end 306 and is very flexible in operation. The receive block 304 internally converts data received from the receive front end 306 into a format suitable for transmission to the RXPP 104. According to one embodiment of the invention, the receive block converts 32 bit DDR data into 64 bit SDR data for transmission. The receive block 304 may also perform other tasks as required according to various embodiments of the invention such as verifying and extracting XGMII tokens, realigning bytes such that the start of packet (“SOP”) token is placed in a “lane zero” position, verifying SOP and EOP framing, detecting giant packets, verifying and optionally stripping packet cyclic redundancy checks, tracking the full suite of RMON statistics, and other useful operations.
  • The receive block 304 also generates flow control packets via the pause and flow control sync block 332. The receive block 304 operates off of the recovered source synchronous clocks contained in the incoming data packets received from the PHY transceiver 300. Other components of the MAC 102, including the transmit block 328, however, are operating off of an internal core clock generated locally. Although these two clocks are nominally the same frequency, there is some variance since they are not really the same clock and therefore tend to “drift” over time. This difference between the two clocks requires periodic synchronization of the receive block 304 and the transmit block 328 for the purposes of passing flow control messages to generate pause frames and avoid network congestion.
  • In such a scenario, the receive block 304 receives an incoming message from a remote source (to which the transmit block 328 is sending data) indicating that the remote source is becoming congested and requesting that the transmit block 328 pause transmission for a requested interval. The pause and flow control sync block 332 synchronizes the receive block 304 clock with the transmit block 328 clock to permit the receive block 304 to pass the pause frame request to the transmit block 328 and reduce the network congestion. Conversely, in the unlikely event that the receive FIFO RAM 308 becomes congested, the pause and flow control sync block 332 would synchronize the two clocks to permit the receive block 304 to instruct the transmit block 328 to start issuing flow control pause frames to a remote sender to reduce network congestion in the MAC 102.
  • The receive block 304 passes processed data to the receive FIFO RAM 308 via the write port 310 of the receive FIFO RAM 308 which enables the receive block 304 to process the next packet without waiting for the receive FIFO block 314 to become available. The receive FIFO RAM 308 is a two-port memory having a write port 310 that accepts incoming data from the receive block 304 and a read port 312 that transmits data stored in the receive FIFO RAM 308 to the receive FIFO block 314. The write port 310 and the read port 312 operate independently of each other thus permitting more efficient use of the receive FIFO RAM 308 by the receive block 304 and the receive FIFO block 314.
  • The FIFO RAM 308 further permits data flow though the asynchronous boundary. In one embodiment of the invention, the receive block 304 operates at a different speed than the receive FIFO block 314. Thus, the FIFO RAM 308 acts as a bridge, allowing data flow to be synchronized between these asynchronous components. For example, in the Foundry BigIron switch, the receive block 304 operates at a 156.25 MHz clock recovered from the arriving data and the FIFO block 314 operates on a locally generated 156.25 MHz clock that differs slightly and drifts in phase relationship over time.
  • To further reduce processing time, the receive block 304 starts streaming data into the receive FIFO RAM 308 when the receive block detects the start of a packet and stops streaming data into the receive FIFO RAM 308 when the receive block 304 detects the end of the packet. All of the packet processing components of the invention stream data into FIFOs in this manner which greatly reduces processing time since components are not required to wait until an entire packet is finished processing to start copying the packet into a FIFO.
  • The receive FIFO block 314 reads data stored in the receive FIFO RAM 308 via the read port 312. The receive FIFO block 314 also notifies the RXPP 104 that packet data is contained in the receive FIFO RAM 308 and available for transmission. This data is transmitted to the RXPP 104 for further processing. According to one embodiment of the invention, the receive block FIFO 314 transmits 64 bit SDR data to the RXPP 104.
  • In addition to the receive pipeline of the MAC 102 as set forth above, the MAC 102 also contains a transmit pipeline that operates in a similar fashion with similar flexibility. The transmit FIFO block 320 is the interface of the MAC 102 that receives data from the TXPP 120. According to one embodiment of the invention, the transmit FIFO block 320 receives 64 bit SDR data from the TXPP 120.
  • The transmit FIFO block 320 streams received data to the transmit FIFO RAM 322 via the write port 324 of the transmit FIFO RAM 322, enabling the transmit FIFO block 320 to process the next incoming packet without waiting for the transmit block 328 to become available. The transmit FIFO RAM 322 is a two-port memory having a write port 324 that accepts incoming data from the transmit FIFO block 320 and a read port 326 that transmits data stored in the transmit FIFO RAM 322 to the transmit block 328. Similar to the two-port memory comprising the receive FIFO RAM 308, the write port 324 and the read port 326 of the transmit FIFO RAM 322 operate independently of each other, thus permitting pipelining and more efficient use of the transmit FIFO RAM 322 by the transmit FIFO block 320 and the transmit block 328.
  • The transmit block 328 reads data stored in the transmit FIFO RAM 322 via the read port 326. Similar to the receive block 304, the transmit block 328 performs a variety of tasks and is very flexible in operation. The transmit block 328 internally converts data received from TXPP 120 into a format suitable for transmission to the PHY transceiver 300. According to one embodiment of the invention, the transmit block converts 64 bit SDR data into 32 bit DDR data for transmission. The transmit FIFO RAM 322 facilitates this conversion by bridging the asynchronous boundary between the transmit block 328 and the transmit FIFO block 320.
  • The transmit block performs other tasks as required according to embodiments of the invention, such as generating flow control packets to the PHY side sender at the request of the TXPP 120 (and in addition to internal flow control requests generated by the receive block 304 via the pause and flow control sync 332 when the receive FIFO RAM 308 is full) to avoid network congestion, calculating and optionally appending a cyclic redundancy check to a packet, determining and inserting XGMII tokens, and tracking the full suite of RMON statistics. In one embodiment of the invention, the transmit block 328 stores data in a programmable FIFO buffer used for data rate matching which allows the MAC 102 to connect to a packet processor that is receiving data slower than line rate.
  • The transmit block 328 passes data processed for to the transmit front end 330 thus enabling the transmit block 328 to begin processing the next packet. The transmit front end 330 is an interface that receives data from the transmit block 328 and passes this data to the PHY transceiver 300 for transmission over the network. According to one preferred embodiment of the invention, the transmit front end 330 transmits 32 bit DDR data to the PHY transceiver 300.
  • Building on the illustration presented in FIG. 1, FIG. 4 presents a block diagram depicting one embodiment of the components of the RXPP 104. The RXPP 402 is responsible for packet classification, performing CAM and PRAM lookups, generating hardware packet headers used for internally forwarding packets within the chassis of the network device, such as between blades, and for preparing packet modifications. Components of the RXPP 104 are embodied in the RXPP chip 402. According to one preferred embodiment of the invention, the RXPP chip 402 comprises an FPGA. In an alternate embodiment of the invention, the RXPP chip 402 comprises an ASIC.
  • The XGMAC 404 interface is responsible for requesting data for the RXPP 402 from the MAC 102. When the receive lookup handler 406 is available to parse additional data and the receive data FIFO 438 is available to store additional data, the XGMAC 404 instructs the MAC 102 to begin streaming packet data into the RXPP 104. The XGMAC interface 404 is connected to the MAC 102 and receives data for processing by the RXPP 104. The XGMAC interface 404 also acts as an asynchronous boundary, writing source-synchronous 64-bit data from the MAC 102 in a small internal FIFO, then sending the synchronized data at 66 MHz in 256-bit chunks for subsequent processing.
  • The XGMAC interface 404 sends synchronized data as it is received from the MAC 102 to the receive data FIFO 438, where it is held until CAM and PRAM lookups are performed. The receive data FIFO 438 thus acts as a delay buffer until packet processing is completed and the packet data can start being written by the dual-port interface 440 into the receive dual-port memory 110.
  • While all data related to a packet is streamed to the receive data FIFO 438, the XGMAC interface 404 also parses the incoming data as it is received from the MAC 102 and streams only the packet header information to the receive lookup handler 406 where it will be used to perform CAM and PRAM lookups.
  • The receive lookup handler 406 performs sanity checks on the packet data as it is received from the XGMAC interface 404. For example, the receive lookup handler 406 identifies valid packet contexts by identifying consistent start-of-packet and end-of-packet boundaries. In this respect, the receive lookup handler 406 also monitors a bad packet control signal from the MAC 102 indicating a data fault. If a data fault is detected, the receive lookup handler 406 discards the header data from the bad packet and also flushes any associated data already stored in the receive data FIFO 438 related to the bad packet. In one embodiment of the invention, if packet processing has already started, a data fault flag indicating a bad packet is stored in the receive data FIFO 438. The dual port interface 440 will later discard the packet when the data fault flag is retrieved from the receive data FIFO 438.
  • The receive lookup handler 406 strips VLAN tags, compares the packet MAC destination address against the port MAC address, performs IPv4 TOS field lookups as required, and also checks the protocol used to encode the packet. Examples of encoding protocols include IP, IP ARP, IPv4, IPv6, 802.3, IPX RAW, IPX LLC, IPX 8137, IPX SNAP, Appletalk, Appletalk ARP, NetBios, IP SNAP, and IP ARP SNAP. This information will be used to assemble an internal hardware packet header to be appended to the packet for use in forwarding the data internally throughout the chassis of the network switch. This additional information is passed from the receive lookup handler 406 to the RX scheduler FIFO 407. The RX scheduler FIFO 407 holds this information until the CAM and PRAM lookups are completed on the destination and source addresses extracted by the receive lookup handler 406 from the packet header.
  • Based upon the information extracted, the receive lookup handler 406 forms the CAM lookups and builds part of the hardware packet header for internally forwarding the packet through the chassis of the network device. The internal state of the receive lookup handler 406 containing this information is then split into two CAM lookup FIFOs 408 and 410, which are memory buffers that permit the receive lookup handler 406 to start processing the next packet received from the XGMAC interface 404. Packet processing is thus pipelined, allowing the receive lookup processor 406 to continue processing packets without waiting for either the CAM1 interface 412 or the CAM2 interface 410 to become available. Information relating to the destination address of the packet and other protocol fields from the header related to Layer 3 are passed to CAM1 lookup FIFO 408. Information relating to the source address of the packet and other protocol fields from the header related to Layer 4 are passed to CAM2 lookup FIFO 410. In an alternate embodiment of the invention, the two pipelines are merged into a single pipeline containing a single CAM interface and a single FIFO interface for lookups.
  • The CAM1 interface 412 becomes available, retrieves the data stored in the CAM1 lookup FIFO 408, and submits requests regarding this data to the external ternary CAM1 414 memory bank that contains a data array of values against which to perform lookups. The CAM1 interface 412 is also pipelined and supports dispatching lookups for multiple packets to the external ternary CAM1 414 memory bank since it takes longer than four clocks for the external CAM1 414 to respond.
  • If the lookup generates a match against an entry in the CAM1 414 array, additional forwarding information exists in the PRAM1 426 memory bank regarding the disposition of the packet. Forwarding information might include details such as the destination port of the packet, the port mirror requirement, the packet type, VLAN handling information, packet prioritization data, multicast group membership, replacement destination MAC addresses (used in network routing), and/or other similar packet data known in the art. The CAM1 414 array entry also contains a link to the memory address of the additional forwarding information stored in the PRAM1 426 memory bank. This link is stored in the CAM1 result FIFO 420 until the PRAM1 interface 424 is available to perform lookups.
  • Similarly, the CAM2 interface 416 retrieves source address data from the CAM2 lookup FIFO 410, performs lookups by submitting requests to the external ternary CAM2 memory bank 418, and stores the results of these lookups in the CAM2 result FIFO 422 until the PRAM2 interface 428 is available to perform lookups. According to one embodiment of the invention, the CAM2 interface 416 operates in parallel with the CAM1 interface 412 to allow CAM lookup operations to complete faster.
  • The PRAM1 interface 424 retrieves the data associated with the successful CAM1 interface 412 lookups from the CAM1 result FIFO 420. The PRAM1 interface 424 extracts from this data the link to the memory address of the additional forwarding information stored in the PRAM1 426 memory bank. PRAM1 interface 424 lookup results are stored in the PRAM1 result FIFO so work can immediately start on the next packet. According to one embodiment, PRAM lookups for a packet take 3 clocks. Similarly, and preferably in parallel, the PRAM2 interface 428 retrieves data associated with successful CAM2 interface 416 source address lookups from the CAM2 result FIFO 422, performs lookups to obtain additional forwarding information stored in the PRAM2 430 memory bank, and stores the results in the PRAM2 result FIFO 434.
  • The receive packet evaluator 436 extracts the data from the PRAM1 result FIFO 432, PRAM2 result FIFO 434, and the RX scheduler FIFO 407. The receive packet evaluator 436 uses this information to construct the internal hardware header used to forward a packet through the chassis with the most advanced forwarding in this aspect permitting total destination address/VLAN/TOS replacement and packet header modification to support hardware packet routing. In one embodiment of the invention, the internal hardware header comprises sixteen bytes. The receive packet evaluator 436 also determines the priority level of the packet according to the CAM and PRAM lookups and may optionally adjust the packet priority according to whether the packet is VLAN tagged or contains IPv4 TOS fields. The priority level is inserted into the internal hardware header of the packet.
  • The receive packet evaluator 436 notifies the dual-port interface 440 that processing is complete and passes the new internal hardware header to the dual-port interface 440 for integration with the packet data stored in the receive data FIFO 438. The dual-port interface 440 reads from the receive data FIFO 438, applying packet modifications to incorporate the new hardware packet header and stores this packet data in the receive dual-port memory 110. The dual-port interface 440 also detects the end of packet (“EOP”) signal and issues a receive packet processing completion notification to the backplane manager 112 so the backplane manager 112 will know to retrieve the packet. If a packet is flagged as bad (for example, an invalid cyclic redundancy check) the buffer is instead immediately recycled for the next packet and the current packet is deleted.
  • FIG. 5 presents a block diagram depicting the operations of the RXPP 402 presented in FIG. 4 more discretely. Data flow commences with the receive lookup handler 501 receiving packet data from the XGMAC interface 404 as illustrated in FIG. 4. The XGMAC interface 404 parses data received from the MAC 102 and sends only the packet header information to the receive lookup handler 501.
  • The receive port tracker 502 examines the port information contained in the packet header to ensure that any VLAN information tags contained in the packet header will be accepted at the destination address port. If the destination address port is not configured to accept the packet header VLAN information or lack thereof, then the receive lookup handler 501 either sets an error bit in the packet header if debugging is supported or the packet is discarded. Alternatively, the receive lookup handler 501 will strip the VLAN tag from its field in the packet and store the VLAN tag in the internal hardware packet header for future use.
  • The receive lookup handler 501 checks the protocol used to encode the packet and classifies the packet accordingly in block 504. Examples of encoding protocols include IP, IP ARP, IPv4, IPv6, 802.3, IPX RAW, IPX LLC, IPX 8137, IPX SNAP, Appletalk, Appletalk ARP, NetBios, IP SNAP, and IP ARP SNAP. This information is used to assemble an internal hardware packet header to be appended to the packet for use in forwarding the data internally throughout the chassis of the switch. This additional information is passed from the receive lookup handler 501 to the RX scheduler FIFO 522. The RX scheduler FIFO 522 holds this information until the CAM and PRAM lookups are completed on the destination and source addresses extracted by the receive lookup handler 501 from the packet header.
  • The receive lookup handler 501 also forms the CAM lookups and builds part of the hardware packet header in block 506. The receive lookup handler 501 extracts source and destination address information from the packet header for use in the CAM lookups. The internal state of the receive lookup processor 501 containing this information is then passed to the CAM lookup FIFO 508, which is a memory buffer that permits the receive lookup processor 501 to start processing the next packet received from the XGMAC interface 404. Packet processing is thus pipelined allowing the receive lookup processor 501 to continue efficiently processing packets without waiting for the CAM interface 509 to become available.
  • When the CAM interface 509 becomes available, it fetches the address data stored in the CAM lookup FIFO 508 as shown in block 510. The CAM interface 509 dispatches requests regarding data in block 512 to the external ternary CAM memory 516 that contains a data array of values against which to perform lookups. The CAM interface 509 is pipelined and supports cycling lookups for multiple packets to the external ternary CAM 516 memory since it takes longer than four clocks for the external CAM 516 to respond. Block 514 illustrates a programmable delay incorporated into the CAM interface 509 pipeline that compensates for this delay while the CAM lookup is being performed.
  • If the lookup generates a match against an entry in the CAM array 516, additional forwarding information regarding disposition of the packet is available in the PRAM memory 530. Forwarding information might include details such as the destination port of the packet, the port mirror requirement, the packet type, VLAN handling information, packet prioritization data, multicast group membership, and/or other similar packet data known in the art. The CAM array 516 entry also contains a link to the memory address of the additional forwarding information stored in the PRAM memory 530. This link is returned by the CAM memory 516 as shown in block 518 and stored in the CAM result FIFO 520 until the PRAM interface 523 is available to perform lookups.
  • When the PRAM interface 523 becomes available, it fetches the link to the address in the PRAM memory 530 that is stored in the PRAM lookup FIFO 520 as shown in block 524. In block 526, the PRAM interface 523 dispatches requests to retrieve the additional forwarding information for the packet to the external PRAM memory 530. The PRAM interface 523 is pipelined and supports cycling lookups for multiple packets to the external PRAM memory 530 since it takes multiple clocks for the external PRAM memory 530 to return results from a lookup. Block 528 illustrates a programmable delay incorporated into the PRAM interface 523 pipeline that compensates for this delay while the PRAM lookup is being performed. The external PRAM 530 returns the additional forwarding information in block 532 and these results are stored in the PRAM result FIFO 534 until the receive packet evaluator 535 is available.
  • In block 536, the receive packet evaluator 535 fetches data from the PRAM result FIFO 534 and the receive scheduler FIFO 522. The receive packet evaluator 535 evaluates this information in block 538 and uses the results to construct the internal hardware packet header in block 540. The internal hardware packet header is used to forward the packet through the chassis among other blades inserted into slots on the backplane. The most advanced forwarding in this aspect permits total destination address/VLAN/TOS replacement and packet header modification to support hardware packet routing. In one embodiment of the invention, the internal hardware header comprises sixteen bytes. The receive packet evaluator 535 notifies the dual-port interface 542 that processing is complete and passes the new internal hardware header to the dual-port interface 542 for integration with the packet data stored in the receive data FIFO 438, as illustrated in FIG. 4. The dual-port interface 542 reads from the receive data FIFO 438 applying packet modifications to incorporate the new hardware packet header for internally forwarding the packet through the chassis of the switch and stores this packet data in the receive dual-port memory 110. The receive dual-port memory is organized as four large FIFOs corresponding to four exemplary priority levels. The dual-port interface 440 also detects the end of packet (“EOP”) and issues a receive packet processing completion notification to the backplane manager 112 so the backplane manager 112 will know to retrieve the packet. If a packet is flagged as bad (for example, an invalid cyclic redundancy check) the packet is deleted and the buffer is recycled for the next packet.
  • Transport within a blade continues with FIG. 6, which presents a block diagram depicting the components of the backplane manager 112 as illustrated in FIG. 1. Components of the backplane manager 602 are embodied in the backplane manager chip. According to a embodiment of the invention, the backplane manager chip 602 comprises an FPGA.
  • The backplane manager 602 is responsible for retrieving data from the receive dual-port memory 610, determining backplane destinations for this data, and sending this data to those destinations. The backplane manager 112 also manages four large FIFOs stored in the external dual-port memory 610. These FIFOs store data according to priority levels by which the data is to be processed by the backplane manager 112.
  • The receive done handler 604 receives EOP information from the receive packet processor 104, including information regarding packet length and packet priority. This information is used to assist the receive done handler 604 in tracking receive dual-port memory 110 utilization for the four priority levels and scheduling packets for dispatch by the transmit queue dispatch 606. If the backplane manager 602 or the receive dual-port memory FIFOs 610 are running low on resources, the receive done handler 604 sends a throttle control back to the receive packet processor 104.
  • The transmit queue dispatch 606 is responsible for ordered packet dispatch from the four priority levels of the receive dual-port memory FIFOs 610. The transmit queue dispatch 606 receives packet length and priority information from the receive done handler 606 and uses this information to schedule packet retrieval from the dual-port RAM 610 by the dual-port interface 608 according to prioritization algorithms contained in the transmit queue dispatch 606.
  • According to one embodiment of the invention, absolute priority is used with higher priority packets being unconditionally transmitted before any packets of lower priority. Absolute priority, however, is not always desirable. In another embodiment, some fraction of the transmission bandwidth available to the backplane manager 112 is dedicated to lower priority packet transmission regardless of whether higher priority packets are also pending because packets are often received by the invention faster than they can be transmitted. If some bandwidth were not allocated to lower priority packets in this manner, a bottleneck might be created with lower priority packets not being transmitted due to higher priority packets monopolizing all available transmission bandwidth. Packets are thus scheduled and posted for use by the transmit queue dispatch 606.
  • The dual-port interface 608 fetches data from the receive dual-port memory 610 based on instructions received by the transmit queue dispatch 606. At the start-of-packet boundary, the dual-port interface 608 extracts a forwarding identifier (“FID”) from the packet and sends the FID to the FID lookup interface 612. The FID is an abstract chassis/system wide number used to forward packets. Each packet type has a FID to instruct the blade how to handle a given type of packet. This allows each blade in the chassis to look at the FID separately to decide how to individually forward the packet.
  • The FID lookup interface 612 translates the FID received from the dual-port interface 608 into a port mask by performing a lookup against addresses stored in the external FID RAM 614. The port mask is a multi-bit field representing a port on the blade and also other possible backplane slot destinations in the device chassis. According to one embodiment, the port mask is an 8-bit field representing a 10 Gigabit Ethernet port on the blade and seven other possible backplane slot destinations.
  • The FID lookup takes a number of clock cycles to complete during which time read data is posted to the delay FIFO 616 by the dual-port interface 608. According to one embodiment of the invention, the FID lookup by the FID lookup interface 612 into the external FID RAM 614 requires a delay of six clocks to complete in order to resume processing the data.
  • The FID lookup is completes and the results are passed from the FID lookup interface 612 to the merge port mask 618. Read data stored in the delay FIFO 616 is also passed to the merge port mask 618. The merge port mask 618 integrates the read data with the appropriate FID lookup port mask result and other port masks as set forth below to ensure that the data is transmitted to all intended destinations.
  • The merge port mask 618 takes the FID lookup port mask result and combines it with CPU and monitor information stored in configuration registers of the backplane manager. For example, a FID indicates a physical destination or possibly a list of destinations, but the receive packet processor 104 might have determined that the CPU also needs a copy of the data and therefore sets the CPU flag for combination with the FID lookup port mask by the merge port mask 618. Alternatively, when a packet needs to be sent to a monitor port for network debugging or similar purpose, the monitor port mask is combined with the FID port mask. The merge port mask 618 thus generates a “qualified” port mask indicating all destinations for which the packet data is intended.
  • The merge port mask 618 may also apply source port suppression. In certain situations, the blade that receives the data packet is listed as part of a FID port mask; source port suppression conditionally prevents the blade from retransmitting packets it just received. For example, this might occur in a broadcast situation where packets with unknown addresses are sent to all ports. Once all port mask data is combined with packet data, the merge port mask 618 stores the final result in the receive data FIFO 620 enabling the merge port mask 618 to process the next packet without waiting for the backplane FIFO dispatch 624 to become available.
  • The backplane FIFO dispatch 624 reads data from the receive data FIFO 620, duplicating the data for each destination indicated in the qualified port mask. The backplane FIFO dispatch 624 restructures the data into a format required by the backplane, generates backplane state and slot information, and posts the results into the backplane data FIFO 626. The backplane data FIFO 626 also acts as an asynchronous boundary between the backplane manager 602 core clock and the actual backplane clock. By posting the results in the backplane data FIFO 626, the backplane FIFO dispatch 624 can process the next packet without waiting for the backplane dispatch 628 to become available. In one embodiment of the invention, data posted to the backplane data FIFO 626 is equivalent to two backplane transfers since the backplane manager runs at approximately one-half the clock speed of the backplane interface 114.
  • The backplane dispatch 628 reads data from the backplane data FIFO 626 and outputs the data to the backplane via the backplane interface 114. According to one embodiment, the backplane dispatch 628 reads data from the backplane data FIFO 626 suitable for more than one transfer because the ratio of the backplane interface 114 clock speed and the clock speed of the backplane manager 602 is not identical. In such an embodiment, the backplane dispatch 628 reads the number of transfers from the backplane data FIFO 626 that fully utilizes the transmission capacity of the backplane interface 114. For example, if the clock speed of the backplane interface 114 is double that of the backplane manager 602, then the backplane dispatch 628 will read two transfers from the backplane data FIFO.
  • The backplane dispatch 628 also monitors backplane status and directs backplane transmission rates since it is possible for a backplane slot destination to become congested or otherwise unavailable. For example, if a plurality of blades comprising a single chassis are devoting all of their transmission capacities to a single blade, then they may overload the destination blade. Such a case might occur when two blades both transmit at 8 Gbps to a single destination blade that, according to the capacity of a backplane slot, can only receive 8 Gbps it total. The two blades would have to throttle back transmissions to the destination blade to 4 Gbps to avoid congestion.
  • Data is received from the backplane by the transmission accumulator 116 as presented in FIG. 1. Turning to FIG. 7, the transmission accumulator 116 collects packets from the backplane and organizes them into per-source, per priority transmit FIFOs stored in the transmit dual-port memory 118. Components of the transmission accumulator are embodied in the transmission accumulator chip 702. According to one embodiment of the invention, the transmission accumulator chip 702 comprises an FPGA.
  • Data is received from the backplane by the backplane front end 704. The backplane front end passes received data to the backplane slot receive accumulator 706. The backplane slot receive accumulator 706 is divided into a series of equal storage structures or memory buffers, with one buffer allocated for each slot or source on the chassis of the device. According to one embodiment of the invention, the backplane slot receive accumulator 706 is divided into eight buffers for receipt of data.
  • When a particular quantity of data is received into one of the backplane slot receive accumulator 706 buffers, the backplane slot receive accumulator 706 notifies the backplane data polling logic 708 to indicate the buffer and priority of the data being stored. In one embodiment of the invention, the backplane slot receive accumulator 706 waits to notify the backplane data polling logic 708 until 32 bytes of data have been received in a bucket and transfers between the two components thus comprise 32 bytes. If the backplane slot receive accumulator 706 is full, then the transmission accumulator is congested and no longer accepts data until the congestion is relieved.
  • The backplane data polling logic 708 reads data from the backplane slot receive accumulator 706 and organizes data according to source and priority. If packets are aborted from the backplane, the backplane data polling logic 708 deletes the packet in order to avoid propagation of the packet to the TXPP 120.
  • The backplane data polling logic 708 processes the data and the final result is stored in the backplane receive FIFO 710, enabling the backplane data polling logic 708 to process the next packet without waiting for the dual-port interface 712 to become available. The backplane receive FIFO 710 also permits dataflow through the asynchronous boundary between the backplane data polling logic block 708 and the dual-port interface 712.
  • The dual-port interface 712 reads data from the backplane receive FIFO 710 and stores this packet data in the transmit dual-port memory 118. The dual-port interface 712 also detects valid end-of-packet (“EOP”) indications and notifies the TXPP 120 via transmission of an EOP message that a packet is available in the transmit dual-port memory 118. The transmit dual-port memory 118 also comprises a series of FIFOs similar to the receive dual-port memory 110. Instead of only four total FIFOs, however, the transmit dual-port memory 118 has four FIFOs for each buffer of the backplane slot accumulator 706, thereby comprising 28 FIFOs for these buffers, plus an additional four FIFOs for the OAR path, yielding a total of 32 FIFOs.
  • Transmission continues in FIG. 8, which depicts a block diagram of the components of the transmit packet processor 120 as illustrated in FIG. 1 a. Components of the TXPP 120 are embodied in the TXPP chip 800. According to an embodiment of the invention, the TXPP chip 800 comprises an FPGA. The TXPP 800 is responsible for retrieving data from the transmit dual-port memory 803, determining network destinations for this data and sending data to identified destinations. The TXPP 120 strips hardware header forwarding information used to route packets throughout the chassis of the switch and replaces this information with header forwarding information necessary to route packets over the network. The TXPP 120 also manages the FIFOs priority queues stored in the transmit dual-port memory 803. These FIFOs store data according to priority levels by which the data is to be processed by the TXPP 800.
  • The transmit done handler 801 receives EOP information from the TX accumulator 116, including information regarding packet length and packet priority. This information is used to assist the transmit done handler 801 in tracking transmit dual-port memory 803 utilization for the four priority levels and scheduling packets for dispatch in the transmit queue dispatch 802. The transmit done handler 801 notifies the transmit queue dispatch 802 regarding packet availability and priority.
  • The transmit queue dispatch 802 is responsible for ordered packet retrieval and dispatch from the four priority levels of the transmit dual-port memory 803 FIFOs. According to one embodiment of the invention, absolute priority is used with higher priority packets being unconditionally transmitted before any packets of lower priority. Absolute priority, however, is not always desirable. In alternative embodiments, some fraction of the transmission bandwidth available to the TXPP 120 is dedicated to lower priority packet transmission regardless, of whether higher priority packets are also pending because packets are often received by the invention faster than they can be transmitted. If some bandwidth were not allocated to lower priority packets in this manner, a bottleneck might be created with lower priority packets not being transmitted due to higher priority packets monopolizing all available transmission bandwidth. Packets are thus scheduled and posted for use by the dual-port handler 804.
  • The dual-port handler 804 fetches the data from the transmit dual-port memory 803 according to instructions received from the transmit queue dispatch 802. At the start-of-packet boundary, the dual-port handler 804 extracts the FID from the packet and sends the FID to the FID lookup block 808. The dual-port handler 804 also extracts any VLAN tags from the packet and sends this information to the multicast start offset lookup block 806.
  • In the FID lookup block 808, the FID received from the dual-port handler 804 is used to perform a lookup against a FID table. The FID lookup block 808 functions similarly to the interaction between the FID lookup interface 612 and the FID RAM 614 as presented in FIG. 6. Accordingly, the results obtained from the FID table indicate how the packet should be handled for transmission by the receiving blade. For example, the FID might indicate that although the packet may have arrived at the blade, the packet should not be transmitted by the blade. This might occur in a broadcast situation where a packet is broadcast to all blades within a chassis. If the FID lookup block 808 determines that a packet has been erroneously received in this manner, the packet is deleted and no longer processed by the TXPP 120. In this sense, the FID lookup block 808 also functions as a transmit filter to ensure that only valid packets are actually sent out over the network.
  • Results of the FID lookup are stored in the delay FIFO 810. This permits the FID lookup block 808 to begin processing the next packet without waiting for the context track and internal header removal block 814 to become available. Pipelining processing data in this manner allows packet processing operations by the TXPP 120 to complete faster.
  • While the FID lookup block 808 is processing the FID data, the multicast start offset lookup block 806 is processing any VLAN tags received from the dual-port handler 804. A VLAN is a local area network identifier that maps locations based on a basis other than physical location. For example, devices attached to a VLAN might be grouped according to department, division, application, etc. Devices that are part of the same VLAN behave as if they were connected to the same wire even though they may actually be physically connected to different segments of a LAN. VLANs are configured using software protocols rather than in hardware and are therefore extremely flexible with respect to implementation. For example, a computer may be moved to a different physical location on the same VLAN without any hardware reconfiguration.
  • VLAN tags placed in a header field indicate whether a packet is intended for routing over a VLAN. Additionally, the VLAN tag in the header may also indicate that a packet is intended for VLAN multicasting. VLAN multicasting occurs when a packet is sent over a VLAN to more than one destination address. Since the header of each packet must be changed to reflect each destination address during VLAN multicasting, this process can be very resource intensive when performed using software.
  • The multicast start offset lookup block 806 supports hardware VLAN multicast replication. The multicast start offset lookup block 806 examines the VLAN tag extracted from the packet header and performs a lookup against a table stored in RAM in the multicast start offset lookup block 806. If the packet VLAN tag matches an entry in the table, additional information pertaining to that VLAN is available at an address location in a memory array stored in the multicast replacement lookup block 812. For example, multicast replacement lookup block 812 might contain information to assist with setting unique VLAN ID values, VLAN priorities, and TXA/SAS/srcport suppression behaviors for each packet transmitted over the VLAN.
  • The multicast start offset lookup block 806 takes the address to the memory array location of the multicast replacement lookup block 812 and stores this result in the delay FIFO 810. This permits the multicast start offset lookup block 806 to begin processing the next packet without waiting for the context track and internal header removal block 814 to become available. Pipelining processing in this manner allows packet processing operations by the TXPP 120 to complete faster.
  • In addition to enabling pipelining, the delay FIFO 810 also stores values from the FID lookup block 808 and the multicast start offset lookup block 806 for retrieval by the multicast replacement lookup block 812 and the context track and internal header removal block 814. The multicast replacement lookup block 812 retrieves the results of the multicast start offset lookup block 806 calculations from the delay FIFO 810 for processing packets subject to VLAN routing. The multicast replacement lookup block 812 takes the address of the memory array location contained in the multicast replacement lookup block 812 and retrieves the additional information that is stored at that location pertaining to routing over the VLAN tag referenced in the packet header. This information is passed to the context track and internal header removal block 814 for incorporation into the outgoing packet header.
  • Taking the results from the delay FIFO 810 and the multicast replacement lookup block 812, the context track and internal header removal block 814 removes the internal hardware header from the packet and begins the process of assembling an outgoing packet header suitable for transmission over the network. Those skilled in the art will recognize that a number of manipulations to the outgoing packet header must take place before this can occur. The context track and internal header removal block 814 passes information regarding any data offset to the header which may have occurred to the barrel shifter 816. The context track and internal header removal block 814 passes information regarding the TXA/PTYPE to the SA substitution and L3 assist block 818. The context track and internal header removal block 814 passes information regarding the packet VLAN ID and the VLAN tag status to the VLAN insertion block.
  • The barrel shifter 816 normalizes any changes to the packet header that occurred during internal routing through the chassis. One function of the internal hardware header of a packet is to permit the CPU to add an encapsulation to a packet. Encapsulation is used by the CPU to complete operations more efficiently by avoiding having to copy the entire packet into CPU memory and then writing the packet back to the buffer pool. Instead, the CPU performs a small modification to the packet header. For example, this might occur when the CPU determines that a packet must be forwarded, but that the CPU must first add data to the header before forwarding can take place. Alternatively, the CPU might also remove data from the header temporarily to assist with forwarding.
  • During this process, the CPU might move data within the packet header into a non-standard format. For example, the destination address might appear at the wrong location within the packet for transmission over the network. The barrel shifter 816 analyzes the composition of the packet header and shifts the data within the header to normalize it and correct for any CPU modifications that might have occurred. When the barrel shifter 816 completes operations on the packet header, the packet header data is then in a standard format and is passed to the SA substitution and L3 assist block 818 for further processing.
  • The SA substitution and L3 assist block 818 performs further modifications on the packet header to prepare the packet for transmission over the network. The SA substitution and L3 assist block 818 replaces the MAC address that is required for routing packets. In an Ethernet environment, each packet header contains a destination address and a source address. The source address must be changed on transmit to reflect which port the packet is being broadcast from. The SA substitution and L3 assist block 818 also modifies other Layer 3 header fields as required, such as changing the IPv4/IPX time to live value or the checksum.
  • The packet is passed to the VLAN insertion block 820 for further processing. VLAN tags that were removed on receipt anywhere in the chassis are stored in the internal hardware header for future use on transmission. The VLAN insertion block 820 takes the internal hardware header information that is passed from the context track and internal header removal block 814 and reintroduces this information into the outgoing packet header as appropriate. This information includes the packet VLAN ID and the Tag Status.
  • When the outgoing header packet is reassembled for transmission over the network, the packet is stored in the TX FIFO 822 prior to being passed to the XGMAC interface 824. The TX FIFO 822 enables the VLAN insertion block 820 to begin processing the next packet without having to wait for the XGMAC interface to become available and enables faster operation by the VLAN insertion block 820.
  • Additionally, the TX FIFO 822 permits data flow though asynchronous boundaries. In some embodiments of the invention, the TXPP 120 operates at a different speed than the MAC 102. Data flow must be synchronized between asynchronous components so the TX FIFO 822 acts as a bridge between these components. For example, in the Foundry BigIron switch, the MAC 102 operates at a 156.25 MHz clock and the TXPP operates at only a 66 MHz clock.
  • While the invention has been described and illustrated in connection with preferred embodiments, many variations and modifications as will be evident to those skilled in this art may be made without departing from the spirit and scope of the invention, and the invention is thus not to be limited to the precise details of methodology or construction set forth above as such variations and modification are intended to be included within the scope of the invention.

Claims (20)

1. A system to provide data transmission speeds at or in excess of 10 gigabits per second between one or more source devices and one or more destination devices, the system comprising:
a media access control (MAC) interface to facilitate receipt and transmission of packets over a physical interface;
a first processor coupled to the MAC interface to receive packets from the MAC interface and to perform initial processing of packets, and further to dispatch packets to a first memory;
a second processor to retrieve packets from the first memory and to compute an appropriate destination, and further to dispatch packets to a backplane;
a third processor to receive packets from the backplane and to organize the packets for transmission, and further to dispatch packets to a second memory; and
a fourth processor coupled to the MAC interface and to retrieve packets from the second memory and to schedule the transmission of packets to the MAC interface for transmission to one or more destination devices.
2. The system of claim 1 wherein the first and second memories are dualport memories.
3. The system of claim 1, further comprising a plurality of FIFO structures within each of said processors to provide for pipelined processing of data.
4. The system of claim 3 wherein said processors are configured to process a packet in four clocks.
5. The system of claim 3 wherein the first processor comprises a receive packet processor to transfer data in at least 64 bit groupings.
6. The system of claim 5 wherein the receive packet processor is configured to perform CAM and PRAM lookups for received packets.
7. The system of claim 6, further comprising FIFO structures coupled between CAM and PRAM lookups to accommodate pipelined processing.
8. The system of claim 7 wherein the receive packet processor is configured to perform multiple CAM and PRAM lookups in parallel.
9. The system of claim 8 wherein at least one of the multiple CAM lookups is a source address CAM lookup and at least one other of the multiple CAM lookups is a destination address CAM lookup.
10. The system of claim 8 wherein at least one of the multiple PRAM lookups is a source address PRAM lookup and at least one other one of the multiple PRAM lookups is a destination address PRAM lookup.
11. The system of claim 1 wherein the second processor comprises a backplane manager to issue packets on the backplane.
12. The system of claim 11 wherein the backplane manager is configured to compute a packet priority.
13. The system of claim 12 wherein the packet priority is computed based on absolute priority.
14. The system of claim 12, further comprising means for providing low priority packets a guaranteed fraction of bandwidth.
15. The system of claim 1 wherein the third processor comprises a transmission accumulator to organize packets into a plurality of per-source and per-priority FIFOs.
16. The system of claim 15 wherein the FIFOs are contained in the second memory.
17. The system of claim 15 wherein the transmission accumulator is configured to delete aborted packets.
18. The system of claim 17 wherein the backplane manager is configured to issue an end-of-packet message to the fourth processor to indicate that a complete packet is available for retrieval.
19. The system of claim 1 wherein the fourth processor is a transmission packet processor to implement a priority scheduler.
20. The system of claim 19 wherein the priority scheduler is configured to determine an order in which packets are transmitted from the transmission packet processor to the MAC interface, based on their priority.
US12/702,031 2002-05-06 2010-02-08 System architecture for very fast ethernet blade Abandoned US20100246588A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US12/702,031 US20100246588A1 (en) 2002-05-06 2010-02-08 System architecture for very fast ethernet blade

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US10/139,831 US7266117B1 (en) 2002-05-06 2002-05-06 System architecture for very fast ethernet blade
US11/828,246 US7738450B1 (en) 2002-05-06 2007-07-25 System architecture for very fast ethernet blade
US12/702,031 US20100246588A1 (en) 2002-05-06 2010-02-08 System architecture for very fast ethernet blade

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
US11/828,246 Continuation US7738450B1 (en) 2002-05-06 2007-07-25 System architecture for very fast ethernet blade

Publications (1)

Publication Number Publication Date
US20100246588A1 true US20100246588A1 (en) 2010-09-30

Family

ID=38456958

Family Applications (3)

Application Number Title Priority Date Filing Date
US10/139,831 Active 2024-05-09 US7266117B1 (en) 2002-05-06 2002-05-06 System architecture for very fast ethernet blade
US11/828,246 Expired - Fee Related US7738450B1 (en) 2002-05-06 2007-07-25 System architecture for very fast ethernet blade
US12/702,031 Abandoned US20100246588A1 (en) 2002-05-06 2010-02-08 System architecture for very fast ethernet blade

Family Applications Before (2)

Application Number Title Priority Date Filing Date
US10/139,831 Active 2024-05-09 US7266117B1 (en) 2002-05-06 2002-05-06 System architecture for very fast ethernet blade
US11/828,246 Expired - Fee Related US7738450B1 (en) 2002-05-06 2007-07-25 System architecture for very fast ethernet blade

Country Status (1)

Country Link
US (3) US7266117B1 (en)

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100161894A1 (en) * 2004-10-29 2010-06-24 Foundry Networks, Inc. Double density content addressable memory (cam) lookup scheme
US7948872B2 (en) 2000-11-17 2011-05-24 Foundry Networks, Llc Backplane interface adapter with error control and redundant fabric
US7978702B2 (en) 2000-11-17 2011-07-12 Foundry Networks, Llc Backplane interface adapter
US7978614B2 (en) 2007-01-11 2011-07-12 Foundry Network, LLC Techniques for detecting non-receipt of fault detection protocol packets
US8037399B2 (en) 2007-07-18 2011-10-11 Foundry Networks, Llc Techniques for segmented CRC design in high speed networks
US8090901B2 (en) 2009-05-14 2012-01-03 Brocade Communications Systems, Inc. TCAM management approach that minimize movements
US8149839B1 (en) 2007-09-26 2012-04-03 Foundry Networks, Llc Selection of trunk ports and paths using rotation
US8170044B2 (en) 2002-05-06 2012-05-01 Foundry Networks, Llc Pipeline method and system for switching packets
US8194666B2 (en) 2002-05-06 2012-06-05 Foundry Networks, Llc Flexible method for processing data packets in a network routing system for enhanced efficiency and monitoring capability
US8238255B2 (en) 2006-11-22 2012-08-07 Foundry Networks, Llc Recovering from failures without impact on data traffic in a shared bus architecture
US8448162B2 (en) 2005-12-28 2013-05-21 Foundry Networks, Llc Hitless software upgrades
US8493988B2 (en) 2004-03-26 2013-07-23 Foundry Networks, Llc Method and apparatus for aggregating input data streams
US8599850B2 (en) 2009-09-21 2013-12-03 Brocade Communications Systems, Inc. Provisioning single or multistage networks using ethernet service instances (ESIs)
US8671219B2 (en) 2002-05-06 2014-03-11 Foundry Networks, Llc Method and apparatus for efficiently processing data packets in a computer network
US8718051B2 (en) 2003-05-15 2014-05-06 Foundry Networks, Llc System and method for high speed packet transmission
US8730961B1 (en) 2004-04-26 2014-05-20 Foundry Networks, Llc System and method for optimizing router lookup
US9946466B2 (en) 2014-08-19 2018-04-17 Samsung Electronics Co., Ltd. Mobile electronic device including embedded memory

Families Citing this family (34)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7382736B2 (en) 1999-01-12 2008-06-03 Mcdata Corporation Method for scoring queued frames for selective transmission through a switch
US7454500B1 (en) 2000-09-26 2008-11-18 Foundry Networks, Inc. Global server load balancing
US7002980B1 (en) 2000-12-19 2006-02-21 Chiaro Networks, Ltd. System and method for router queue and congestion management
US7649885B1 (en) 2002-05-06 2010-01-19 Foundry Networks, Inc. Network routing system for enhanced efficiency and monitoring capability
US7266117B1 (en) * 2002-05-06 2007-09-04 Foundry Networks, Inc. System architecture for very fast ethernet blade
US7688823B2 (en) * 2002-06-04 2010-03-30 Alcatel-Lucent Usa Inc. Efficient mechanism for wire-tapping network traffic
US7289498B2 (en) * 2002-06-04 2007-10-30 Lucent Technologies Inc. Classifying and distributing traffic at a network node
US7552275B1 (en) * 2006-04-03 2009-06-23 Extreme Networks, Inc. Method of performing table lookup operation with table index that exceeds CAM key size
US7606968B2 (en) * 2006-05-08 2009-10-20 Mcdata Corporation Multi-level content addressable memory
US7903654B2 (en) 2006-08-22 2011-03-08 Foundry Networks, Llc System and method for ECMP load sharing
US8615008B2 (en) 2007-07-11 2013-12-24 Foundry Networks Llc Duplicating network traffic through transparent VLAN flooding
US8271859B2 (en) 2007-07-18 2012-09-18 Foundry Networks Llc Segmented CRC design in high speed networks
US8248928B1 (en) 2007-10-09 2012-08-21 Foundry Networks, Llc Monitoring server load balancing
US8190881B2 (en) 2007-10-15 2012-05-29 Foundry Networks Llc Scalable distributed web-based authentication
KR101408390B1 (en) * 2008-04-28 2014-06-17 휴렛-팩커드 디벨롭먼트 컴퍼니, 엘.피. Adjustable server-transmission rates over fixed-speed backplane connections within a multi-server enclosure
US7995621B2 (en) * 2008-10-01 2011-08-09 Nortel Netwoeks Limited Techniques for time transfer via signal encoding
US8780911B2 (en) * 2009-10-08 2014-07-15 Force10 Networks, Inc. Link aggregation based on port and protocol combination
CN101692209B (en) * 2009-11-09 2011-11-30 盛科网络(苏州)有限公司 Circuit design method and device for simulating TCAM by using embedded SRAM of FPGA
US9137166B2 (en) * 2010-01-28 2015-09-15 Brocade Communications Systems, Inc. In-order traffic aggregation with reduced buffer usage
US20130016725A1 (en) * 2010-12-24 2013-01-17 Telefonaktiebolaget L M Ericsson (Publ) Method and system for intra-node header compression
US9258273B2 (en) * 2013-10-03 2016-02-09 Mcafee, Inc. Duplicating packets efficiently within a network security appliance
US9565138B2 (en) 2013-12-20 2017-02-07 Brocade Communications Systems, Inc. Rule-based network traffic interception and distribution scheme
US9648542B2 (en) 2014-01-28 2017-05-09 Brocade Communications Systems, Inc. Session-based packet routing for facilitating analytics
US9692715B2 (en) 2014-02-21 2017-06-27 Cavium, Inc. Multiple ethernet ports and port types using a shared data path
US10771475B2 (en) 2015-03-23 2020-09-08 Extreme Networks, Inc. Techniques for exchanging control and configuration information in a network visibility system
US10911353B2 (en) 2015-06-17 2021-02-02 Extreme Networks, Inc. Architecture for a network visibility system
US9866478B2 (en) 2015-03-23 2018-01-09 Extreme Networks, Inc. Techniques for user-defined tagging of traffic in a network visibility system
US10129088B2 (en) 2015-06-17 2018-11-13 Extreme Networks, Inc. Configuration of rules in a network visibility system
US10057126B2 (en) 2015-06-17 2018-08-21 Extreme Networks, Inc. Configuration of a network visibility system
US10530688B2 (en) 2015-06-17 2020-01-07 Extreme Networks, Inc. Configuration of load-sharing components of a network visibility router in a network visibility system
US10091075B2 (en) 2016-02-12 2018-10-02 Extreme Networks, Inc. Traffic deduplication in a visibility network
US10999200B2 (en) 2016-03-24 2021-05-04 Extreme Networks, Inc. Offline, intelligent load balancing of SCTP traffic
US10567259B2 (en) 2016-10-19 2020-02-18 Extreme Networks, Inc. Smart filter generator
US11212257B2 (en) 2018-06-22 2021-12-28 Aeronix, Inc. Multi-level secure ethernet switch

Citations (96)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3866175A (en) * 1974-04-24 1975-02-11 Ncr Co Data communication system between a central computer and a plurality of data terminals
US4723243A (en) * 1985-12-02 1988-02-02 Advanced Micro Devices, Inc. CRC calculation machine with variable bit boundary
US4807280A (en) * 1987-09-18 1989-02-21 Pacific Bell Cross-connect switch
US4896277A (en) * 1988-05-03 1990-01-23 Thermo King Corporation Method of mapping refrigerated containers in a power line carrier based monitoring system
US4985889A (en) * 1988-02-04 1991-01-15 Sprint International Communications Corporation Data packet switching
US5101404A (en) * 1988-08-26 1992-03-31 Hitachi, Ltd. Signalling apparatus for use in an ATM switching system
US5195181A (en) * 1992-01-10 1993-03-16 Digital Equipment Corporation Message processing system having separate message receiving and transmitting processors with message processing being distributed between the separate processors
US5280582A (en) * 1989-08-29 1994-01-18 Digital Equipment Corporation No-owner frame and multiple token removal mechanism for token ring networks
US5282196A (en) * 1991-10-15 1994-01-25 Hughes Aircraft Company Bursted and non-bursted data router
US5287477A (en) * 1991-08-07 1994-02-15 Hewlett-Packard Company Memory-resource-driven arbitration
US5299190A (en) * 1992-12-18 1994-03-29 International Business Machines Corporation Two-dimensional round-robin scheduling mechanism for switches with multiple input queues
US5299195A (en) * 1991-01-18 1994-03-29 National Semiconductor Corporation Repeater interface controller with multiple port node interfaces
US5390173A (en) * 1992-10-22 1995-02-14 Digital Equipment Corporation Packet format in hub for packet data communications system
US5392279A (en) * 1992-01-17 1995-02-21 Fujitsu Limited Interconnection network with self-routing function
US5490258A (en) * 1991-07-29 1996-02-06 Fenner; Peter R. Associative memory for very large key spaces
US5598410A (en) * 1994-12-29 1997-01-28 Storage Technology Corporation Method and apparatus for accelerated packet processing
US5600795A (en) * 1993-08-28 1997-02-04 U.S. Philips Corporation Local network operating in asynchronous transfer mode (ATM) generating control cell containing information about the user, address of the station, and user-related identification
US5721819A (en) * 1995-05-05 1998-02-24 Silicon Graphics Corporation Programmable, distributed network routing
US5732080A (en) * 1995-07-12 1998-03-24 Bay Networks, Inc. Method and apparatus for controlling data flow within a switching device
US5734826A (en) * 1991-03-29 1998-03-31 International Business Machines Corporation Variable cyclic redundancy coding method and apparatus for use in a multistage network
US5862350A (en) * 1994-12-22 1999-01-19 Intel Corporation Method and mechanism for maintaining integrity within SCSI bus with hot insertion
US5864555A (en) * 1996-07-05 1999-01-26 General Datacomm, Inc. Method and apparatus for generating a proxy connection endpoint for operation administration and management (OAM) asynchronous transfer mode (ATM) cells
US5867675A (en) * 1996-08-06 1999-02-02 Compaq Computer Corp Apparatus and method for combining data streams with programmable wait states
US5870538A (en) * 1995-07-19 1999-02-09 Fujitsu Network Communications, Inc. Switch fabric controller comparator system and method
US5872783A (en) * 1996-07-24 1999-02-16 Cisco Systems, Inc. Arrangement for rendering forwarding decisions for packets transferred among network switches
US5875200A (en) * 1990-11-08 1999-02-23 Cirrus Logic, Inc. Reed-Solomon code system employing k-bit serial techniques for encoding and burst error trapping
US6016310A (en) * 1997-06-30 2000-01-18 Sun Microsystems, Inc. Trunking support in a high performance network device
US6023471A (en) * 1997-10-07 2000-02-08 Extreme Networks Network interconnect device and protocol for communicating data among packet forwarding devices
US6031843A (en) * 1996-11-21 2000-02-29 Alcatel Data Networks Inc. Digital communications switching fabric
US6035414A (en) * 1996-11-08 2000-03-07 Hitachi, Ltd. Reliability of crossbar switches in an information processing system
US6038288A (en) * 1997-12-31 2000-03-14 Thomas; Gene Gilles System and method for maintenance arbitration at a switching node
US6172990B1 (en) * 1997-06-19 2001-01-09 Xaqti Corporation Media access control micro-RISC stream processor and method for implementing the same
US6178520B1 (en) * 1997-07-31 2001-01-23 Lsi Logic Corporation Software recognition of drive removal or insertion in a storage system
US6181699B1 (en) * 1998-07-01 2001-01-30 National Semiconductor Corporation Apparatus and method of assigning VLAN tags
US6185222B1 (en) * 1995-09-28 2001-02-06 Cisco Technology, Inc. Asymmetric switch architecture for use in a network switch node
US6195335B1 (en) * 1997-06-27 2001-02-27 International Business Machines Corporation Data switch
US6201492B1 (en) * 1999-05-28 2001-03-13 Cirrus Logic, Inc. Techniques for converting a plurality of channels continuously in an A/D converter
US6335935B2 (en) * 1998-07-08 2002-01-01 Broadcom Corporation Network switching architecture with fast filtering processor
US20020001307A1 (en) * 2000-05-20 2002-01-03 Equipe Communications Corporation VPI/VCI availability index
US6343072B1 (en) * 1997-10-01 2002-01-29 Cisco Technology, Inc. Single-chip architecture for shared-memory router
US20020012585A1 (en) * 2000-06-09 2002-01-31 Broadcom Corporation Gigabit switch with fast filtering processor
US6351143B1 (en) * 2000-05-01 2002-02-26 Xilinx, Inc. Content-addressable memory implemented using programmable logic
US6356942B1 (en) * 1998-05-13 2002-03-12 Axis Ab Integrated circuit and method for bringing an integrated circuit to execute instructions
US6356550B1 (en) * 1999-07-30 2002-03-12 Mayan Networks Corporation Flexible time division multiplexed bus using sonet formatting
US6363077B1 (en) * 1998-02-13 2002-03-26 Broadcom Corporation Load balancing in link aggregation and trunking
US6505281B1 (en) * 1998-06-02 2003-01-07 Raymond C. Sherry Hard disk drives employing high speed distribution bus
US20030009466A1 (en) * 2001-06-21 2003-01-09 Ta John D. C. Search engine with pipeline structure
US20030012198A1 (en) * 2001-07-11 2003-01-16 Nec Corporation Packet processing unit
US6510138B1 (en) * 1999-02-25 2003-01-21 Fairchild Semiconductor Corporation Network switch with head of line input buffer queue clearing
US20030033435A1 (en) * 2001-07-17 2003-02-13 Hanner Brian D. System and method for multidimensional data compression
US6522656B1 (en) * 1994-09-12 2003-02-18 3Com Corporation Distributed processing ethernet switch with adaptive cut-through switching
US20030043848A1 (en) * 2001-08-30 2003-03-06 Sonksen Bradley Stephen Method and apparatus for data item processing control
US20030043800A1 (en) * 2001-08-30 2003-03-06 Sonksen Bradley Stephen Dynamic data item processing
US6532229B1 (en) * 1999-01-15 2003-03-11 Brian Weatherred Johnson Low cost link aggregation method and system
US20030048785A1 (en) * 2001-08-28 2003-03-13 International Business Machines Corporation Network processor with single interface supporting tree search engine and CAM
US20030061459A1 (en) * 2001-09-27 2003-03-27 Nagi Aboulenein Method and apparatus for memory access scheduling to reduce memory access latency
US6678248B1 (en) * 1997-08-29 2004-01-13 Extreme Networks Policy based quality of service
US6681332B1 (en) * 2000-03-13 2004-01-20 Analog Devices, Inc. System and method to place a device in power down modes/states and restore back to first mode/state within user-controlled time window
US6683872B1 (en) * 1998-07-31 2004-01-27 The University Of Tokyo Variable rate digital switching system
US6687217B1 (en) * 1999-05-04 2004-02-03 Carriercomm, Inc. Method of and system for one plus one protection for radio equipment
US6687247B1 (en) * 1999-10-27 2004-02-03 Cisco Technology, Inc. Architecture for high speed class of service enabled linecard
US20040022263A1 (en) * 2002-08-02 2004-02-05 Xiaodong Zhao Cross point switch with out-of-band parameter fine tuning
US6691202B2 (en) * 2000-12-22 2004-02-10 Lucent Technologies Inc. Ethernet cross point switch with reduced connections by using column control buses
US20040028060A1 (en) * 2002-08-08 2004-02-12 Samsung Electronics Co., Ltd. Link state synchronization method and apparatus on ad-hoc network, and data structure therefor
US6697359B1 (en) * 1999-07-02 2004-02-24 Ancor Communications, Inc. High performance switch fabric element and switch systems
US6696917B1 (en) * 2000-09-21 2004-02-24 Nortel Networks Limited Folded Clos architecture switching
US6697368B2 (en) * 2000-11-17 2004-02-24 Foundry Networks, Inc. High-performance network switch
US6839346B1 (en) * 1999-04-05 2005-01-04 Nec Corporation Packet switching apparatus with high speed routing function
US6842422B1 (en) * 1999-06-15 2005-01-11 Marconi Communications, Inc. Data striping based switching system
US20050010849A1 (en) * 2003-03-18 2005-01-13 Cisco Technology, Inc., A Corporation Of California Method and system for emulating a Fibre Channel link over a SONET/SDH path
US20050010630A1 (en) * 2003-05-13 2005-01-13 International Business Machines Corporation Method and apparatus for determining a remainder in a polynomial ring
US6854117B1 (en) * 2000-10-31 2005-02-08 Caspian Networks, Inc. Parallel network processor array
US6856600B1 (en) * 2000-01-04 2005-02-15 Cisco Technology, Inc. Method and apparatus for isolating faults in a switching matrix
US6859438B2 (en) * 1998-02-03 2005-02-22 Extreme Networks, Inc. Policy based quality of service
US20050041684A1 (en) * 1999-10-01 2005-02-24 Agilent Technologies, Inc. Multi-channel network monitoring apparatus, signal replicating device, and systems including such apparatus and devices, and enclosure for multi-processor equipment
US6982974B1 (en) * 1999-01-15 2006-01-03 Cisco Technology, Inc. Method and apparatus for a rearrangeably non-blocking switching matrix
US6990102B1 (en) * 2001-05-10 2006-01-24 Advanced Micro Devices, Inc. Parallel lookup tables for locating information in a packet switched network
US6993032B1 (en) * 2000-06-29 2006-01-31 Intel Corporation Buffer arrangements to support differential link distances at full bandwidth
US6996663B1 (en) * 2001-09-12 2006-02-07 Cisco Technology, Inc. Method and apparatus for performing address translation using a CAM
US20060031610A1 (en) * 2004-08-03 2006-02-09 Liav Ori R FIFO sub-system with in-line correction
US20060034452A1 (en) * 2002-06-20 2006-02-16 Hitachi, Ltd. Code calculating device
US7005812B2 (en) * 2001-07-23 2006-02-28 Lawrence Hardy Mitchell Commutation converter for a brushless servo motor
US7176911B1 (en) * 1999-02-02 2007-02-13 Olympus Corporation Solid-state imaging apparatus
US20070038798A1 (en) * 2004-09-10 2007-02-15 Bouchard Gregg A Selective replication of data structures
US7185141B1 (en) * 2001-12-27 2007-02-27 Netlogic Microsystems, Inc. Apparatus and method for associating information values with portions of a content addressable memory (CAM) device
US7185266B2 (en) * 2003-02-12 2007-02-27 Alacritech, Inc. Network interface device for error detection using partial CRCS of variable length message portions
US20080002707A1 (en) * 2002-05-06 2008-01-03 Davis Ian E Flexible method for processing data packets in a network routing system for enhanced efficiency and monitoring capability
US7324509B2 (en) * 2001-03-02 2008-01-29 Broadcom Corporation Efficient optimization algorithm in memory utilization for network applications
US20080031263A1 (en) * 2006-08-07 2008-02-07 Cisco Technology, Inc. Method and apparatus for load balancing over virtual network links
US20080037544A1 (en) * 2006-08-11 2008-02-14 Hiroki Yano Device and Method for Relaying Packets
US20080049742A1 (en) * 2006-08-22 2008-02-28 Deepak Bansal System and method for ecmp load sharing
US7649885B1 (en) * 2002-05-06 2010-01-19 Foundry Networks, Inc. Network routing system for enhanced efficiency and monitoring capability
US7657703B1 (en) * 2004-10-29 2010-02-02 Foundry Networks, Inc. Double density content addressable memory (CAM) lookup scheme
US20100034215A1 (en) * 2000-11-17 2010-02-11 Foundry Networks, Inc. Backplane Interface Adapter with Error Control
US20100046521A1 (en) * 2003-05-15 2010-02-25 Foundry Networks, Inc. System and Method for High Speed Packet Transmission
US20110002340A1 (en) * 2002-05-06 2011-01-06 Foundry Networks, Inc. Pipeline method and system for switching packets

Family Cites Families (216)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4683564A (en) 1982-11-22 1987-07-28 Data Switch Corporation Matrix switch system
DE3616141A1 (en) 1986-05-14 1987-11-19 Siemens Ag Contour collimator for radiation therapy (radiotherapy, X-ray therapy)
US4791629A (en) 1986-06-02 1988-12-13 Ibm Corporation Communications switching system
CA1292054C (en) 1987-05-15 1991-11-12 Hitachi, Ltd. Packet switching equipment and a packet switching method for controlling packet switched networks
US5136584A (en) 1990-07-11 1992-08-04 At&T Bell Laboratories Hardware interface to a high-speed multiplexed link
FR2666472B1 (en) 1990-08-31 1992-10-16 Alcatel Nv TEMPORARY INFORMATION STORAGE SYSTEM INCLUDING A BUFFER MEMORY RECORDING DATA IN FIXED OR VARIABLE LENGTH DATA BLOCKS.
JP2770936B2 (en) 1990-12-18 1998-07-02 インターナショナル・ビジネス・マシーンズ・コーポレイション Method of creating communication network and communication channel
US5224108A (en) 1991-01-23 1993-06-29 Mcdysan David E Method and apparatus for call control signaling
US5307345A (en) 1992-06-25 1994-04-26 Digital Equipment Corporation Method and apparatus for cut-through data packet transfer in a bridge device
GB2273224B (en) 1992-12-05 1997-01-22 Netcomm Ltd An ATM Cell switch suitable for multicast switching
ATE209414T1 (en) 1992-12-18 2001-12-15 Cit Alcatel ATM SWITCH AND ATM SWITCH ELEMENT WITH ROUTING LOGIC
US5323386A (en) 1993-02-01 1994-06-21 Trw Inc. Expandable high speed serial data switch
US5406643A (en) 1993-02-11 1995-04-11 Motorola, Inc. Method and apparatus for selecting between a plurality of communication paths
FR2703545B1 (en) 1993-03-31 1995-05-12 Alcatel Nv Asynchronous switching node dynamically distributing cells to outputs constituting a so-called irregular group.
US5365512A (en) 1993-07-02 1994-11-15 Ericsson Ge Mobile Communications Inc. Multisite trunked RF communication system with reliable control messaging network
US5408469A (en) 1993-07-22 1995-04-18 Synoptics Communications, Inc. Routing device utilizing an ATM switch as a multi-channel backplane in a communication network
DE4328862A1 (en) 1993-08-27 1995-03-02 Sel Alcatel Ag Method and device for buffering data packets and switching center with such a device
US5640504A (en) 1994-01-24 1997-06-17 Advanced Computer Applications, Inc. Distributed computing network
EP0748540A4 (en) 1994-03-03 1997-10-29 Proxim Inc Frequency hopping medium access control protocol
US6151301A (en) 1995-05-11 2000-11-21 Pmc-Sierra, Inc. ATM architecture and switching element
GB9408574D0 (en) 1994-04-29 1994-06-22 Newbridge Networks Corp Atm switching system
US5555244A (en) 1994-05-19 1996-09-10 Integrated Network Corporation Scalable multimedia network
US5606359A (en) 1994-06-30 1997-02-25 Hewlett-Packard Company Video on demand system with multiple data sources configured to provide vcr-like services
US5907660A (en) 1994-09-21 1999-05-25 Mitsubishi Denki Kabushiki Kaisha Digital video signal playback device with special playback data being in the form of a still image slice data
EP0716370A3 (en) 1994-12-06 2005-02-16 International Business Machines Corporation A disk access method for delivering multimedia and video information on demand over wide area networks
JPH08167907A (en) 1994-12-15 1996-06-25 Nec Corp Atm cell exchange
US5619497A (en) 1994-12-22 1997-04-08 Emc Corporation Method and apparatus for reordering frames
US5566170A (en) 1994-12-29 1996-10-15 Storage Technology Corporation Method and apparatus for accelerated packet forwarding
US5550816A (en) 1994-12-29 1996-08-27 Storage Technology Corporation Method and apparatus for virtual switching
US5546385A (en) 1995-01-19 1996-08-13 Intel Corporation Flexible switching hub for a communication network
US5666353A (en) 1995-03-21 1997-09-09 Cisco Systems, Inc. Frame based traffic policing for a digital switch
US5936939A (en) 1995-05-22 1999-08-10 Fore Systems, Inc. Digital network including early packet discard mechanism with adjustable threshold
US5646878A (en) 1995-06-02 1997-07-08 Motorola, Inc. Content addressable memory system
US5663952A (en) 1995-07-07 1997-09-02 Sun Microsystems, Inc. Checksum generation circuit and method
US6147996A (en) 1995-08-04 2000-11-14 Cisco Technology, Inc. Pipelined multiple issue packet switch
US6427185B1 (en) 1995-09-29 2002-07-30 Nortel Networks Limited Method and apparatus for managing the flow of data within a switching device
US5745708A (en) 1995-09-29 1998-04-28 Allen-Bradley Company, Inc. Method for and apparatus for operating a local communications module in arbitrating for mastership of a data transfer across a back plane bus in industrial automation controller
US5802065A (en) 1995-10-23 1998-09-01 Kawasaki Steel Corporation Data receiving device
US5838684A (en) 1996-02-22 1998-11-17 Fujitsu, Ltd. Low latency, high clock frequency plesioasynchronous packet-based crossbar switching chip system and method
GB2312353B (en) 1996-04-16 2000-12-06 Gpt Ltd Digital telecommunications transmision systems
US5835496A (en) 1996-04-30 1998-11-10 Mcdata Corporation Method and apparatus for data alignment
US6243667B1 (en) 1996-05-28 2001-06-05 Cisco Systems, Inc. Network flow switching and flow data export
US5751710A (en) 1996-06-11 1998-05-12 Cisco Technology, Inc. Technique for connecting cards of a distributed network switch
US6369855B1 (en) 1996-11-01 2002-04-09 Texas Instruments Incorporated Audio and video decoder circuit and system
FI103312B1 (en) 1996-11-06 1999-05-31 Nokia Telecommunications Oy Switching matrix
US6493347B2 (en) 1996-12-16 2002-12-10 Juniper Networks, Inc. Memory organization in a switching device
US6094434A (en) 1996-12-30 2000-07-25 Compaq Computer Corporation Network switch with separate cut-through buffer
US6076115A (en) 1997-02-11 2000-06-13 Xaqti Corporation Media access control receiver and network management system
US6222845B1 (en) 1997-02-25 2001-04-24 Cascade Communications Corp. System and method for providing unitary virtual circuit in digital network having communication links of diverse service types
US5920886A (en) 1997-03-14 1999-07-06 Music Semiconductor Corporation Accelerated hierarchical address filtering and translation using binary and ternary CAMs
US6000016A (en) 1997-05-02 1999-12-07 Intel Corporation Multiported bypass cache in a bypass network
US5907566A (en) 1997-05-29 1999-05-25 3Com Corporation Continuous byte-stream encoder/decoder using frequency increase and cyclic redundancy check
JP3116860B2 (en) 1997-06-05 2000-12-11 日本電気株式会社 Data processing device, data processing system, data processing method, information storage medium
US6118787A (en) 1997-06-27 2000-09-12 Advanced Micro Devices, Inc. Apparatus and method for regulating assigned bandwidth in high speed packet switched networks
US6081512A (en) 1997-06-30 2000-06-27 Sun Microsystems, Inc. Spanning tree support in a high performance network device
US6128666A (en) 1997-06-30 2000-10-03 Sun Microsystems, Inc. Distributed VLAN mechanism for packet field replacement in a multi-layered switched network element using a control field/signal for indicating modification of a packet with a database search engine
US6081522A (en) 1997-06-30 2000-06-27 Sun Microsystems, Inc. System and method for a multi-layer network element
US6088356A (en) 1997-06-30 2000-07-11 Sun Microsystems, Inc. System and method for a multi-layer network element
US5920566A (en) 1997-06-30 1999-07-06 Sun Microsystems, Inc. Routing in a multi-layer distributed network element
US5909686A (en) 1997-06-30 1999-06-01 Sun Microsystems, Inc. Hardware-assisted central processing unit access to a forwarding database
JP2978844B2 (en) 1997-07-01 1999-11-15 日本電気株式会社 Back pressure type ATM switch
US6304903B1 (en) 1997-08-01 2001-10-16 Agilent Technologies, Inc. State machine for collecting information on use of a packet network
US6108306A (en) 1997-08-08 2000-08-22 Advanced Micro Devices, Inc. Apparatus and method in a network switch for dynamically allocating bandwidth in ethernet workgroup switches
US6157641A (en) 1997-08-22 2000-12-05 Cisco Technology, Inc. Multiprotocol packet recognition and switching
US6333929B1 (en) 1997-08-29 2001-12-25 Intel Corporation Packet format for a distributed system
US6272144B1 (en) 1997-09-29 2001-08-07 Agere Systems Guardian Corp. In-band device configuration protocol for ATM transmission convergence devices
US6226680B1 (en) 1997-10-14 2001-05-01 Alacritech, Inc. Intelligent network interface system method for protocol processing
US6480477B1 (en) 1997-10-14 2002-11-12 Innowave Eci Wireless Systems Ltd. Method and apparatus for a data transmission rate of multiples of 100 MBPS in a terminal for a wireless metropolitan area network
US6434620B1 (en) 1998-08-27 2002-08-13 Alacritech, Inc. TCP/IP offload network interface device
US6320859B1 (en) 1997-10-31 2001-11-20 Nortel Networks Limited Early availability of forwarding control information
US6125417A (en) 1997-11-14 2000-09-26 International Business Machines Corporation Hot plug of adapters using optical switches
US6144668A (en) 1997-11-26 2000-11-07 International Business Machines Corporation Simultaneous cut through and store-and-forward frame support in a network device
EP0920152A3 (en) 1997-11-27 2002-09-18 Hitachi, Ltd. Method and apparatus for multiplex transmission
US6067606A (en) 1997-12-15 2000-05-23 Intel Corporation Computer processor with dynamic setting of latency values for memory access
US6160809A (en) 1997-12-17 2000-12-12 Compaq Computer Corporation Distributed packet data with centralized snooping and header processing router
US6549519B1 (en) 1998-01-23 2003-04-15 Alcatel Internetworking (Pe), Inc. Network switching device with pipelined search engines
EP0938212A1 (en) 1998-02-19 1999-08-25 International Business Machines Corporation Process and system of flow control for a switching system
US6496502B1 (en) 1998-06-29 2002-12-17 Nortel Networks Limited Distributed multi-link trunking method and apparatus
US6658002B1 (en) 1998-06-30 2003-12-02 Cisco Technology, Inc. Logical operation unit for packet processing
US7346063B1 (en) 1998-07-08 2008-03-18 Broadcom Corporation Memory management unit for a network switch
US6424659B2 (en) 1998-07-17 2002-07-23 Network Equipment Technologies, Inc. Multi-layer switching apparatus and method
EP0986226B1 (en) 1998-09-11 2007-01-17 Hitachi, Ltd. Ip packet communication apparatus
US6324651B2 (en) 1998-11-12 2001-11-27 International Business Machines Corporation Method and apparatus for saving device state while a computer system is in sleep mode
US6470436B1 (en) 1998-12-01 2002-10-22 Fast-Chip, Inc. Eliminating memory fragmentation and garbage collection from the process of managing dynamically allocated memory
US6934305B1 (en) 1999-01-15 2005-08-23 Cisco Technology, Inc. Method and apparatus for detecting errors in a backplane frame
US6424658B1 (en) 1999-01-29 2002-07-23 Neomagic Corp. Store-and-forward network switch using an embedded DRAM
US6556208B1 (en) 1999-03-23 2003-04-29 Intel Corporation Network management card for use in a system for screen image capturing
US6490280B1 (en) 1999-03-31 2002-12-03 Advanced Micro Devices, Inc. Frame assembly in dequeuing block
US6654346B1 (en) 1999-07-19 2003-11-25 Dunti Corporation Communication network across which packets of data are transmitted according to a priority scheme
US6813266B1 (en) 1999-05-21 2004-11-02 Advanced Micro Devices, Inc. Pipelined access to address table in a network switch
US6460088B1 (en) 1999-05-21 2002-10-01 Advanced Micro Devices, Inc. Method and apparatus for port vector determination at egress
JP3449294B2 (en) 1999-06-18 2003-09-22 日本電気株式会社 Multiprotocol processing device, line interface, and multiprotocol switch system having the same
WO2001015388A1 (en) 1999-08-20 2001-03-01 Fujitsu Limited Atm communication monitor
EP1208677B1 (en) 1999-09-03 2012-05-02 Broadcom Corporation Apparatus and method for enabling voice over ip support for a network switch
JP4421156B2 (en) 1999-10-29 2010-02-24 富士通株式会社 Image reproducing apparatus and image recording / reproducing apparatus
JP4234866B2 (en) 1999-10-29 2009-03-04 富士通株式会社 Optical node, optical node for transmission, optical node for reception, and failure recovery method for wavelength path network
JP3482996B2 (en) 1999-12-03 2004-01-06 日本電気株式会社 ATM switch
US6661791B1 (en) 1999-12-28 2003-12-09 Mosaid Technologies, Inc. Method and apparatus for generating forward overrides in a packet switch
JP3386117B2 (en) 2000-01-11 2003-03-17 日本電気株式会社 Multilayer class identification communication device and communication device
GB2358332B (en) 2000-01-14 2002-05-29 Marconi Comm Ltd Method of communicating data in a communication system
US6778546B1 (en) 2000-02-14 2004-08-17 Cisco Technology, Inc. High-speed hardware implementation of MDRR algorithm over a large number of queues
US6813243B1 (en) 2000-02-14 2004-11-02 Cisco Technology, Inc. High-speed hardware implementation of red congestion control algorithm
US6765866B1 (en) 2000-02-29 2004-07-20 Mosaid Technologies, Inc. Link aggregation
US6643269B1 (en) 2000-03-03 2003-11-04 Luminous Networks, Inc. Routing switch automatically identifying network topology
US6654342B1 (en) 2000-03-07 2003-11-25 Cisco Technology, Inc. Accumulating and distributing flow control information via update messages and piggybacked flow control information in other messages in a packet switching system
US6798740B1 (en) 2000-03-13 2004-09-28 Nortel Networks Limited Method and apparatus for switch core health monitoring and redundancy
US6700894B1 (en) 2000-03-15 2004-03-02 Broadcom Corporation Method and apparatus for shared buffer packet switching
US6751224B1 (en) 2000-03-30 2004-06-15 Azanda Network Devices, Inc. Integrated ATM/packet segmentation-and-reassembly engine for handling both packet and ATM input data and for outputting both ATM and packet data
US6798933B2 (en) 2000-04-14 2004-09-28 Shipley Company, L.L.C. Fiber optic array switch
US6807179B1 (en) 2000-04-18 2004-10-19 Advanced Micro Devices, Inc. Trunking arrangement in a network switch
US6574195B2 (en) 2000-04-19 2003-06-03 Caspian Networks, Inc. Micro-flow management
KR20000064177A (en) 2000-05-01 2000-11-06 권용순 mobile station
US7075926B2 (en) 2000-05-24 2006-07-11 Alcatel Internetworking, Inc. (Pe) Programmable packet processor with flow resolution logic
US6946948B2 (en) 2000-06-06 2005-09-20 Vitesse Semiconductor Corporation Crosspoint switch with switch matrix module
US6658591B1 (en) 2000-06-08 2003-12-02 International Business Machines Corporation Recovery from data fetch errors in hypervisor code
US6654373B1 (en) 2000-06-12 2003-11-25 Netrake Corporation Content aware network apparatus
US6654370B1 (en) 2000-06-30 2003-11-25 Marconi Communications, Inc. Backplane synchronization in a distributed system with clock drift and transport delay
US7103041B1 (en) 2000-06-30 2006-09-05 Marconi Intellectual Property (Ringfence), Inc. Optimization of number of transceivers used in a switch
US6463063B1 (en) 2000-06-30 2002-10-08 Marconi Communications, Inc. Receiver makes right
US6473433B1 (en) 2000-06-30 2002-10-29 Marconi Communications, Inc. Queue resynch: synchronous real-time upgrade of a distributed switching system
US6831932B1 (en) 2000-07-14 2004-12-14 Level 3 Communications, Inc. Transfer of SONET traffic over a packet-switched network
US6792484B1 (en) 2000-07-28 2004-09-14 Marconi Communications, Inc. Method and apparatus for storing data using a plurality of queues
US6721313B1 (en) 2000-08-01 2004-04-13 International Business Machines Corporation Switch fabric architecture using integrated serdes transceivers
US6804731B1 (en) 2000-08-11 2004-10-12 Paion Company, Limited System, method and article of manufacture for storing an incoming datagram in switch matrix in a switch fabric chipset system
US7283547B1 (en) 2000-08-11 2007-10-16 Ericsson Ab Switch fabrics logical synchronization utilizing propagation lockdown
US6865153B1 (en) 2000-09-20 2005-03-08 Alcatel Stage-implemented QoS shaping for data communication switch
US7171487B2 (en) 2000-09-22 2007-01-30 International Business Machines Corporation Method and system for application specific packet forwarding
US6633952B2 (en) 2000-10-03 2003-10-14 Broadcom Corporation Programmable refresh scheduler for embedded DRAMs
US6553370B1 (en) 2000-10-04 2003-04-22 Lsi Logic Corporation Flexible search engine having sorted binary search tree for perfect match
US6745277B1 (en) 2000-10-04 2004-06-01 Force10 Networks, Inc. Intelligent interleaving scheme for multibank memory
US6792502B1 (en) * 2000-10-12 2004-09-14 Freescale Semiconductor, Inc. Microprocessor having a content addressable memory (CAM) device as a functional unit therein and method of operation
US7080238B2 (en) 2000-11-07 2006-07-18 Alcatel Internetworking, (Pe), Inc. Non-blocking, multi-context pipelined processor
US6925516B2 (en) 2001-01-19 2005-08-02 Raze Technologies, Inc. System and method for providing an improved common control bus for use in on-line insertion of line replaceable units in wireless and wireline access systems
US20020089937A1 (en) 2000-11-16 2002-07-11 Srinivasan Venkatachary Packet matching method and system
US7356030B2 (en) 2000-11-17 2008-04-08 Foundry Networks, Inc. Network switch cross point
US7236490B2 (en) 2000-11-17 2007-06-26 Foundry Networks, Inc. Backplane interface adapter
JP2004537871A (en) 2000-11-17 2004-12-16 ファウンドリー ネットワークス インコーポレーティッド High performance network switch
US20020091884A1 (en) 2000-11-17 2002-07-11 Andrew Chang Method and system for translating data formats
US6735218B2 (en) 2000-11-17 2004-05-11 Foundry Networks, Inc. Method and system for encoding wide striped cells
US6629099B2 (en) 2000-12-07 2003-09-30 Integrated Silicon Solution, Inc. Paralleled content addressable memory search engine
US6654862B2 (en) 2000-12-29 2003-11-25 Ncr Corporation High performance disk mirroring
US20020126672A1 (en) 2001-01-10 2002-09-12 Nelson Chow Method and apparatus for a flexible and reconfigurable packet classifier using content addressable memory
US6732219B1 (en) 2001-02-23 2004-05-04 Hewlett-Packard Development Company, L.P. Dynamic allocation of devices to host controllers
US6959007B1 (en) 2001-03-01 2005-10-25 Lsi Logic Corporation High speed network protocol stack in silicon
US7193997B2 (en) 2001-03-19 2007-03-20 International Business Machines Corporation Packet classification
US6957258B2 (en) 2001-03-28 2005-10-18 Netrake Corporation Policy gateway
US6961340B2 (en) 2001-04-06 2005-11-01 Texas Instruments Incorporated AAL2 receiver for filtering signaling/management packets in an ATM system
US6937606B2 (en) * 2001-04-20 2005-08-30 International Business Machines Corporation Data structures for efficient processing of IP fragmentation and reassembly
US8218555B2 (en) 2001-04-24 2012-07-10 Nvidia Corporation Gigabit ethernet adapter
US6829682B2 (en) 2001-04-26 2004-12-07 International Business Machines Corporation Destructive read architecture for dynamic random access memories
US6804220B2 (en) 2001-05-07 2004-10-12 Qualcomm Incorporated Method and apparatus for generating control information for packet data
US7206283B2 (en) 2001-05-15 2007-04-17 Foundry Networks, Inc. High-performance network switch
US7239642B1 (en) 2001-07-16 2007-07-03 Network Appliance, Inc. Multi-protocol network interface card
US7218637B1 (en) 2001-07-20 2007-05-15 Yotta Networks, Llc System for switching data using dynamic scheduling
US20030035371A1 (en) 2001-07-31 2003-02-20 Coke Reed Means and apparatus for a scaleable congestion free switching system with intelligent control
US6868476B2 (en) 2001-08-27 2005-03-15 Intel Corporation Software controlled content addressable memory in a general purpose execution datapath
US6996117B2 (en) 2001-09-19 2006-02-07 Bay Microsystems, Inc. Vertical instruction and data processing in a network processor architecture
US7355970B2 (en) 2001-10-05 2008-04-08 Broadcom Corporation Method and apparatus for enabling access on a network switch
US20030074657A1 (en) 2001-10-12 2003-04-17 Bramley Richard A. Limited time evaluation system for firmware
US7210003B2 (en) 2001-10-31 2007-04-24 Netlogic Microsystems, Inc. Comparand generation in a content addressable memory
US7286534B2 (en) 2001-11-02 2007-10-23 Infineon Technologies North America Corporation SRAM based cache for DRAM routing table lookups
JP3627697B2 (en) 2001-11-16 2005-03-09 日本電気株式会社 Packet destination search system
US7236492B2 (en) 2001-11-21 2007-06-26 Alcatel-Lucent Canada Inc. Configurable packet processor
JP3898498B2 (en) 2001-12-06 2007-03-28 富士通株式会社 Server load balancing system
US6754881B2 (en) 2001-12-10 2004-06-22 International Business Machines Corporation Field programmable network processor and method for customizing a network processor
US8195705B2 (en) 2001-12-11 2012-06-05 International Business Machines Corporation Hybrid search memory for network processor and computer systems
US6920154B1 (en) 2001-12-17 2005-07-19 Supergate Technology Usa, Inc. Architectures for a modularized data optimization engine and methods therefor
US6721229B1 (en) 2001-12-19 2004-04-13 Network Equipment Technologies, Inc. Method and apparatus for using SDRAM to read and write data without latency
US6944731B2 (en) 2001-12-19 2005-09-13 Agere Systems Inc. Dynamic random access memory system with bank conflict avoidance feature
US6839797B2 (en) 2001-12-21 2005-01-04 Agere Systems, Inc. Multi-bank scheduling to improve performance on tree accesses in a DRAM based random access memory subsystem
US7212536B2 (en) 2001-12-27 2007-05-01 Alcatel-Lucent Canada Inc. User priority mapping in bridged VLANS
US7237058B2 (en) 2002-01-14 2007-06-26 Netlogic Microsystems, Inc. Input data selection for content addressable memory
ATE338979T1 (en) 2002-02-06 2006-09-15 Koninkl Philips Electronics Nv ADDRESS SPACE, BUS SYSTEM, STORAGE CONTROL AND FACILITY SYSTEM
US6781990B1 (en) 2002-02-11 2004-08-24 Extreme Networks Method and system for managing traffic in a packet network environment
US20030152096A1 (en) 2002-02-13 2003-08-14 Korey Chapman Intelligent no packet loss networking
US7719980B2 (en) 2002-02-19 2010-05-18 Broadcom Corporation Method and apparatus for flexible frame processing and classification engine
US6836808B2 (en) 2002-02-25 2004-12-28 International Business Machines Corporation Pipelined packet processing
US7190696B1 (en) 2002-03-05 2007-03-13 Force10 Networks, Inc. System and method for distributing packets among a plurality of paths to a destination
US7126956B2 (en) 2002-03-05 2006-10-24 Applied Micro Circuits Corporation System to provide fractional bandwidth data communications services
US7161948B2 (en) 2002-03-15 2007-01-09 Broadcom Corporation High speed protocol for interconnecting modular network devices
US7478167B2 (en) 2002-03-18 2009-01-13 Nortel Networks Limited Resource allocation using an auto-discovery mechanism for provider-provisioned layer-2 and layer-3 virtual private networks
US6842791B2 (en) 2002-03-20 2005-01-11 Intel Corporation Method and apparatus for memory efficient fast VLAN lookups and inserts in hardware-based packet switches
JP2003289359A (en) 2002-03-28 2003-10-10 Hitachi Kokusai Electric Inc Mobile telephone set
DE10219154A1 (en) 2002-04-29 2003-11-20 Siemens Ag Procedure for monitoring the availability of connections in MPLS networks
US20120155466A1 (en) 2002-05-06 2012-06-21 Ian Edward Davis Method and apparatus for efficiently processing data packets in a computer network
US7266117B1 (en) * 2002-05-06 2007-09-04 Foundry Networks, Inc. System architecture for very fast ethernet blade
EP1502396A1 (en) 2002-05-08 2005-02-02 Siemens Aktiengesellschaft Method for assisting equivalent circuits in mpls networks
US7151797B2 (en) 2002-05-14 2006-12-19 Limberg Allen Leroy Adaptive K-factor-improvement filter for receiver of radio signals subject to multipath distortion
US7366100B2 (en) 2002-06-04 2008-04-29 Lucent Technologies Inc. Method and apparatus for multipath processing
US20030227913A1 (en) 2002-06-05 2003-12-11 Litchfield Communications, Inc. Adaptive timing recovery of synchronous transport signals
WO2004004251A1 (en) 2002-07-01 2004-01-08 Ipsquare Inc. Semiconductor circuit device, packet processing method, management system, management method, and packet processing method
US6671275B1 (en) 2002-08-02 2003-12-30 Foundry Networks, Inc. Cross-point switch with deadlock prevention
US20040054867A1 (en) 2002-09-13 2004-03-18 Paulus Stravers Translation lookaside buffer
US7277425B1 (en) 2002-10-21 2007-10-02 Force10 Networks, Inc. High-speed router switching architecture
US7802049B2 (en) 2002-10-30 2010-09-21 Intel Corporation Links having flexible lane allocation
US7403536B2 (en) 2002-12-19 2008-07-22 International Business Machines Corporation Method and system for resequencing data packets switched through a parallel packet switch
US7505416B2 (en) 2003-03-31 2009-03-17 Finisar Corporation Network tap with integrated circuitry
US7561590B1 (en) 2003-05-05 2009-07-14 Marvell International Ltd. Network switch having virtual input queues for flow control
US7558268B2 (en) 2003-05-08 2009-07-07 Samsung Electronics Co., Ltd. Apparatus and method for combining forwarding tables in a distributed architecture router
US7391769B2 (en) 2003-06-27 2008-06-24 Lucent Technologies Inc. Packet aggregation for real time services on packet data networks
US20050138276A1 (en) 2003-12-17 2005-06-23 Intel Corporation Methods and apparatus for high bandwidth random access using dynamic random access memory
KR100600673B1 (en) 2003-12-18 2006-07-18 한국전자통신연구원 A method for requesting and reporting channel quality information in wireless system and apparatus thereof
JP2007525883A (en) 2004-01-14 2007-09-06 リヴァーストーン ネットワークス インコーポレーテッド Processing usage management in network nodes
US7249306B2 (en) 2004-02-20 2007-07-24 Nvidia Corporation System and method for generating 128-bit cyclic redundancy check values with 32-bit granularity
US8806103B2 (en) 2004-04-28 2014-08-12 Hewlett-Packard Development Company, L.P. System and method for interleaving memory
US8054751B2 (en) 2004-05-10 2011-11-08 Alcatel Lucent Remote access link fault indication mechanism
US7228509B1 (en) 2004-08-20 2007-06-05 Altera Corporation Design tools for configurable serial communications protocols
US8259562B2 (en) 2004-10-07 2012-09-04 Cisco Technology, Inc. Wiring closet redundancy
US7680107B2 (en) 2004-11-30 2010-03-16 Broadcom Corporation High speed trunking in a network device
US7499395B2 (en) 2005-03-18 2009-03-03 Cisco Technology, Inc. BFD rate-limiting and automatic session activation
KR100687739B1 (en) 2005-03-29 2007-02-27 한국전자통신연구원 Method for monitoring link performance and diagnosing active status of link for Ethernet Passive Optical Network
US8730814B2 (en) 2005-05-25 2014-05-20 Alcatel Lucent Communication network connection failure protection methods and systems
US7424666B2 (en) 2005-09-26 2008-09-09 Intel Corporation Method and apparatus to detect/manage faults in a system
US8131628B2 (en) 2006-01-29 2012-03-06 INXS Technologies Ltd. System, method, and computer program for realizing a unified brokerage access channel
US20070288690A1 (en) 2006-06-13 2007-12-13 Foundry Networks, Inc. High bandwidth, high capacity look-up table implementation in dynamic random access memory
US7827471B2 (en) 2006-10-12 2010-11-02 Intel Corporation Determining message residue using a set of polynomials

Patent Citations (99)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3866175A (en) * 1974-04-24 1975-02-11 Ncr Co Data communication system between a central computer and a plurality of data terminals
US4723243A (en) * 1985-12-02 1988-02-02 Advanced Micro Devices, Inc. CRC calculation machine with variable bit boundary
US4807280A (en) * 1987-09-18 1989-02-21 Pacific Bell Cross-connect switch
US4985889A (en) * 1988-02-04 1991-01-15 Sprint International Communications Corporation Data packet switching
US4896277A (en) * 1988-05-03 1990-01-23 Thermo King Corporation Method of mapping refrigerated containers in a power line carrier based monitoring system
US5101404A (en) * 1988-08-26 1992-03-31 Hitachi, Ltd. Signalling apparatus for use in an ATM switching system
US5280582A (en) * 1989-08-29 1994-01-18 Digital Equipment Corporation No-owner frame and multiple token removal mechanism for token ring networks
US5875200A (en) * 1990-11-08 1999-02-23 Cirrus Logic, Inc. Reed-Solomon code system employing k-bit serial techniques for encoding and burst error trapping
US5299195A (en) * 1991-01-18 1994-03-29 National Semiconductor Corporation Repeater interface controller with multiple port node interfaces
US5734826A (en) * 1991-03-29 1998-03-31 International Business Machines Corporation Variable cyclic redundancy coding method and apparatus for use in a multistage network
US5490258A (en) * 1991-07-29 1996-02-06 Fenner; Peter R. Associative memory for very large key spaces
US5287477A (en) * 1991-08-07 1994-02-15 Hewlett-Packard Company Memory-resource-driven arbitration
US5282196A (en) * 1991-10-15 1994-01-25 Hughes Aircraft Company Bursted and non-bursted data router
US5195181A (en) * 1992-01-10 1993-03-16 Digital Equipment Corporation Message processing system having separate message receiving and transmitting processors with message processing being distributed between the separate processors
US5392279A (en) * 1992-01-17 1995-02-21 Fujitsu Limited Interconnection network with self-routing function
US5390173A (en) * 1992-10-22 1995-02-14 Digital Equipment Corporation Packet format in hub for packet data communications system
US5299190A (en) * 1992-12-18 1994-03-29 International Business Machines Corporation Two-dimensional round-robin scheduling mechanism for switches with multiple input queues
US5600795A (en) * 1993-08-28 1997-02-04 U.S. Philips Corporation Local network operating in asynchronous transfer mode (ATM) generating control cell containing information about the user, address of the station, and user-related identification
US6522656B1 (en) * 1994-09-12 2003-02-18 3Com Corporation Distributed processing ethernet switch with adaptive cut-through switching
US5862350A (en) * 1994-12-22 1999-01-19 Intel Corporation Method and mechanism for maintaining integrity within SCSI bus with hot insertion
US5598410A (en) * 1994-12-29 1997-01-28 Storage Technology Corporation Method and apparatus for accelerated packet processing
US5721819A (en) * 1995-05-05 1998-02-24 Silicon Graphics Corporation Programmable, distributed network routing
US5732080A (en) * 1995-07-12 1998-03-24 Bay Networks, Inc. Method and apparatus for controlling data flow within a switching device
US5870538A (en) * 1995-07-19 1999-02-09 Fujitsu Network Communications, Inc. Switch fabric controller comparator system and method
US5872769A (en) * 1995-07-19 1999-02-16 Fujitsu Network Communications, Inc. Linked list structures for multiple levels of control in an ATM switch
US6185222B1 (en) * 1995-09-28 2001-02-06 Cisco Technology, Inc. Asymmetric switch architecture for use in a network switch node
US5864555A (en) * 1996-07-05 1999-01-26 General Datacomm, Inc. Method and apparatus for generating a proxy connection endpoint for operation administration and management (OAM) asynchronous transfer mode (ATM) cells
US5872783A (en) * 1996-07-24 1999-02-16 Cisco Systems, Inc. Arrangement for rendering forwarding decisions for packets transferred among network switches
US5867675A (en) * 1996-08-06 1999-02-02 Compaq Computer Corp Apparatus and method for combining data streams with programmable wait states
US6035414A (en) * 1996-11-08 2000-03-07 Hitachi, Ltd. Reliability of crossbar switches in an information processing system
US6031843A (en) * 1996-11-21 2000-02-29 Alcatel Data Networks Inc. Digital communications switching fabric
US6172990B1 (en) * 1997-06-19 2001-01-09 Xaqti Corporation Media access control micro-RISC stream processor and method for implementing the same
US6195335B1 (en) * 1997-06-27 2001-02-27 International Business Machines Corporation Data switch
US6016310A (en) * 1997-06-30 2000-01-18 Sun Microsystems, Inc. Trunking support in a high performance network device
US6178520B1 (en) * 1997-07-31 2001-01-23 Lsi Logic Corporation Software recognition of drive removal or insertion in a storage system
US6678248B1 (en) * 1997-08-29 2004-01-13 Extreme Networks Policy based quality of service
US6343072B1 (en) * 1997-10-01 2002-01-29 Cisco Technology, Inc. Single-chip architecture for shared-memory router
US6023471A (en) * 1997-10-07 2000-02-08 Extreme Networks Network interconnect device and protocol for communicating data among packet forwarding devices
US6038288A (en) * 1997-12-31 2000-03-14 Thomas; Gene Gilles System and method for maintenance arbitration at a switching node
US6859438B2 (en) * 1998-02-03 2005-02-22 Extreme Networks, Inc. Policy based quality of service
US6363077B1 (en) * 1998-02-13 2002-03-26 Broadcom Corporation Load balancing in link aggregation and trunking
US6356942B1 (en) * 1998-05-13 2002-03-12 Axis Ab Integrated circuit and method for bringing an integrated circuit to execute instructions
US6505281B1 (en) * 1998-06-02 2003-01-07 Raymond C. Sherry Hard disk drives employing high speed distribution bus
US6181699B1 (en) * 1998-07-01 2001-01-30 National Semiconductor Corporation Apparatus and method of assigning VLAN tags
US6335935B2 (en) * 1998-07-08 2002-01-01 Broadcom Corporation Network switching architecture with fast filtering processor
US6335932B2 (en) * 1998-07-08 2002-01-01 Broadcom Corporation High performance self balancing low cost network switching architecture based on distributed hierarchical shared memory
US6683872B1 (en) * 1998-07-31 2004-01-27 The University Of Tokyo Variable rate digital switching system
US6982974B1 (en) * 1999-01-15 2006-01-03 Cisco Technology, Inc. Method and apparatus for a rearrangeably non-blocking switching matrix
US6532229B1 (en) * 1999-01-15 2003-03-11 Brian Weatherred Johnson Low cost link aggregation method and system
US7176911B1 (en) * 1999-02-02 2007-02-13 Olympus Corporation Solid-state imaging apparatus
US6510138B1 (en) * 1999-02-25 2003-01-21 Fairchild Semiconductor Corporation Network switch with head of line input buffer queue clearing
US6839346B1 (en) * 1999-04-05 2005-01-04 Nec Corporation Packet switching apparatus with high speed routing function
US6687217B1 (en) * 1999-05-04 2004-02-03 Carriercomm, Inc. Method of and system for one plus one protection for radio equipment
US6201492B1 (en) * 1999-05-28 2001-03-13 Cirrus Logic, Inc. Techniques for converting a plurality of channels continuously in an A/D converter
US6842422B1 (en) * 1999-06-15 2005-01-11 Marconi Communications, Inc. Data striping based switching system
US6697359B1 (en) * 1999-07-02 2004-02-24 Ancor Communications, Inc. High performance switch fabric element and switch systems
US6356550B1 (en) * 1999-07-30 2002-03-12 Mayan Networks Corporation Flexible time division multiplexed bus using sonet formatting
US20050041684A1 (en) * 1999-10-01 2005-02-24 Agilent Technologies, Inc. Multi-channel network monitoring apparatus, signal replicating device, and systems including such apparatus and devices, and enclosure for multi-processor equipment
US6687247B1 (en) * 1999-10-27 2004-02-03 Cisco Technology, Inc. Architecture for high speed class of service enabled linecard
US6856600B1 (en) * 2000-01-04 2005-02-15 Cisco Technology, Inc. Method and apparatus for isolating faults in a switching matrix
US6681332B1 (en) * 2000-03-13 2004-01-20 Analog Devices, Inc. System and method to place a device in power down modes/states and restore back to first mode/state within user-controlled time window
US6351143B1 (en) * 2000-05-01 2002-02-26 Xilinx, Inc. Content-addressable memory implemented using programmable logic
US20020001307A1 (en) * 2000-05-20 2002-01-03 Equipe Communications Corporation VPI/VCI availability index
US20020012585A1 (en) * 2000-06-09 2002-01-31 Broadcom Corporation Gigabit switch with fast filtering processor
US6993032B1 (en) * 2000-06-29 2006-01-31 Intel Corporation Buffer arrangements to support differential link distances at full bandwidth
US6696917B1 (en) * 2000-09-21 2004-02-24 Nortel Networks Limited Folded Clos architecture switching
US6854117B1 (en) * 2000-10-31 2005-02-08 Caspian Networks, Inc. Parallel network processor array
US6697368B2 (en) * 2000-11-17 2004-02-24 Foundry Networks, Inc. High-performance network switch
US20100034215A1 (en) * 2000-11-17 2010-02-11 Foundry Networks, Inc. Backplane Interface Adapter with Error Control
US6691202B2 (en) * 2000-12-22 2004-02-10 Lucent Technologies Inc. Ethernet cross point switch with reduced connections by using column control buses
US7324509B2 (en) * 2001-03-02 2008-01-29 Broadcom Corporation Efficient optimization algorithm in memory utilization for network applications
US6990102B1 (en) * 2001-05-10 2006-01-24 Advanced Micro Devices, Inc. Parallel lookup tables for locating information in a packet switched network
US20030009466A1 (en) * 2001-06-21 2003-01-09 Ta John D. C. Search engine with pipeline structure
US20030012198A1 (en) * 2001-07-11 2003-01-16 Nec Corporation Packet processing unit
US20030033435A1 (en) * 2001-07-17 2003-02-13 Hanner Brian D. System and method for multidimensional data compression
US7005812B2 (en) * 2001-07-23 2006-02-28 Lawrence Hardy Mitchell Commutation converter for a brushless servo motor
US20030048785A1 (en) * 2001-08-28 2003-03-13 International Business Machines Corporation Network processor with single interface supporting tree search engine and CAM
US20030043848A1 (en) * 2001-08-30 2003-03-06 Sonksen Bradley Stephen Method and apparatus for data item processing control
US20030043800A1 (en) * 2001-08-30 2003-03-06 Sonksen Bradley Stephen Dynamic data item processing
US6996663B1 (en) * 2001-09-12 2006-02-07 Cisco Technology, Inc. Method and apparatus for performing address translation using a CAM
US20030061459A1 (en) * 2001-09-27 2003-03-27 Nagi Aboulenein Method and apparatus for memory access scheduling to reduce memory access latency
US7185141B1 (en) * 2001-12-27 2007-02-27 Netlogic Microsystems, Inc. Apparatus and method for associating information values with portions of a content addressable memory (CAM) device
US7649885B1 (en) * 2002-05-06 2010-01-19 Foundry Networks, Inc. Network routing system for enhanced efficiency and monitoring capability
US20110002340A1 (en) * 2002-05-06 2011-01-06 Foundry Networks, Inc. Pipeline method and system for switching packets
US20080002707A1 (en) * 2002-05-06 2008-01-03 Davis Ian E Flexible method for processing data packets in a network routing system for enhanced efficiency and monitoring capability
US20060034452A1 (en) * 2002-06-20 2006-02-16 Hitachi, Ltd. Code calculating device
US20040022263A1 (en) * 2002-08-02 2004-02-05 Xiaodong Zhao Cross point switch with out-of-band parameter fine tuning
US20040028060A1 (en) * 2002-08-08 2004-02-12 Samsung Electronics Co., Ltd. Link state synchronization method and apparatus on ad-hoc network, and data structure therefor
US7185266B2 (en) * 2003-02-12 2007-02-27 Alacritech, Inc. Network interface device for error detection using partial CRCS of variable length message portions
US20050010849A1 (en) * 2003-03-18 2005-01-13 Cisco Technology, Inc., A Corporation Of California Method and system for emulating a Fibre Channel link over a SONET/SDH path
US20050010630A1 (en) * 2003-05-13 2005-01-13 International Business Machines Corporation Method and apparatus for determining a remainder in a polynomial ring
US20100046521A1 (en) * 2003-05-15 2010-02-25 Foundry Networks, Inc. System and Method for High Speed Packet Transmission
US20060031610A1 (en) * 2004-08-03 2006-02-09 Liav Ori R FIFO sub-system with in-line correction
US20070038798A1 (en) * 2004-09-10 2007-02-15 Bouchard Gregg A Selective replication of data structures
US7657703B1 (en) * 2004-10-29 2010-02-02 Foundry Networks, Inc. Double density content addressable memory (CAM) lookup scheme
US20080031263A1 (en) * 2006-08-07 2008-02-07 Cisco Technology, Inc. Method and apparatus for load balancing over virtual network links
US20080037544A1 (en) * 2006-08-11 2008-02-14 Hiroki Yano Device and Method for Relaying Packets
US20080049742A1 (en) * 2006-08-22 2008-02-28 Deepak Bansal System and method for ecmp load sharing
US20110044340A1 (en) * 2006-08-22 2011-02-24 Foundry Networks, Llc System and method for ecmp load sharing

Cited By (34)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7948872B2 (en) 2000-11-17 2011-05-24 Foundry Networks, Llc Backplane interface adapter with error control and redundant fabric
US9030937B2 (en) 2000-11-17 2015-05-12 Foundry Networks, Llc Backplane interface adapter with error control and redundant fabric
US7978702B2 (en) 2000-11-17 2011-07-12 Foundry Networks, Llc Backplane interface adapter
US8964754B2 (en) 2000-11-17 2015-02-24 Foundry Networks, Llc Backplane interface adapter with error control and redundant fabric
US8619781B2 (en) 2000-11-17 2013-12-31 Foundry Networks, Llc Backplane interface adapter with error control and redundant fabric
US8514716B2 (en) 2000-11-17 2013-08-20 Foundry Networks, Llc Backplane interface adapter with error control and redundant fabric
US8170044B2 (en) 2002-05-06 2012-05-01 Foundry Networks, Llc Pipeline method and system for switching packets
US8989202B2 (en) 2002-05-06 2015-03-24 Foundry Networks, Llc Pipeline method and system for switching packets
US8671219B2 (en) 2002-05-06 2014-03-11 Foundry Networks, Llc Method and apparatus for efficiently processing data packets in a computer network
US8194666B2 (en) 2002-05-06 2012-06-05 Foundry Networks, Llc Flexible method for processing data packets in a network routing system for enhanced efficiency and monitoring capability
US8718051B2 (en) 2003-05-15 2014-05-06 Foundry Networks, Llc System and method for high speed packet transmission
US8811390B2 (en) 2003-05-15 2014-08-19 Foundry Networks, Llc System and method for high speed packet transmission
US9461940B2 (en) 2003-05-15 2016-10-04 Foundry Networks, Llc System and method for high speed packet transmission
US8493988B2 (en) 2004-03-26 2013-07-23 Foundry Networks, Llc Method and apparatus for aggregating input data streams
US9338100B2 (en) 2004-03-26 2016-05-10 Foundry Networks, Llc Method and apparatus for aggregating input data streams
US8730961B1 (en) 2004-04-26 2014-05-20 Foundry Networks, Llc System and method for optimizing router lookup
US7953922B2 (en) 2004-10-29 2011-05-31 Foundry Networks, Llc Double density content addressable memory (CAM) lookup scheme
US7953923B2 (en) 2004-10-29 2011-05-31 Foundry Networks, Llc Double density content addressable memory (CAM) lookup scheme
US20100161894A1 (en) * 2004-10-29 2010-06-24 Foundry Networks, Inc. Double density content addressable memory (cam) lookup scheme
US9378005B2 (en) 2005-12-28 2016-06-28 Foundry Networks, Llc Hitless software upgrades
US8448162B2 (en) 2005-12-28 2013-05-21 Foundry Networks, Llc Hitless software upgrades
US8238255B2 (en) 2006-11-22 2012-08-07 Foundry Networks, Llc Recovering from failures without impact on data traffic in a shared bus architecture
US9030943B2 (en) 2006-11-22 2015-05-12 Foundry Networks, Llc Recovering from failures without impact on data traffic in a shared bus architecture
US9112780B2 (en) 2007-01-11 2015-08-18 Foundry Networks, Llc Techniques for processing incoming failure detection protocol packets
US8155011B2 (en) 2007-01-11 2012-04-10 Foundry Networks, Llc Techniques for using dual memory structures for processing failure detection protocol packets
US7978614B2 (en) 2007-01-11 2011-07-12 Foundry Network, LLC Techniques for detecting non-receipt of fault detection protocol packets
US8395996B2 (en) 2007-01-11 2013-03-12 Foundry Networks, Llc Techniques for processing incoming failure detection protocol packets
US8037399B2 (en) 2007-07-18 2011-10-11 Foundry Networks, Llc Techniques for segmented CRC design in high speed networks
US8149839B1 (en) 2007-09-26 2012-04-03 Foundry Networks, Llc Selection of trunk ports and paths using rotation
US8509236B2 (en) 2007-09-26 2013-08-13 Foundry Networks, Llc Techniques for selecting paths and/or trunk ports for forwarding traffic flows
US8090901B2 (en) 2009-05-14 2012-01-03 Brocade Communications Systems, Inc. TCAM management approach that minimize movements
US9166818B2 (en) 2009-09-21 2015-10-20 Brocade Communications Systems, Inc. Provisioning single or multistage networks using ethernet service instances (ESIs)
US8599850B2 (en) 2009-09-21 2013-12-03 Brocade Communications Systems, Inc. Provisioning single or multistage networks using ethernet service instances (ESIs)
US9946466B2 (en) 2014-08-19 2018-04-17 Samsung Electronics Co., Ltd. Mobile electronic device including embedded memory

Also Published As

Publication number Publication date
US7738450B1 (en) 2010-06-15
US7266117B1 (en) 2007-09-04

Similar Documents

Publication Publication Date Title
US7738450B1 (en) System architecture for very fast ethernet blade
US8989202B2 (en) Pipeline method and system for switching packets
US9461940B2 (en) System and method for high speed packet transmission
US9509639B1 (en) Switch device having a plurality of processing cores
US8005084B2 (en) Mirroring in a network device
US6515993B1 (en) Method and apparatus for manipulating VLAN tags
US6584101B2 (en) Communication method for packet switching systems
US6680945B1 (en) Method and apparatus for support of tagging and untagging per VLAN per port
US6424659B2 (en) Multi-layer switching apparatus and method
JP5760083B2 (en) Method and apparatus for fast switching from a primary multicast tree to a standby multicast tree
JP4057067B2 (en) Mechanism for replacing packet fields in multi-layer switching network elements
US9544216B2 (en) Mesh mirroring with path tags
US7570639B2 (en) Multicast trunking in a network device
US20060114907A1 (en) Cut-through switching in a network device
US7436775B2 (en) Software configurable cluster-based router using stock personal computers as cluster nodes
US20100040074A1 (en) Multi-speed cut through operation in fibre channel switches
ES2340954T3 (en) MULTI-PROTOCOL ENGINE FOR RECONFIGURABLE BITS CURRENT PROCESSING IN HIGH SPEED NETWORKS.
US7756124B2 (en) Encapsulating packets for network chip conduit port
US7710994B1 (en) Systems and methods for efficient multicast handling
US6778547B1 (en) Method and apparatus for improving throughput of a rules checker logic
US6693906B1 (en) Apparatus and method for buffer-free evaluation of packet data bytes with multiple min terms
US7009973B2 (en) Switch using a segmented ring

Legal Events

Date Code Title Description
AS Assignment

Owner name: FOUNDRY NETWORKS, LLC, CALIFORNIA

Free format text: CHANGE OF NAME;ASSIGNOR:FOUNDRY NETWORKS, INC.;REEL/FRAME:024733/0739

Effective date: 20090511

STCB Information on status: application discontinuation

Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION