CA2157319C - Communication network design method and communication equipment - Google Patents

Abstract

In order to establish virtual channel networks on an ATM network, a different virtual channel handler interconnection network for each media or service consisting of information to be transferred is established on a virtual path network; and virtual channel networks corresponding to these media or services are formed by means of these respective virtual channel handler interconnection networks, whereby a plurality of independent virtual channel networks for different media or services are established simultaneously on a single physical network.

Description

COMMUNICATION NETWORK DESIGN
METHOD AND C011~CATION EQUIPMENT
$ACKGROUND OF THE 1)VVENTTON
1. Field of the Invention This invention is utilized in asynchronous transfer mode (AT1VI) communication networks, and relates in particular, to techniques for actualizing a different network topology for each media or service on a single physical network and communication equipment associated therewith.

2. Descri lion of the Related Art FIG. I and FIG. 2 show the configurations of two conventional communication networks. The network shown in FIG. 1 may be utilized for telephony, and comprises: local switches 11 and 13 accvnunodating subscriber lines; transit switch 12 interconnecting local switches 11 and 13;
transmission links 1~ interconnecting local switches 11 and 13 and transit switch 12; and cross-connect switches 15. FIG. 2 illustrates a communication system typified by Ethernet, wherein cross-connect switches 21 and transmission finks 22 are connected in a single loop or a single bus, and add/drop multiplexers 23 are providod for inserting and breaching off information via a subscriber lint 20.
With conventional network configurations, two communication networks have been necessary for the different topologies of the two services mentioned above. This is so because a service, such as telephony, occupies considerable resources due to the fact that each call is at a constant bit rate (C$R) and has a comparatively long connection time. Also, traffic is dispersed over a wide area, and as a result, transit switches are required due to the necessity o~ concentrating circuits for long-distance communi-cation. On the other hand, other services transmit data at a variable bit rate (VBR) with the traffic itself in the form of bursts, and although the peak rate is high, the average rate is low and some degree of delay is allowed. A
network utilized for data transmission, therefore, has a topology whereby a bus or a loop is shared as a single resource.

Because services with different tragic characteristics have thus used physical networks with separate topologies, there is the problem that there will be a different optimum topology every time a new service or media appears in this multimedia age, resulting in the need for many networks.
To put it another way, each service will require its own coaxial or apfiical fibre cable. However, this arrangement lacks flexibility.
The present invention provides a communication network design method which: 1) physically connects, by means of a transmission link network, a plurality of virtual chamiel handlers (VCHs) which perform switching in units of v'srhial channels (VCs), and a virtue! path handler (VPI~ which cbanges over information transfer routes in units of virtual paths (VPs) that accommodate a plurality of VCs; 2) establishes a YP
network on a fission link netvvork, the VP network being formed from VPs which logically coanca a plurality of VCHs; and 3) establishes a VC
network on the VP network, the VC network being formed from VCs which logically coronet subscribers. The comznunication network desigt method is eharactesized in that a different VCH interconnection network fvr each media or service is established on the VP network, and a VC network corresponding to a given media or service is formed by means of each VCH
interconnection network, whereby a plurality of independent VC networks for different media or services are established simultaneously on a single physical network.
In other words, in order to actualize an ideal topology for each media or servicx on a single physical network, logical connections between VCHs arc established far each media or service in an ATM network comprising a VC network and a VP network. The VCs of these media or services are established an the VCH interconnection networks thereby obtained.

Z hC term "lnedla" UsCd h~eln p~S t0 a CO~n11n1Catloll SyStenl with a distinctive set of commuaic~tion standards, quality. triode, capacity, and so forth,. For example, "media" encompasses voice comraimicadon and image communication; the latocr is subdivided into still and moving image.
Still imagcs are ftuther subdividcd into ra~rded images (FAQ and display images (LCD), while moving images are subdivided into s~ndard 'f V
images and simplified TV . A plurality of media generally cocxist in a single communication system (this is tamed "multimedia") and each media uses a different kind of signal and has a different signal quality requirement. The quality that users require for each mcdia includes specified valucs for call blocking probability, cell loss probabfiity, and delay, for example.
The VC and VP networks of sa ATM network corrcspond, respectively, to ~tlte circuit and path netarorks in a conventional netvvork.
The VP network is a network which is established logically on a physical network, and it can logically all VCHs in a mesh The VC nctwork is a network which logically oonnocts sabscribcrs an the VP network by means of VCs.
A VCH interconnection network, as used in this invention, is formed by means of a logical network which connects subscriber VCHs used by a given media or service and a plurality of transit VCRs. VCs for this media or service are establishcd by selection on this VCH interconnection network only. A VCH interconnection network for a given media or service corresponds to the VC network of that media or service. However, since a VC network is a network with subscribers as the ternninal points, it establishes connections between subscribers and VCHs as well as between VCHs; whereas a VCH interconnection nctyvork is a network which establishes logical connections between VCHs. Although a VCH
interconnection network does not take connections to subscribers into consideration, designing VCH interconnection networks suited to each media or service is equivalent to designing VC networks suited to the same mcdia oT SCTVICeS.

_ ~~~~3.~~
AlI VCbTs can be cod in a mesh by the a VP which does not pass through any other VCH can be established between any two given VCHs. The establishment of a VCI:I interconnection network may be performed by assigning different VPs to each media or service consisting of information to be transferred. Alternatively, it is also feasible to establish a VCH interconnection network by aDowing a plurality of media or services with the same destination VCH to be assigned to the same ms's. In this case, priority control should be performed bdvveen VPs with the same destination in accordance with a priority ranking of the media or services.
The network to~logies of a plurality of VCH interconnection networks for diffezem kinds of media rn' service may also be dcsi~n~
simultaneously after taking their averall optimi~ion into consideration.
Alternatively, they may be designed in a particular order.
When designing in a partiGUiar omdt~, an optimized network topology may be obtained by calculating the network resource efficiency from the working YP ratio, which is defined as the ratio of working VP capacity to total VP capacity (i.e., both working and back up VPs). and from the VP
use efficiency, which has been calculated from the traffic type, traffic demand and required cell loss probability in VCs established between _ .
subscxibcrs utilizing this VCH interconnoction network. In this case, the number of disjoint routes D (i.e., the number of paths between origin and destination VCHs which do not pass through the same YPHs or VP links) may be set as a parameter of trial network topologies, and satisfying the given number of disjoint routes D may be set as a constraint. A trial network topology, wherein the total length of all the VP links is minimized, is then formed, and the number of disjoint routes D is changed when the trial network topology is changed. Tl~e network resource efficiency may then be calculated for the trial network topology obtained.
Network topologies can also be designed simultaneously so that overall network cost is minimized. In this case, connection or disconnection of VP connections between VCHs for establishing a VCH interconnection network may be performed in parallel with establishing VCH
interconnection networks for other media or services.
When a plurality of network topologies are desigacd in a particular order, the design priority order should be determined in the order' of decreasing network cost of the media or services supported by the VGH
interconnection networks in question- After network topologies have been designed in this manner, the proportion of overall network cost accounted for by the network cost of the VCH interconnection network for each media or service can be calculated and the design priority order can be obtained in accordance with the results of this calculation. If this design priority order is different from a previous trial, the design of ac~rork topologies can be repeated in accordance with the new design priority order_ A limit may be set for the number of such repetitions.
It is also feasible to classify media or services into a priority group and a non priority gmup according to their network cost. The design of VCH intercormection network topologies for the priority group is performed in the order of decreasing network cost, while the design of VCH
interconnection network topologies for the non priority group is performed simultaneously so that overall network cost is minimized.
When designing network topology, it is desirable to set an upper limit for the number of VPs passed through from an origin VCH to a destination VCH.
The present invention also provides communication equipment which has: 1) a plurality of VCHs which perform switching in units of VCs; Z) a VPH which changes over routes in units of VPs which accommodate a plurality of VCs; 3) a transmission Iink network which physically connects the plurality of VCHs and the VPH; attd 4) a higher order means which establishes a VP network on the transmission link network, the VP network being formed from VPs which logically connect the plurality of VCHs, and which also establishes a VC network on this vP network, the VC network being formed from VCs which Logically connect subscribers. The communication equipment is characterized in that the higher order means ~s-0 includes: 1) a VCH interconnection network establishing means which establishes, on the VP network, a different VCH interconnection network for each media or service;
and 2) a means which forms a VC network corresponding to a given media or service on the basis of each VCH interconnection network, whereby a plurality of VC
networks for different media or services are established simultaneously on a single physical network.
According to one aspect of the invention, there is provided a method of designing an ATM network comprising the steps of: physically connecting, by means of a transmission link network, a plurality of virtual channel handlers having subscribers connected thereto, said plurality of virtual channel handlers switching data to be transmitted over said transmission link network in units of virtual channels, and a virtual path handler which changes information transfer routes between virtual channel handlers in units of virtual paths that accommodate a plurality of said virtual channels; .
establishing a virtual path network on said transmission link network, said virtual path network being formed from virtual paths which logically connect said plurality of virtual channel handlers; establishing a virtual channel network on said virtual path network, said virtual channel network being formed from virtual channels which logically connect subscribers on said transmission link network; establishing a different virtual channel handler interconnection network on said transmission link network for each media having information to be transferred -on said transmission link network;
and forming said virtual channel network corresponding to a given media by means of each virtual channel handler interconnection network, so that a plurality of independent virtual channel networks for different media are established simultaneously on a single physical network.
According to a further aspect of the invention, there is provided:
communication ATM network comprising: a plurality of virtual channel handlers which perform switching in units of virtual channels; at least one subscriber connected to each virtual channel handler; a virtual path handler which changes over routes of data in units of virtual paths which accommodate a plurality of said virtual channels; a transmission link network which physically connects said plurality of virtual channel handlers and said virtual path handlers; and a control means which establishes a virtual path network on said transmission link network, said virtual path network being formed from virtual paths which logically connect said plurality of virtual channel handlers, and which also establishes a virtual channel network on said virtual path 0 network, said virtual channel network being formed from virtual channels which logically connect said subscribers on said transmission link network;
said control means including: a virtual control handler interconnection network establishing means which establishes, on said virtual path network, a different virtual channel handler interconnection network for each media having information to be transferred; and a means for establishing virtual channel networks corresponding to said media by means of respective virtual channel handler interconnection networks, whereby different virtual channel networks for different media are established simultaneously on a single physical network.
BRIEF DESCRIPTION OF THE DRAWINGS
Specific embodiments of the invention will now be described by way of example only, with the use of drawings in which:
FIG. 1 shows a conventional communication network configuration;
FIG. 2 shows another conventional communication network configuration;
FIG. 3 is a block diagram showing communication equipment according to a first embodiment of the present invention;
FIGS. 4(a)-4(c) show examples of logical network topologies for various services;
FIGS. 5(a)-5(d) show configurations for actualizing a plurality of topologies on a single physical network;
FIG. 6 shows an example of a VCH configuration;
FIG. 7 shows another example of a VCH configuration;
FIG. 8 explains the manner in which VPs are disposed on a transmission link;
FIG. 9 shows another example of a VCH configuration;
FIG. 10 is a flowchart showing a conventional method for establishing a VC
network in an ATM network;
- 6~ -_ 2~r~~~.~~
FIG. 11 is a flowchart showing a method for obtaining a different optimum VC network for each media;
FIG. 12 is a flowchart showing an example of a method for forming a trial VCH interconnection network topology;
FIG. 13 is a flowchart showing an example of a method for changing the trial VCH interconnection nctwark topology;
FIGS. 14(a~14(d) show trial nttwork topologies when the number of disjoint routes D has been sct as a parameter;
FIG. 15 shows an example- of resalts from an optimization of VCH
interconnection network topology;
FIG. 16 shows another example of results from as optimization of VCH interconnection network topology;
FIG. 17 shows the ovaall flow of a method for establishing a plurality of VCH intcrconnedion that minimize cost;
FIG. 18 shows the dean flow of tech VCH interconnection network topology;
FIG. 19 shows the design flow of a VCH interconnection artwork topology for one given media;
FIG. 20 is a flowchart sbowing a method for designing a VCH
2p interconnection network topology for each media after setting the design priority order of the VCH interconnection networJcs in the order of decreasing network cost;
FIG. 2I is a flowchart showing snottier method for designing a VCH
interconnection network topology for each media;
FIG. 22 shows another example of a method for designing a VCH
interconnection network topology for a particular media;
FIG. 23 shows still another method for designing VCH
interconnection network topologics;
FIG. 24 shows a method for simultaneously designing VCH
interconnection network topologies for a plurality of media;
FIG. 25 shows an cxamplc of a method for designing VCH
interconnection network topologies simultaneously for each of M mcdia;

FIG. 26 shows the flow of a for determining the order of pairs of VCHs between which VP connections will be added or deleted;
FIG. 27 explains the method illustrated in FIG. 26;
FIG. 28 shows the processing flow for estabhc~ VCH
S interconnection networks by combining the method for designing a VCH
interconnection network topology for each media in a particular order with the method for designing simultaneously;
FIG. 29 explains the design method illustrated in FIG 28;
FIG. 30 shows another method for establishing VCfI interconnection networks by combining the method for designing a VCx interconnection network topology for each media in a particular order with the method for simultaneously designing a plurality of VCH intercont~cGon network topologies; and FIG. 31 shows still another method for establishing VCH
I S interconnection networks.
DETAILED DESCRIPZTON OF THE
PRESENTLY PREFERRED EXEMPL Y EMBODTMENTS
FIG. 3 is a block diagram showing communication equipment according to a first embodiment of this invention. This equipment includes:
a plurality of virhsal chazmel handlers VCHl-VCHS which .perform switching in units of vCs that logically coy subscribers; VPH1-VPH3 which switch the routes of these VCs in units of VPs which accommodate a plurality of VCs; a transmission link network formed from transmission Iinks 31--37 which physically connect VCHI-VCHS and VPHl-VPH3; and an operation system 4pS 27 which establishes, on this transmission link network, a VP network 25 formed from VPs that logically connect VCHl-VCHS, and also establishes, on this VP network, a VC network 76 formed from VCs. OpS can establish a VP between any two VCHs that dots not pass through any other VCH. It can also simultaneously establish a plurality of VC networks for different media or services on a single physical _g_ _ 2~~~~.~~
network by establishing, on the VP network, a differcnt VCH
interconnection network for each mcdia ~ service consisting of information to be transferred, and by using each VCH interconnection network to form a VC network corresponding to that media or service.
VCHI-YCHS are respectivcly arranged in various regions, and any given VCHs can be cvmnectcd directly via the VP network. Although that are circumstances in which a connection betvveen VCHs will involve passing through one or more of VPHI-VPH3, VPIiI-VPH3 do not need to be switched in VC units and generally operate faster and with simpler hardware than VCHl-VCHS. Accordingly, a YP network is which the routes of the paths are switched by VPH1 YPH3 is a fast, high-quality network with little cell loss.
FIGS. 4(a~4(c) show examples, realizable by means of this invention, of logical VCH interconnection network topologies for various services. In the examples s)wwn hcre, the VCH intercout~ction netvuorks are all formod. fiom VPs. Category 1 shown in FIG. 4(a) is a star topology as typified by telephony; Category 2 shown in FIG. 4(b) is a loop topology as typificxi by electronic mail and data transfer systems; while Category 3 shaven in FIG. 4(c) is a mesh topology as utilized for on-lint banking, etc., wherc cspocially high reliability is requircd.
FIGS. 5(a~-5(d) show the configurations which realize the plurality of VCH interconnection ueiwork topobgics shown in FIGS. 4(a~4(c) on a singlc physical network. FIG. 5(a) shows the capon of the physical network, while FIGS. 5(b)-5(d) show the respective VCH interconnection network topologies. In Category 1, VCHS is taken as a transit VCH and VCHI-YCH4 are taken as total VCHs. VCHS is connected to VGHl-VCH4 by virtual paths VPI-1 to VPi-4, and alI traffic among VCHI VCfI4 can be connected via VCHS. In Category 2, which is completely independent from the VPs of Category 1, the loop topology is realized by means of virtual paths VPZ-1 to VP2-5. 1n the case of Category 3, the mesh topology is realized by moans of viraial paths VP3-1 to VP3-10, independently of Categories 1 and 2. VCI-Is can thus be connected using ~~~7~i VPs in such manner that the desired topologies are obtained. In these topologies, VCHS (by way of example) functions as a transit VCH in Category 1; as an add-drop VCH in Category 2; and as a local VCH in rY 3. -S FIG. 6 shows an example of the configuration of a VCH boa. This VCH bas a service distribution part 61 which distributes cells with the same destination received on line 56. It also has, for each category of topology, a shaper 62 which temporarily stores distributed cells and then sends them to a respective separaroc VP via a transmission line ~8. Each VCH assigns a VP to tech service and performs at the speed reqvmd for each service category. A VP is established for each destinarion: i.e., for tech destination VCH, and for each service.
FIG. 7 shows another example of the configuration of a VCH 60b.
This VCH has, in addition to the configaration shown in FIG. 6, a read control part 63 which controls the rtading of shapers 62. This read control part C3 performs read control in acxordanve with the priority or non priority of each service category with the same desatnation. For example, when telephone tra~'te, which is real-time in nature, has in~ased, read control part 63 increases the bandwidth of the corresponding VP and makes the traffic of other categories wait, the other categories being such that waiting is permissible. The manner in which VPs are disposed on a transmission link is shown in FIG. 8_ Priority control of this sort is priority control among VPs with the same destination, and does not aged rcsaurce rnanage~meat in the VP
network. - It enables the quality of service (delay and cell loss) of a plurality of media or services to be controlled e~ciently at a switch.
In the foregoing embodiment, an explanation was given of the case where an individual VP is assigned to each media, i.e., to each service category. However, a plurality of media with the same destination can also be assigned to the same VP_ A VCH 60c configuration in such a case is shown in FIG. 9.
_ lfl _ VCH 60G shown in FIG. 9 has: a shaper 62-1 for each category; VP
shaper 62-2 for tech destination; and read control part 63 which controls the reading of shapers 62-1. For each category, a shaper 62-1 controls the quality of service for that category. VP shaper 62-2 sends cells that have bees read from a shaper 62-1 for tech category to a common VP. In this case, Category 1 may, far example, be assigned to virtual channel identifiers ~L.~~ n I n n I ~n~ ~e Category 2- may be assigned to VCIS "1 O I "-"200".
This results in one VP per destination at the transmission lick level, which means that the number of VPs can be decreased and the communication network can be realized with smaller YPHs.
In the embodiments explained above, an ideal ~twork topology is realized for each service or media in a single physical network. Because the quality of service can be controlled as occasion demands for each service, multimedia can lx handled economically and flexibly.
I S Next, explanations will be gives of pmcedurcs for establishing the VCH interconnection networks dcscn'bed about.
FIG. 10 shows, as an example of a meths for establishing a VCH
interconnection netv~rork utilizing a known method for establishing a YC
network. In order to establish a VCH interconnection network, the initial values for a trial VCH interconnection network topology are first input in step 100. The quality required by the users in this example includes a relatively high degree of reliability. As an example, the network's reliability will be considered as the recovery ratio of working YPs when a fault has occurred. In an ATM network, the reliability required for tech media is different. The back up VP capacity is assigned so as to guarantee the highest required reliability, i.e., tta: recovery ratio of working VPs when a fault has occurred. The back-up VP capacity is set within the range of the given trial VCH int~nnection network topology. In step IOZ, working VP ratio is defined as:

total ~rldng VP capac=ty ( total working YP capacity + t~ntal back-up Vp capacity) The working VP ratio is thus obtained from the working VP capacity and the back-up VP capacity, which have boezi set so as to guarantee the -required reliability. It is then decided whether the working VP ratio is at a maadmum in step 103. If it is not at a maximum, the trial VCH
interconncction network topology is changed in step 104 and the wo~ang VP ratio is obtained again in the same manner. If the working VP ratio is a maximum, the VCH interconnecttion network topology will be an optimum network topology (see step 10~ and the design nnetbod is finished.
However, in an ATM network in which a plurality of mcdia coexist, the method illustrated in FIG. I0 establishes a VCH interconnection network suited to the media which reqtureg the highest reliability, which means that excessive network resources are given to media which do not require such reliability. In addition, VCs that handle burst type tragic and VCs that handle non-burst tragic are multiplexed onto the same VPs, and if the traffic type (i.e., whether in the farm of bursts or not), VC number, r cell loss probability, etc., of VCs established between subscribers using this VCH interconnection network are not reflected in the network resource efficiency, it will be impossible to make an accurate evaluation of the network resource eff raeacy.
FIG. I1 shows a utethod fox obtaining a different optimum VCH
interconnection network for each media, thereby overcoming these defects.
In this method, a different VCH intenonnection network is configured in accordance with the required Quality, traffic type and traffic demand of each media, so that network resource e~ciency is maximized. In establishing a VCH interconnecrioa network for each media, network resource efficiency is calculated not just from the working VP ratio, but also from the VP use efficiency, which has been calculated from the traffic type (whether in the form of bursts or nvt), traffic demand (VC number) and required cell loss probabiiity'(required quality) of VCs established between subscribers on the basis of the VCH interconnection network. Network resource efficiency thus calculated is used to evaluate the Lion, and the trial VCH
interconnoction network topology is c~nged on this basis. The mettrod illustrated in FIG. 11 can be carried oot by a computer programmed to implement the processes illuin FIG. 11. Such a computer would include an input device for inputting initial values, a Cerrtral Processing Unit (CPU) and an output device, such as a display.
This method will now be cxplsinod in greatea detail below. In the optimization of a VCH intcroonnoetion network for a given media, the initial values for a trial VCfi intea~con network topology arc first input in step 110, and VC routes are ~ablished for this trial VCH
inter~nnection network topology. VC mutes may be considered as taking the shortest path between origin and destination VCfIs. VC mutes pertaining to the same media and passing beiwocn two VCHs may be multiplexed onto the same VP. The VP capacity inquired for this trial connection is calculated in steps 112-1 i4 from the traffic type (whether in t-he form of bursts yr net), VC numbs and cell loss probability of the VCs;
and the VP use efficiency is found at the same time. The VC member ix _. equivalent to the tragic demand of the media, while cell loss probability is part of the quality required of the modes. The use efficiency of any given VP is defined as:
sum of mean rates of VCs accommodated in given VP
reguired YPrapacity When considered in terms of the whle network, the overall VP use efficiency will be:
(s~cm of mean rates of YCs accommodated in a YP) requirrd YP capacity where E indicates the sum total for all the VPs. When the reliability required of a media is considered as the recovery ratio in the event of a fault in a VP, the back up VP c~pis set so as to be sufficient to guarantee the required recovery ratio in an alternative path, in order to achieve recovery between faulty VP terminals. An alternative path may, for example, be considered as btit~g the sbortest path within the Limits of the given VCH interconnection fork topology.
The working VP ratio is defined as:
total workia$ VP capacity total working Yl' capacity + focal back-up YP capacity The overall network resource efficiency is defined as:
working YP ratio x VP ttse e,,~ency which expresses the state of utilization of network resources. In step 1 I6, it is decided whether this network resotu~ce efficiency is at a maximum, and if it is not at a maximum, the trial. VCH interconnection network topology is changed in step I I 8 and the same process is repeated rf the network resource e~eieney is at a nnaximum, the given VCH interconnection network topology becomes tl~ optima network topology for that media.
See step 120. Optimization of a VCH interconnection network for a new media is begun if in step 122, it is d~ermined that optimization is not complete for all media and the same process is performed until it is determined in step 122 that optimi~on of VCH interconnection networks is completed for all media.
Thus, because a different VCH i~crconnection network for each media is optimized, the VCH interconnection networks are not all designed to suit the media that requires the higbest quality (reliability and cell loss probability), and excess network resources are not given to media that do not require such high quality. Moreover, network resource efficiency is calculated not just from the working VP ratio, but also from the VP use efficiency as ~~a~d from the traffic type (whether in the form of bursts or not), traffic demand (VC number) and required cell loss probability of VCs established between subscribers utilizing these VCH interconnection networks. As a result, VCH interconnection networks can be optimized in correspondence with required quality, traffic type and traffic demand, which differ for digerent media.
FIG. 12 and FIG. I3 show examples of how a trial VCH
interconnection network topology is formed in step 110 of FIG. 11 so that ~~~7~~9 network resource e~ciency can be calculated. FIGS. 14(ar(d) stww trial network topologies when the number of disjoint routes D bas been set as a parameter. The number of disjoint routes D between VCHs is set as a parameter when the initial values for a trial VCH interconnection network topology are given or when a trial VCH interconnection netv~~ork topology is changed. The number of disjoint routes between VCHs is the uumbcrr of paths between origin and destination VCHs which do not pass through the same VCH or VP link In the present case, the number of disjoint routes will be considered to be the same between all VCHs.
In the mctlwd shown in FIG. 12, a given numberr of disjoint routes D is set as a parameter of trial VCH intcrcotuiection network topology in step 130, and satisfying this number of disjoint routes D will be tak~cn as a constraint. In step 132, initial values for a trial VCH interconnection network topology are formed so that the total length of all the VP Iiuks is minimized. The network resource efficiency of the trial VCH
interconnection network topology obtained is calculated using the method described above. When the trial VCH interconnection network topology is changed, the number of disjoint routes D is also changed, as shown in steps 140 and 142 in FiG_ 13.
By thus setting the number of disjoint routes D as a constraint, the member of possible trial VCH interconnection network topaiogies is re, so that optimization of VCH interconnection networks can be sped up.
FIG. 15 and FiG. 16 show the results of optimizations of VCH
interconnection network topologies in which the numtxs of disjoint routes D
is set as a parameter of VCH interconnection network topology, and network resource e~cieney is maximized. FIG. 15 shows the optimum number of disjoint routes for guaranteeing a 100'/° recovery ratio, while FIG. 16 shows the optimum number of disjoint routes for guaranteeing a 30% recovery ratio. It is assumed in both cases that VC mean raft is 150 kbps, burst length is 4 kbyte, and cell loss probability is 10~. The horizon-tal axis represents traffic demand T, while the vertical axis represents peak -i5-ratio = VC peak rate/VC mean rats. The number and arrangement of VCHs ~s as in the nctvvork models illushated in FIGS. 14(a)-14(d). Traffic demand T was defin~l as the member of VCs required between origin and destination VCHs, and it was assumed that the number of VCs required was the same all origin and destination VCHs.
As has been explained above, by optimizing a digere~ VCH
interconnection network for each media, the VCH inbemonnection networks are not all designed to suit the m;cdia that requires the highest quality (rcliabifrty and cell loss proba'bility), and excess network resot~ are not given to media that do not require such high quality, Moreover, network resource efficiency is calculated not just from the working VP ratio, but also fiom the VP use e~ciency as calculatod fmm the tragic type, tragic demand and required cell loss probability of VCs estabiishcd between sub-scribers utilizing these VCH interconnection networks. As a result, VCH
irrterconncction networks can be op~timmod in corndence with required quality, trafl~c type and tragic demand, which differ for different media. In addition, by setting the munber of disjoint mutes D as a constraint, the number of possible trial VCH interconnection network topologies is reduced, so that optinnization of VCH interconnection nctvvorks can be speeded up.
Thus far, explanations have hair given of methods for maximizing netwrork resource efl~ciency in terms of VPs. However, it is also feasible to design VCH interconnection networks so as to minimize the network cost of all elements that are required for establishing VCH interconnection networks, l.c., of all elements, including VCHs and VCH interconnections, that are used in establishing individual VCs. Methods for accomplishing this will now be explained.
FTG. 17 shows the overall flow of a method for establishing a plurality of VCH interconnection netvwrks that minimize network cost.
FIG. 18 shows the design flow of each VCH intereom~ection network topology. FIG. 19 shows the design Sow of a VCH interconnection network topology for one given media. The processes illustrated in these Figures can be implemented on an appropriately Programmed computer.
A method for designing VCH interconnection network topologies for a plurality of media will be explained with reference to FIG. 17. First; the S capacity C (a vector quantity) of the ATM physical network is input to operation system OpS in step 200. This capacity C includes information regarding the capacity of each VCH and each VPH, the capacity of each transmission link, and so forth. When there are M media, the characteristics of each of these M media are input to the OpS. Specifically, items thus input include the required quality (call blocking probability, cell loss probability, delay), traffic characteristics (V$R or CBR), traffc demand (geographical distribution of traffic), and so forth. Next, the order of the media for which VCH intercomneetion network topologies are to be designed is determined is step 202. This order can be determined randomly.
Alvely, the order can be detesmincd in the probable order of decreasing proportion of overall network cost accounted for by the media is question. A media which Mmes m° in this order will hereinafter be termed "media m".
A VCH interconnection nchwork topology is then designed using a computer, for example, for each media in accordance with the determined order step 204. This design procedure will be explained with reference to FIG. 18. First, m is ~initiali~ed to "1" in step 208. Next, the capacity C of the physical network and the quality required for media m are set as eoa~traints, and a VCH interconnection network topology for media m is designed so that network cost of n~dia m is minimised in step 210. When a VCH interconnection network topology for media m is designed, capacity C,~ (a vector quantity) exhibited by media m in the ATM physical network can be obtained. This capacity C" includes the capacity of transmission links, VCHs and YPHs, ere. occupied by media m. Next, in step 212, Cm is subtracted from capacity C of the physical network and this is taken as the new C. 'this altered C bocomes a constraint of the physical network when a YCH interconnection network topology is designed for the next media - m--After m has been increased by 1 in step 214, if m is less than or equal to the number of media M as determined in step 216, that VCH
interconnection network topology is designed. The remaining capacity C of the physical network (i.e., excluding use of previously established media) is set as a constraint, and a VCH intercozmection network topology is designed for the next media m in the order.
If this processing is repeated until m becomes larget.than M, the design of all the VCH interconnection network topologies will be finished.
This means that VCH interconnection network to~logies have been respectively obtained for all the M media, as shown in FIG. 17. YCH
intcr~onnection networks are then established on the VP network in accord ance with these topologies in step 206.
An example of a mahod for designing a VCH inbGroovn~ectioa network topology for media m in step 210 of FIG. 18 will be explained 1 S with reference to FIG. 19. First, the initial values for the trial VCH
nneetion network topology ~" for media m are input in step 220. A
VCH imerconncction network topology can be expres~d, for example, by an adjacency matrix A, wberc A is an NxN matrix and N is the number of YCHs. The (t, j) element of A, ay, is defined as a,~=1 if there is a VP
connection from VCHi to VCHj, and as a~ if there is no VP conncxtion from VGHi to VCHj. Adjacency matrix A is generally taken as a variable when a VCI-I interconnection network is established, and the VCH
interconnection network topology is determined on this basis. The minimum network cost Z"m, is initialized as infinity.
Next, in step 222, VF routes and capacities and VC routes that will guarantee tbc quality required of media m (e.g., call blocking probability, cell loss probability, and delay) and result in lower network cost are obtained, and network cost Z for media m is calculated. It is assumed that a VP is established only when a VP connection. between VCHs has been established in the aforementioned VCH interconnection network topology A~, (i.e., only when a;~~l). It is also assumed that the system in use is the one where YCs belonging to different media are accommodated in separate VPs. Accordingly, when considering the accommodation of VCs pertaining to a given media in a VP, no account hgs to be taken of VCs of other media accomm~ted in the same VP. There may be one or more than one VP
between VCHs. The VC routes are selects within the limits of the VP
connections represented by the given YCH interconnection network topology. The network cost Z for media m may comprise the sum of value Zl, which is calculated from the quantity of facilities used in the ATM
physical network for media m, and a petlalty function Zi. In other words, Z=Z,+ZZ.
Speei~cally. the "guantity of facilities used in the ATM physical network for media m" mesas the queu~ity of fa~dlities used for transmission Iinks, YCHs and VPHs. The penalty function is formulated so that it is "0"
when the constraints on physical network capacity and require quality, are met, and so that its value ZZ incrcas~es is con~espondence with the degree to which these constraints are not met. Although the value of ZZ may become larger than "0" dunng the process whereby the design Of VCH
interconnection network topology is Optimized, Z.~ has to become "0" at the point in time when the optimization pancess is f nishod.
Next, ii is decided in step 224 wbethcr the value of Z obtained in the manner described above is smalltr than Z",~. If it is, the value of Z is substituted for Z,";", and Am is substituted for A,~ in step 226. Froces~sing then proceeds to step 228. If Z is not smaller than Z~", processing proceeds to step 228 without changing the values of Z"~, and Ap,~ In step 228, it is decided whether all possible VCH interconnection network topologies A.m have been tried. If so, the design of a VCH interconnection network topology for media m is terminated and processing ret<uns to the steps , shown in FIG. 18. However, if this is not so, VCH interconnection network topology A", is changed in step 230 and the processing described -above is repeated.
By designing VCH interconnection network topologies after determining the order of the mtslia, YCH interconnection networks that are suited to the characteristics of the media can be est~ablist~ed and efficient use can be made of network resources.
FIG. 20 is a flowchart capable of being computer implemented showing a method for designing a VCH irnerconnection network topolegy for each media after setting a design priority order for the YCH
interconnection nehvorks in the order of decreasing network cost.
The steps beginning with the input of ATM physical network capacity C (step 200) to obtaining a VCH interconnection network topology for each media (step 20G) are the same as in the method illusin FIG..
17. At this point, the network cost of each VCH interconnection network and t3ic overall network cost are known. The media are, therefore, arranged in the order of ding proportion of overall network cost accounted for by the YCH inbercomnec,~tion network cost of each media is step 230. It is then decided in step 232 whether the order has changed as a result of this rearrangement. If it has, the design of a VCH interconnection netwark topology for each media is rid. If the order has not changed, this means that the order of the media has converged and that the desired VCH
intercomteetion network topology for each media has been obtained.
liocause the method illustrated in FTG. 20 arranges the design order in the order of decreasing proportion of overall network cost accounted for by tbc VCH interconnection network cost of each media, and designs a VCH interconnection network to~logy for each media in accordance with this order, it can give priority to establishing VCH interconnection networks for media that have a greater effect on overall network cost. 'this method ZS can therefore reduce overall network cost and establish YCH interconnection networks that are suited to media characteristics, thereby making e~cient use of network resources.
FIG. 21 is a flowchart capable of being computer-implemented showing another method for designing YCH interconnection network topologies for each media. This method differs from the method shown in FIG. 20 in that a limit is imposed on the number of times the order can be changed. Namely, the number of times that VCH interconnection networl~

topologies are designed in a particular order is restricted beforehaad to K
After an initial order has been set in steps 200 and 202, a count is initialized to k=1 in step 240, whercupon the value of k is increased by 1 in stop 242 each timc the order i$ restranged. It is then derided whether ~k.
I~ in step 244 this is so, the processing shifts to the decision regarding whether the order has chaaged. If not, it is decided t3~at the design of a VCH interooimection nto~logy for each media is finished and processing is tem~ated.
The number of times that the design of a VCH interconncction network topology far each modia is rcpeated is thus predetermined, and even if the order has changed, if this munber of times is eroded, it is decidcd that the desiga of a VCH interconnection network topology for tech media is f nished, whereupon processiwg is tezminatod. Bccause a restriction is imposed on the calculation time spcnt designing the VCH
interconnection network topologies, the cammuiricalion network can be established snore efficiently.
FIG. 22 shows a method used for step Z10 of FIG. 18 for designing a VCH interconnection t~tvuvrk topology for mcdia m which diffcrs from the design method illustrated is FIG. 19. According to this method, an z0 optim l solution for a VCH interconnection network is found using I trial VCH interconnection network topologies. In the subsequent explanation, the subscript will be omitted from the symbol A~, which indicates the VCH
interconnection network topology for the m'~ m~ia.
First, the number of repetitions s of the trial VCH interconnection network topology is sct to "1" and Z~ is set to infinity in step 300. Initial values for the trial VCH interconnection network topology A for media m are input as the initial topology in step 302. VP routes and capacities and VC routes that will guarantct the quality required of media m (e.g., call blocking probability, cell loss probability, and dclay) and rcsult in lower network cost are obtained, and network cost Z for media m are calculated in seep 302. The methods used for establishing VP routes and capacities and VC routes, as well as for calculating network cost Z, are the same as those _ 2-~~73.~~
explained with reference to FIG. 19. The value of Z obtained is substituted for Z"~. A VCH interconnection network topology can be expressed, for example, by an adjacency maW x A, where A is an NxN fix, and N is the number of VCHs. The (i,j) element of A, a~, is defined as a~-=I if there is a VP connection from VCI~ to VCHj, and as a~~ if there is no VP.
connection from VCHi to VCHj. Adjacency matrix A is generally taken as a variable when a VCH interconnection network is established, and the VCH intcrcvnnection network topology is detcr~nined on this basis.
BY ~Y ~ $ VP connections between VCHs, the method illustrated in FIG. 22 optimizes VCH interconnection network topology so that network cost for media m is minimized.
To achieve this, first an order is determined in step 304 for pairs of VCHs between which VP connexions arc to be added or deleted. This order can be determined for evcample by arranging the VCH pairs randomly.
I S VCH pairs (ij) are then selxted in accordance with the order thereby determined, and the value of a~ is invcr~d. In other words, if a,3=I it is changed to "0", while if a~.~ it is changed to "1 ". VP routs and opacities and VC routes that will guamntx the quality required of media m (e.g.. call blocking probability, cell loss probability, and delay) and result in lower network cost are obtained for the altered VCH interconnection network topology, and network cost Z for media m is calculated in step 306. It is then decided in step 308 whether Z.~,>Z. If so, the value of Z is substitntied for Z""~ in step 310. This signifies that it is possible that the VCH
interconnection network topology at that point in time will become an optimal solution. If Z,"", is not greater than Z, the value of a~ is restored to its previous value in step 312. It is then decided in step 314 whether the foregoing processing has been performed for all VCH pairs. rf it has, processing proceeds to step 316. If not, the processing returns to the calculation of network cost Z step 306, and the same processing is performed on the next VCH pair (i~j) in the order. When the foregoing processing has been performed on all VCH pairs, it is decided in step 316 whether the VCH interconnection network topology has converged.

e~ ~ c~
Convergence of a VCH intercom~ection network topology means that there is no change, at the point in time of this convergence decision, in the previous VCH interconnection network topology for which the order of the VCId pairs is determined. If it is decided that the VCH intcrconion network topology has not converged, processing returns to step 304. If it is decided that the VCH interconnection network topology bas comergcd; it is then decided in step 318 whether Z~Z";~. If so, Z~ and A,~ are updated with the values of Z,~;" and A, rvely in step 320. The value of i is increased by 1 in step 322 and it is decided whether Izi in step 324. If so, processing returns to step 302. If it is not so, this processing is terminated.
By adding or deleting YP connections for T trial VCH
interconnection network topologies, the method illustrated in FIG. 22 enables an optimal solution for VCH ittterconncction nctvvork tapology to be obtained e~ciently.
FIG. 23 shows still another method used in step 210 of FIG. 18 for designing VCH intemonneetion netvvodc topologies. This method diflbrs from the method illustzated in FIG. 22 in that it performs a decision and related pmc~ssing regarding a delay condition.
Namely, after network cost Z has been calculated in step 304, if is decided whether a delay condition is satisfied in step 340. In the method shown in FIG. 19 or the method shown in FIG. 2Z, failure to meet the delay conditions was reflected in the value of a penalty function Z2. In the method shown in FIG. 23, in addition to providing a penalty function, delay is reduced by establishing a VP beh~en VCHs which are not directly connected by a VP.
In other words, when the delay condition is not satisfied, it is decided in step 342 whether a~.=1. If a;~ 1, the value Z",m is updated to Z in step 344, irrespective of the value of the network cost. If ay is not equal to 1, the value of a,~ is restored and a~ is set to 1 in step 346. If ay=1, it signif es that a VP has been established between VCHi and VCHj, and became there is no intermediary VCH, delay can be reduced.

_ 2~ 73~
By incorporating a decision and related processing regarding a delay condition into the algorithm for optimizing VCH interconnection network topology, an optimal solution for VCH interconnection network topology can be obtained efficie~ly. .
Explanations have been given, with r~rference to FIG. 17-FIG. 23, of methods for designing a plu~radity of VCH intec~conneetion network topologies in an order that wih minimize network cost. As opposed to. this, .
it is also feasible to design a plurality of VCH invcrcouncction network topologies simultaneously. Such metlwds will be explained below. .
I0 FIG. 24 shows a method for designing VCH interconnection network topologies simultaneously for a plurality of media. This method is also capable of being implemented on a computer. In this method, capacity C (a vector quantity) of the ATM physical network is first input to operation system tJpS in step 400. Capacity C includes information regarding the 15 capacity of each VCI'rI, the opacity of each transnais.~tion link, and so forth.
When there arc M media, the characteristics of each of these M media are input to the OpS. Specifically, items thus input include the required quality (call blocking probability, cell loss probability, delay), traffic characteristics (VBR or CBR), traffic demand (geographical distribution of trat~c), and so 20 forth.
Next, rather than designing the VCH interconnection network topologies for each of the M media in a particular order, they are designed simultaneously in step 402. An example of this design method in step 402 will be explained with refer~ce to FIG. 25.
25 First, the number of repetitions i of the trial VCH interconnection network topology is set to "1 ", and ~~ is set to infinity in step 410. For ah m (m--1, 2, ..., M), initial values for the trial VCH interconnection network topology A for media m are input as the initial topology in step 412. Next, for all m, VP routes and capacities and VC routes are also obtained in step 30 412 that will guarantee the quality required of media m (e.g., call blocking probability, cell loss probability, and delay) and result in a smaller sum total of the VCH interconnection network costs arising from each media, i.e., in ~.~5~3.~~
lower overall network cost. Overall netvwork cost z is then calculated. VP
routes and capacities and YC routes, and overall network cost Z, are obtained by applying the methods explained with reference to FIG. 19 to all media is succession. The value of z obtained is substituted for Z",;". ~
VCH interconnection network topology can be expressed by an NxN matrix (where N is the number of VCHs). In the present case, a VCH
interconnection network topology for media m will be represented by adjaccacy matrix Am. The (i,y~ element of Am, ans., is defined as amp=1 if there is a VP co~ec~tion from VCHi to VCHj, and as am~-0 if there is no Vi' connection from VCTTt to VCTTj.
By rcpeaGedly adding aad deleting YP connections between VCHs in VC networks for all media, the method illustrated in FIG. 25 optimizes VCH imercanaection network topologies and establishes all VCH
urterconnoction networks so that the overall network cost resulting from ail media is mi~nized. After the calculation of network cost Z, the order of the VCH pairs between which VP con~neQions will be added and deleted is dctermia~ for all media in step 414. An example of the method used in step 4I4 for determining this order will be explained with reference to FIG.
26 and FIG. 27.
First; as Procedure 1 in step 440, the order of VGH pairs is deter-mined for each media. An -example far the case where there are three media is shown in FIG. 27. Next, as Pmccdure 2 in step 442, the order in which the media arc distributed is determined. In the example given in FIG. 27, this order is: media 2, media l, media 3, media 2,.... When there are three media, the order of the f rst three YCH pairs may be determined by selecting from the media in random order. This random order is then repeated every three pairs. Next, as Procedure 3 in step 444, VCH pairs amp are assigned in accordance with the order determined, by procedures 1 and 2. An example of this is shovm in FIG. 27. VCH pairs am;~ arc .
selected in turn in accordance with the order determined in Procedure 3, and the value of amn is inverted in step 416. In other words, if amv-=1, it is changed to "0", and if 2~=0, it is changed to "1". VP routes and capacities and VC routes that will guarat~ee the quality required of media m (e.g., call blocking probability, cell loss probability, and delay) and result in lower network cost arc obtained for the attered VCH interconnection network topology for media m; a VCH interconnection network configuration is S obtained for media m; and overall network cost Z is calculated in step 416.
At this point, VCH interconnection networks for media other than media m do not change. It is then decided in step 418 whether Z",~Z. If so, the value of Z is substituted for Z~;" in step 420. This signifies that it is possible that the YCH intc~mection network topology at that point in time will become an optimal solution. If Z"~" is pot greater than Z, the value of atn~ is restore to its value prior to the change in step 422. It is then decided whether the foregoing processing has been performed for VCFi pairs of all the media in step 424. If it has, processing proceeds to step 426_ If not, the same processing is performed on the next VCH pair (i~j) of media m in the sequence. Nexxt, for all m, it is decided in step 426 whether the VCH intcrco~eetion network topology Am for media m bas converged Convergence of a YCH interoom~ion network topology means that there is no change, at the point in time of this decision, in VCH interconnection network topologies Am (for all m, where m=1, 2,..., M) prior to starting processing step 414 for determining tlbe order of the VCH pairs. If it is decided that there has been convergencx, the processing proceeds to step 428. If it is decided that convergence has not occurred, the processing rchuns to determining the order of YCH pairs between which VP
connections will be adds or deleted, and the same processing as described above is repeated. In the subsequent processing step 428, it is decided whether Z~Z,~;". If so, Z~ and A~ arc updated with the values of Z,a;"
and Am, respectively in sbcp 430. The value of l is increased by I in step 432 and it is decided in step 434 whether I?i for a predetermined I. if so, processing returns to step 412 and the operations described above are 3U repeated. In the present case, I is the number of times that a trial VCH
interconnection network topology is formed If it is dccidod that I is not greater than or equal to l, this processing step 402 illustrated in FIG. 25 is terminated As a result of this processing, a VCH intercozmectian network for each media can be established for all the media Because a VCI-I interconntctian network topology is designed simultaneously for each media, VCH interconnection networks suited to the characteristics of the respective media can be established while taking into consideration the effects which the respective VCH interconnectipn networks have on each other. As a result, e~ciem use can be made of network resources.
FIG. 28 and FICI. 29 show a mettwd for establishing VC~i interconnection networks by combining the method for designing VCH
interconnection network topologies far respective media in a particular order with the method for designing them simultancoasly. FIG. 28 shows the processing flow, while FIG. 29 explains the design method.
In the method illustrated in FIG Z8, the capacity C (a vector quantity} of the ATM physical is first input to operation system QpS in step 500. Cdr G includes information regarding the c~City of each ~VCH and each VPH, the city of each transmission link, and so forth. When there are M media, tl~ ch2tractaistics of each of these M
media are input to the QpS. Specifically, items thus input include the required quality (call blocking probability, cell loss probability, delay), traffic characteristics (VBIt or CBR), tragic demand (geographical distribution of tragic), and so forth.
Next, M' media are selected fi~om the M media in stop 502. Because priority is given to the design of VCH interconnection network topologies for the M' media, the selection of the M' media may involve, by way of example, selecting media which may have a large influence on network cost. For the M' media that are selected, the media order in which networks are to be established is determined in step 504, a VCH interconnection network topology is designed for each media in this order in step 506, and VCH interconnection networks are established in step 508.
Thereafter, respective YCH interconnection network topologies for the remaining M-M' media are designed in step S 10. At this point, the 2~~731~
capacity of the given physical network is the remaining capacity of the part occupied by the VCH interconnection nclworks of the M modia which have been established as described above. When this processing is completed, a VCH interconnection network for each nlcdia is established for the .
remaining M-M' media in step 512.
By. establishing VCH interoonn~tion networks by combining the method for designing VCH interconnection network topologies for rcxpecfiive media in a particular order with the method for deslgtuag a plurality of VCH imerconneetion topologies simultaneously, media with a high priority ranking eau obtain prefere»tial assignment of network resources, while the remaining media eaa obtain effident assignment of network resources that involves taking into consideration the mutual influence of each VCH unterconacction network.
FIG. 30 shows another method for establishing VCH interconnection networks by combining the method for designing VCH interconnection network topologies for respc~ve malia in a particular order with the method for designing a pltnality of VCH interconnection network to~logics simultaneously. This method differs fra~n the method illustrated in FIG. 28 and FIG. 29 in that M' media for preferential netvv~ork~establishment are selected from the M media, and VCH i~erconnection network topologies arc designed for these M' media in the order of decreasing proportion of overall network cost accounted for by the VCH interconnection network cost of a given media. In otter words, this method differs from the method illustrated in FIG. 28 and FIG. 29 in that the method illustrated in FIG. 20 is utilized.
In this method, media with a high priority ranking can obtain preferential assignment of network resources, while the remaining media can obtain efficient assignmem of network resources that involves taking into consideration the mutual influence of tech VCH interconnection network.
In particular, because the media arc arranged in the order of decreasing proportion of overall network cost accounted for by their respective VCH
interconnection networks, and because VCH interconnection network _2g_ _ 2~ ~,~~~ ~
topologies for the rGxpective media arc designed in accordance with this order, media which have a large ittifluence on overall network cost are e~~shed fast- Accordingly, overall network cost can be reduced; VCH
interconrsection networks suibod to the characteristics of the respective media can be establisind; and ei~cicat use can be made of network resources.
In FIG. 31, in similar manner to the method illustrated in FIG. 28 and FIG. 29 and to the method illustrated in FTG. 30, YCH interconnection networks are established by coznb'ming the method for designing VCH
interconnection network topologies for respective media in a particular order with the method for designing a plurality of VCH interconnection network topalogies simultaneously. However, the method shown in FIG. 31 differs in that the number of times a VCH interconnection network topology is designed for tech media is set in advance: in other words, it differs in that ~e shod illustrated in FIG. 2I is ettilized.
Namely, when estabii~ing networks for M' media, the numbac of times a VCH interconnection network topology is designed for each media is set in advance in step 600, and even though the order has changed, if this number of lames will be exceeded, it is decided that a VCH interconnection network topolagy has been designed for each of the M' media, and the processing shifts to the method for designing the remaining 1VI IVI' VCH
interconnection network topologies simultaneously. Accordingly, because the time spent in calculating and thereby establishing a VCH interconnection network is restricted, communication networks can be established ei$ciently.
As explained above, this invention establishes a different VCH
25. interconnection network for each media and establishes VC networks on the basis of these VCH imercomiection networks. Consequently, the VC
network is not designed to guarantee the highest quality (call bl~king probability, cell loss probability and delay) requited, and axces.sive netw~nrk resources are not given to media which do not require such high quality.
In addition, network cost can be reduced by setting the capacity of the physical network (i.e_, the capacity of transmission links, VCHs and VPHs) and the qualities required for a plurality of media as constraints, and -~9-by establishing a respective VCH interconnection network for each media on the basis of the media ch~:r~cteristics (traffic type, traffic demand) that have been input, so that overall network cost is minimized when the VCH
interconnection network topologies constituted by the plurality of media are combi~d.
As has been explained above, this invention establishes a did VCH interconnection network for tech media in coxrespondcacc with the Q~tY required by the media or the media traffic characteristics. It can therefore utilize this to establish VC networks that connect subscxibers.
Moreover, when a plurality of media coexist in a conununication n~,ork, it can design the network on the basis of minimum ra;twork cost. The constituent elements of the communication network can therefore be utilized effectively. Furthermore, communication networks in which a plurality of media coexist can be designed quickly by means of this invention.
While the present invention has been desen'bed in conjunction with what is prc~ntly considered to be the most practical and prcfcnrcd embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but on the co~rary, is intended to cover various madifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims (15)

THE EMBODIMENTS OF THE INVENTION IN WHICH A EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A method of designing an ATM network comprising the steps of:
physically connecting, by means of a transmission link network, a plurality of virtual channel handlers having subscribers connected thereto, said plurality of virtual channel handlers switching data to be transmitted over said transmission link network in units of virtual channels, and a virtual path handler which changes information transfer routes between virtual channel handlers in units of virtual paths that accommodate a plurality of said virtual channels;
establishing a virtual path network on said transmission link network, said virtual path network being formed from virtual paths which logically connect said plurality of virtual channel handlers;
establishing a virtual channel network on said virtual path network, said virtual channel network being formed from virtual channels which logically connect subscribers on said transmission link network;
establishing a different virtual channel handler interconnection network on said transmission link network for each media having information to be transferred on said transmission link network; and forming said virtual channel network corresponding to a given media by means of each virtual channel handler interconnection network, so that a plurality of independent virtual channel networks for different media are established simultaneously on a single physical network.

2. A communication network design method as set forth in claim 1, wherein said virtual path is established in said step of establishing said virtual path network such that said virtual path includes two (2) virtual channel handlers one at each termination thereof.

3. A communication network design method as set forth in claim 1 or 2, wherein said virtual channel handler interconnection network is established by assigning different virtual paths to each media having information to be transferred.

4. A communication network design method as set forth in claim 1 or 2, wherein said virtual channel handler interconnection network is established by allowing a plurality of media with a same destination virtual channel handler to be assigned a single virtual path.

5. A communication network design method as set forth in claim 4, further comprising a step of controlling virtual channels with a same destination in accordance with a priority ranking of said media having information to be transferred.

6. A communication network design method as set forth in claim 1 or 2, wherein network topologies of a plurality of virtual channel handler interconnection networks for different kinds of media are designed simultaneously after taking their overall optimization into consideration in said step of establishing said virtual channel handler interconnection network.

7. A communication network design method as set forth in claim 6, wherein said overall optimization of said communication network topology is obtained by calculating a network resource efficiency from a working virtual path ratio, which is defined as a ratio of working virtual path capacity to total capacity of working and back-up virtual paths and from a virtual path use efficiency, which is calculated from a traffic type, traffic demand and required cell loss probability of virtual channels.

8. A communication network design method as set forth in claim 7, wherein said network resource efficiency is obtained by:
setting a number of disjoint routes D which corresponds to a number of paths between an origin and a destination virtual channel handler, said virtual path indicating only two virtual channel handlers, one at each termination thereof;
forming said trial network topology such that a total length of all virtual path links is minimized; and changing said number of disjoint routes D when said trial network topology is changed.

9. A communication network design method as set forth in claim 7, wherein connection or disconnection of virtual path connections between virtual channel handlers for establishing said virtual channel handler interconnection network is performed in parallel with establishing virtual channel handler interconnection networks for other media.

10. A communication network design method as set forth in claim 1, wherein a design priority order of said network topologies is set based on decreasing network cost of said media supported by its virtual channel handler interconnection networks.

11. A communication network design method as set forth in claim 1, further comprising the steps of:
calculating a proportion of overall network cost accounted for by said network cost of said virtual channel handler interconnection network for each media;
setting said design priority order in accordance with a result of said calculation; and repeating said design of network topologies in accordance with a new design priority order if said design priority order is different from a previously determined design priority.

12. A communication network design method as set forth in claim 11, wherein a limit is set on a number of repetitions in said repeating step.

13. A communication network design method as set forth in claim 7, wherein network topologies are designed after setting an upper limit for a number of virtual paths passed through from an origin virtual channel handler to a destination virtual channel handler.

14. A communication network design method as set forth in claim 1 or 2, further comprising the step of classifying said different media into a priority group and a non-priority group according to their network cost; and wherein said virtual channel interconnection network for said priority group established in said establishing a virtual channel handler interconnection step is performed in an order of decreasing network cost; and wherein said virtual channel handler interconnection network for said non-priority group is established simultaneously in a manner such that overall network cost is minimized.

15. Communication ATM network comprising:
a plurality of virtual channel handlers which perform switching in units of virtual channels;
at least one subscriber connected to each virtual channel handler;
a virtual path handler which changes over routes of data in units of virtual paths which accommodate a plurality of said virtual channels;
a transmission link network which physically connects said plurality of virtual channel handlers and said virtual path handlers; and a control means which establishes a virtual path network on said transmission link network, said virtual path network being formed from virtual paths which logically connect said plurality of virtual channel handlers, and which also establishes a virtual channel network on said virtual path network, said virtual channel network being formed from virtual channels which logically connect said subscribers on said transmission link network;
said control means including:
a virtual control handler interconnection network establishing means which establishes, on said virtual path network, a different virtual channel handler interconnection network for each media having information to be transferred; and a means for establishing virtual channel networks corresponding to said media by means of respective virtual channel handler interconnection networks, whereby different virtual channel networks for different media are established simultaneously on a single physical network.