JP3485948B2 - Network configuration method and device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a network for carrying a call (traffic) such as a telephone, that is, a subscriber line exchange for accommodating a user (user), and a line for communication via a relay exchange. It relates to a method of constructing a network.

[0002]

2. Description of the Related Art In the network configuration, the form of the network (network topology), the number of transit exchanges, the number of lines between exchanges, the network cost, etc. are obtained. Here, the network topology is a network configuration consisting of nodes and links connecting them, and is represented by a connection matrix between nodes. A node refers to a logical point that converges and relays traffic in a traffic network, and refers to an exchange (subscriber line exchange, transit exchange) in a line network.
Further, a link indicates a traffic link in a traffic network and a line group between exchanges in a line network.

This kind of network construction method is performed by using a computer, and has been conventionally as follows. The line network (network between exchanges) is directly configured (designed), and many complicated design conditions regarding design are set at the same time to proceed with the design. The network creation unit (which creates the network topology and the routing table) and the traffic assignment unit are created as a functionally integrated unit. Here, the routing table is a table showing which route a call takes to the node which is the final destination of the call.

[0006] For the route from the originating switch to the destination switch, a transit switch through which traffic passes is given as an input. The network form (network topology) connecting the exchanges and the call flow route (routing table) are directly input to the system directly. It only deals with small, single-structured networks.

[0005]

However, according to such a conventional method, the designer judges all the complicated design conditions of the network at the same time from the viewpoint of software configuration, and proceeds with the design. , It is difficult to make the optimum network configuration (design). Further, when the network configuration (the number of layers of the network, etc.) and the design algorithm change drastically, the entire design system (device) is often rebuilt. That is, the design system lacks efficiency and flexibility.

The network creating unit and the traffic allocating unit are not functionally clearly separated, and it is necessary to change both the network creating unit and the traffic allocating unit every time the network configuration or the design algorithm is changed. For each of the many routes from the originating switch to the destination switch, the transit switch through which the traffic between them passes is given as an input, so there is no flexibility in changing the route.

In creating the network topology and the routing table, the design system is manually input directly.
The operation for that is very large. Actual networks have different structures for each region, but these cannot be designed in a unified manner. For example, in one area, a call that communicates outside the area is connected to a network outside the area via a transit exchange in a specific building within the area. It is structured so that it can be connected to the outside network, and these cannot be designed in a unified manner.

In addition, even for a network of a nationwide scale, it is divided into individual design systems, so that the entire network cannot be designed centrally and efficiently. There are drawbacks such as. That is, when the network configuration or design algorithm changes, there are many parts to be changed, which cannot be flexibly dealt with, and a complicated and large-scale network like an actual network cannot be designed.

The present invention solves these drawbacks. The object of the present invention is to express a complicated network in a simple and easy-to-understand manner, to flexibly respond to changes in the network configuration and design algorithm, and to significantly reduce the data input operation to the system (device). A large-scale network such as a network is to provide a network configuration method that can be efficiently designed (configured) in a centralized manner.

[0010]

According to the present invention, a network is hierarchized into two, a traffic network and a circuit network,
Design is advanced from the traffic network, and the design result is automatically taken over to the network. -Traffic network This network focuses on the flow of traffic, and consists of a network topology consisting of traffic nodes and traffic links and a routing table that defines the flow of traffic for each traffic node.

A network topology having a circuit network switch as a node and a line group as a link, and a routing table set for each switch. For each of the traffic network and circuit network design,
Functionally clearly separate the network creation unit (which creates the network topology and routing table) from the traffic assignment unit,
The interface between them is represented in the general form of an internode connection matrix and a routing table.

A routing table is defined for each node. When creating a traffic network, the network topology and routing table are created automatically by specifying the traffic area (geographic range of traffic handled by one transit exchange), the number of network layers, and so on. Also, in the design of the circuit network, the traffic network design result is received, and it is automatically generated by designating the necessary design parameters.

In the traffic allocation unit of the circuit network design unit, when traffic is allocated to the network based on the inter-switch connection matrix and the routing table for each switch,
Apply a conditional depth-first search algorithm. The conditional depth-first search can be applied to a directed graph consisting of nodes and links. The algorithm is
When the traffic allocation is completed for all the links input to a node, the traffic allocation for that node is completed, and then the total amount of the input traffic is allocated to each link on the output side at the specified ratio. The feature of the processing is that the traffic allocation is completed by performing such allocation to all the nodes. According to this method, it is possible to significantly reduce the amount of calculation as compared with the case where traffic is individually assigned to all routes between the originator node and the destination node.

A network is represented as a network having a composite structure composed of a combination of one or several sub-networks and element networks constituting the sub-networks. The sub-networks are independent from each other in terms of traffic flow, and can be designed separately, with the subscriber line switch as a base point (the subscriber line switch can be shared between the sub-networks).

The element network is a network having a hierarchical structure, and one or several element networks are connected to form a sub-network. By doing so, it is possible to provide a software configuration capable of uniformly and efficiently designing a complicated and large-scale network such as an actual network.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows the target range of the network configuration (design) of this embodiment. In other words, in this embodiment, the network is the subscriber line exchange LS that accommodates users or more and includes the relay exchange.

In the present invention, the network is hierarchized. That is, for example, as shown in FIG. 2, the conventional network is hierarchized into two networks, that is, a network focused on the flow of traffic (traffic network) and a network connecting the exchanges (line network). The concept of the traffic network is shown, for example, in FIG. 3, where a traffic node (hereinafter referred to as TN) is provided on the subscriber line exchange LS.
And a three-level network with traffic links. T
N is a logical node for relaying traffic, and in the case of the third floor , there are lower traffic nodes LTN1 to LTN4 and upper traffic nodes HTN1 to HTN2. The traffic link indicates a logical route through which traffic flows.

In FIG. 3, the design target area is divided into four lower traffic areas LTA1 to LTA4, the subscriber line exchanges included therein belong to the respective lower traffic areas LTN, and the LTN is further higher traffic node HTN. Shows the situation that belongs to. HTN
The traffic area corresponding to the higher traffic area H
Called TA. Further, in FIG. 3, the section of LTN1 and LTN 3 is a direct link DL (without passing through the TN upper attributable, path for their communication traffic) is set. In addition, the subscriber line exchange is a level 0 node, L
The TN is called a level 1 node and the HTN is called a level 2 node.

FIG. 4 shows the concept of a circuit network corresponding to the traffic network of FIG. Based on the traffic volume at each TN obtained in the traffic network design, find the replacement model and the number of exchanges,
Links are connected between exchanges under specified conditions. The link between the exchanges is called a line group. Here, each T
The N exchanges are divided into two groups, and a line group is set for each exchange from one TN exchange to each of the other TN groups in consideration of reliability. For example, one group of LTN1 is HTN1
Link to each of the two groups, and to each of the two groups of LTN3 by a direct link, the LTN
The other one group is also link-connected to the two groups of HTN1 and is directly linked to the two groups of LTN3. The exchanges of each of the two groups will be housed in different buildings after the circuit network is designed.

With respect to the line group from the LTN to the subscriber line exchange, it is necessary to receive the call to all the subscriber line exchanges belonging to the LTN, so the line group is set for all the subscriber line exchanges. It An example of a network having a composite structure is shown in FIG. FIG. 5A shows a plurality of sub-networks (hereinafter,
It shows that it can be expressed by superimposing A and B (called sub NW). In this case, the subscriber's line switching system has each sub N
Can be connected to W. FIG. 5B shows two element networks having a hierarchical structure (hereinafter referred to as element NW).
It shows one sub-NW configured by network connection of A and B. Generally, a sub NW is one or several elements N
It is configured by connecting W. By expressing the network by such a combination, it becomes possible to design an actual network.

FIG. 6 schematically shows a network design device (system) according to an embodiment of the present invention. This system includes a traffic creating unit 11, a network designing unit 12,
It is composed of a database unit 13 and a network evaluation unit 14, and the network design unit 12 further includes a traffic network design unit 15
And the network design unit 16. The traffic creation unit 11 predicts the traffic in the design target area, and for the traffic of ordinary telephones, creates a traffic matrix showing the traffic exchange status between the subscriber line exchanges, and the traffic of various services such as international calls. ,
Create total outgoing and incoming call volume by service for each subscriber line switch. These are input to the network design unit 12.

The traffic network designing section 15 includes a network creating section 17
And a traffic allocation unit 18. The network creating unit 17 creates a traffic node connection matrix (hereinafter referred to as TN connection matrix) and a routing table for each TN. In the traffic allocation unit 18, the network creation unit 17
Traffic is assigned to the subscriber line exchange based on the TN connection matrix and the routing table created in step 1 and the call volume flowing to the TN and the traffic link is obtained. In this traffic allocation, since there is one route between the originating-side subscriber line exchange and the receiving-side subscriber line exchange, the next network design unit 16
It is easier than the traffic allocation of.

The line network designing unit 16 is similar to the traffic network designing unit 15 in that the network creating unit 19 and the traffic allocating unit 2
1, and the network creating unit 19 further includes a network topology creating unit 22 and a routing table creating unit 23. The network topology creation unit 22 obtains the exchange model and the number of exchanges corresponding to the call volume flowing through the TN, and inputs a link (line group) setting method based on the exchange model and the number of exchanges obtained from the traffic network design unit 15. By doing so, the inter-switch connection matrix is automatically created.

In the routing table creating section 23, the TN
A routing table for each exchange is created from the routing table for each exchange and the connection matrix between exchanges. The traffic allocator 27 allocates the traffic of the subscriber line switch based on the inter-switch connection matrix created by the network creator 19 and the routing table for each switch, and the traffic volume flowing through the line group between each switch and the switch. Is calculated, and the number of lines corresponding thereto is calculated.

The database unit 13 stores the design results of the traffic network design unit 15 and the circuit network design unit 16 as a database. The network evaluation unit 14 calculates a network cost, reliability, etc. in order to evaluate the designed network from various viewpoints. The processing procedure of the network creating unit (creating the network topology and the routing table) 17 of the traffic network designing unit 15 in FIG. 6 is shown in FIGS. 7 and 8, and an image of creating a traffic network, a service network in the traffic network (international call 9 and 10 show images of connection to a network such as a facsimile or a facsimile).

A description will be given below with reference to the flowcharts of FIGS. First, the design target area of the sub NW (the area where the subscriber line exchange exists) is input (design target area of the sub NW in FIG. 9) (S ₁ ). Next, the element NW in the sub NW
The target design area (design target area of the element NWa in FIG. 9) is input, and the number of layers of the element NW and the traffic area handled by the TN of each level (L in FIG. 9).
Input TA1 to LTA4) (S ₂ ).

Based on these data, the TN (including the subscriber line exchange) of each level of the element NW is the higher T
A membership relation table indicating whether or not it belongs to N and an inter-TN connection matrix indicating how each TN of the element NW are connected are created (S ₃ ). This inter-TN connection matrix represents a network in which all the top TNs are connected to a mesh and the TNs below the top TN are connected to a tree structure as shown in FIG.

Next, using the attribution table, the element N
W to create a routing table of (S _4). Further, the direct link setting candidate is calculated and displayed using the membership table and the inter-subscriber line exchange (LS) traffic matrix of the element NW created by the traffic creating section 11 (S ₅ ). The direct link setting candidate is selected when the call volume between the nodes is equal to or greater than the designated reference value.

[0029] than direct link setting candidate, selected direct link setting section (input) to (link connecting LTN1 and LTN3 in Figure 9 DL) (S _6), on this basis, between TN element NW connection matrix and to modify the routing table (S _7). Up to this point, the creation of one element NW is completed. If the one sub-NW are plural elements NW examines if with the process for all the elements NW (S _8),
If it remains, the process returns to step S ₂ and the above is repeated. In the example of FIG. 9, first, the process is performed for the element NWa, and then the same process is performed for the element NWb.

After creating the inter-TN connection matrix and the routing table for each element NW, information specifying the TN for connection to connect the elements NW to the network (LTN in FIG. 9).
4 and LTN) (S ₉ ), and based on that, the element NWs are network-connected by the link ML, and the sub-NW TN-to-TN connection matrix and the routing table reflecting this are created (S ₁₀ ).

The network created up to this point is for general telephone calls only.
Although it was targeted, services such as facsimile and international calls
Connect to various service networks so that calls can be handled
Therefore, a connection matrix between TNs is created as follows. First,
Gateway TN (GTN) for connecting to the service network (Fig. 1
If 0, enter GTN1 to GTN3), then G
Traffic area handled by TN (from GTA1 in Fig. 10)
Enter GTA3). In addition, the sub that connects with GTN
Representative TN of NW (in FIG. 10, representative TN1 to representative TN
3) is designated (S₁₁). Created based on these
Modify the created connection matrix between TN and routing table
(S ₁₂).

[0032] For all of the sub-NW in the case of sub-NW there are multiple checks whether finished the processing (S _13),
If there is any, the process returns to step S ₁ and the above is repeated. This completes the creation of the traffic network. FIG. 11B shows an TN connection matrix for the element NW shown in FIG. 11A as an example of the TN connection matrix created as described above. All the TNs are arranged in the first row and the first column of the table, respectively, and when there is a connection (traffic link) between the row TN and the column TN, 1 is set in the matrix element. For example, since the LTN1 and the HTN1 are connected (showing both the traffic flow in the direction of LTN1 → HTN1 and the traffic flow in the direction of HTN1 → LTN1), the row LTN1 is connected in the inter-TN connection matrix.
Is set to the intersecting matrix element of the column HTN1 and the intersecting matrix element of the row HTN1 and the column LTN1.
In addition, generally, in the connection matrix, the node indicated by the row represents the transmission side.

Element NW for element NT shown in FIG. 12A
12B shows an example of a membership table showing which TN of each level (including the subscriber line exchange LS) other than the highest level of TN belongs to which TN of the higher level. Element NW
Represents the membership relationship between TNs by using the property of having a tree structure. For example, level 1 node TN11
1 and TN112 belong to the level 2 node TN11, and TN11 further belongs to the level 3 node TN1.
Belong to.

Elements NW1 and NW2 as shown in FIG. 13A.
1 shows an example of the configuration of a sub-NW TN connection matrix by concatenation of FIG.
3C. The inter-TN connection matrices of the elements NW1 and NW2 shown in FIG. 13B are combined to create a sub-NW inter-TN connection matrix. The matrix element is T of elements NW1 and NW2.
The matrix elements of the N-to-N connection matrix are transcribed to their corresponding portions, and the connection relation between the network connection nodes connecting the elements NW is added. In this example, since the HTN3 of the element NW1 and the HTN2 'of the element NW2 are linked,
To add a connection relationship between network connection nodes, 1 is added to a matrix element of which row is HTN3 and column is HTN2 ', row is HTN2', and column is HTN2 '.
It is realized by writing 1 in each of the matrix elements of TN3.

FIG. 14 shows an example of creating a routing table of a traffic network in the element NW. The routing table is set for each TN. This table is a "goal"
It is composed of a column and a "next TN" column. In the target field, the LTA (here, the set of subscriber line exchanges included in the lower traffic area), which is the final destination of the call, is set. The LTA as the destination matches the case where the level 1 TN is targeted, except when one switch is shared between different TNs.

In the next TN column, the TN through which the call destined for the destination described in the target column first passes is set. In FIG. 14, the routing table of each TN is created using the attribution table of the element NW. When the target is within its own area, a lower TN is set as the next TN, and when the target is outside the own area, the next TN is the TN except the highest level, the upper TN, and the TN of the highest level.
Then, the same level TN is set.

For example, in the case of the membership table shown in FIG. 14A, in the routing table of the level 3 node TN1, the target is LT as shown in FIG. 14B.
At A111, TN11 is set as the next TN. TN
In the routing table of No. 11, when the target is the LTA 211 outside the own area as shown in FIG. 14C, TN1 is set as the next TN.

FIG. 15 conceptually shows a traffic network creation flow for one sub NW. In the network topology creation unit, by inputting the design target area of the element NW, the traffic area (geographic range of traffic handled by one transit exchange), the number of layers, etc., the membership relationship table and the TN connection matrix for each element NW are input. create. In the routing table creation section, using the attribution table,
A routing table (for each TN) for each element NW is automatically generated.

Next, the network topology creation unit calculates a direct link setting candidate based on the membership relationship table and the traffic matrix between subscriber line exchanges (LS) in the element NW created by the traffic creation unit. By inputting the direct link setting section based on the result, the network topology creating unit automatically corrects the TN connection matrix considering the direct link, and the routing table creating unit automatically corrects the routing table considering the direct link.

Further, by inputting a network connection node for network-connecting the elements NW, the TN connection matrix and routing table of one sub-NW in which the elements NW are respectively network-connected are automatically created. Further, based on the information about the connection to the service network, the TN connection matrix and the routing table are automatically corrected in consideration of the connection of the service network.

The traffic creation section creates traffic of a service network such as a nationwide subscriber line exchange (LS) traffic matrix (for general telephones) and international calls, and also inputs sub-NWs and element NWs. Based on the design target area, the corresponding data is cut out. 16 and 17 show flowcharts of traffic allocation in the traffic network. TN on the route (route) from the originating subscriber line exchange (originating LS) to the destination subscriber line exchange (ending LS), and the originating LS on the traffic link
When allocating the traffic of 1), one outgoing LS belongs to one LTN (level 1 node), and the destination LS also belongs to one LTN, paying attention to the route from the outgoing LS to the incoming LS. Between LS and departure LTN and departure LTN-
By considering the destination LTNs by dividing them into two, it is possible to significantly reduce the calculation amount (even if the number of LSs is several thousand, the number of LTNs is considered to be about several tens). In addition,
Here, the LTN to which the originating LS belongs is defined as the originating LTN, and the set of the originating LS belonging to the originating LTN is defined as the originating LTA. Similarly, if the LTN to which the destination LS belongs is the destination LTN,
A set of destination LSs belonging to the destination LTN is called a destination LTA.

In FIGS. 16 and 17, first, the sub-N corresponding to the network for traffic allocation using the inter-TN connection matrix.
A W traffic counter table (including a node traffic counter table and a link traffic counter table) is created, and the work traffic counter table is created by copying the sub NW traffic counter table for work (S ₁ ). .

Next, one LTA is selected as an originating LTA (S ₂ ), the work traffic counter value is initialized to 0 (S ₃ ), and one LTA different from the originating LTA is selected as a destination LTA. To do. The TN connected to the LS in the originating LTA is defined as the originating LTN. TN connected to LS in destination LTA
Is the arrival LTN (S ₄ ). Using the inter-LS traffic matrix, the originating LTN → arriving LTA call volume is calculated (S ₅ ). Origination LTN → wearing LTA traffic intensity, allocated to each TN, traffic links on route from originating LTN to wear LTN (made using routing table), and sets the traffic counter for work (S _6). The work traffic counter value is added to the sub-NW traffic counter (S ₇ ). All L as a wearing LTA
It is checked whether TA has been selected (S ₈ ), and if there is any remaining, the process returns to step S ₃ and this process is performed for all LTAs (destination LTAs) other than the originating LTA of interest.

Next, one LS is selected from the originating LTAs (S
₉ ) Then, using the inter-LS traffic matrix for the LS, the total outgoing call volume and the total incoming call volume are calculated (S ₁₀ ) and assigned to the links between the LS and the outgoing LTN, respectively. This traffic volume is set in the traffic counter for sub-NW (S
₁₁ ). Determine select all LS from the calling LTA (S _12), the process returns to step S ₉ any remaining performs this processing for all the LS in calling LTA.

The onset Selecting all LS from LTA Determine select all LTA as originating LTA (S _13), if they remain back to step S _2, all of the LTA the above processing as originating LTA Do until you select. In this way, traffic allocation in the traffic network is performed, and the call volume flowing through each TN and traffic link can be obtained.

In the traffic allocation in the traffic network, since there is one route for the call from the originating LS to the terminating LS, it is not necessary to apply the conditional depth-first search algorithm. 18 and 19 show flowcharts of traffic allocation in the circuit network. When allocating the traffic of the originating LS to the exchange and the link (line group) on the route (route) from the originating side subscriber line exchange (originating LS) to the destination side subscriber line exchange (ending LS),
Considering that the number of LTNs is about one tenth of the number of LSs, the originating LTN (the LTN to which the originating LS belongs) → the destination LTN
By applying a conditional depth-first search algorithm (described later) to the section of (the LTN to which the destination LS belongs), the calculation amount can be greatly reduced. The flowcharts of FIGS. 18 and 19 are based on this idea, and will be described below.

In this flowchart, first, a traffic counter table for sub-NW (a node traffic counter table and a link traffic counter table) corresponding to a network to which traffic is allocated is created using the connection matrix between exchanges. Furthermore, the traffic counter table for the sub NW is copied for work, and the traffic counter table for work is created (S ₁ ).

Further, one LTA (here, it represents a set of LSs included in the lower traffic area) is selected as an outgoing LTA, and an LTN directly connected to the LS in the outgoing LTA is set as an outgoing LTN ( S ₂ ). Further, one LTA different from the calling LTA is selected as a destination LTA, and the destination LTA is selected.
The LTN connected to the LS in the
₃ ). In addition, “each exchange (calling side SW) in the calling LTN → each LS (calling side LS) call volume table in the calling LTA”, “each LS in the calling LTA (calling side LS) → calling side, which is necessary for calculation LS
Call volume ratio table ", to create a" call-out SW → called side LS call volume ratio table "(S _4).

Next, one LS is selected from the originating LTAs to be the originating LS (S ₅ ). Using the inter-LS traffic matrix, the call volume of "Calling LS → Calling party LS" is extracted and
→ Called LTA call volume, which is the call volume of the called LS, is calculated and the ratio of the call volume of the called LS to each of the called LS with respect to the call volume of the called LS is shown. Departure LS
-> Call volume ratio of "Calling side LS" is calculated, and "Calling side LS → Calling side L
Described S traffic intensity ratio table "(S _6).

Next, the inter-switch connection matrix and the designated outgoing L
The outgoing LS total call volume is issued by using the “outgoing LS → incoming side SW call volume ratio” that indicates the ratio of the outgoing call volume of S to the exchange (one or more) of the outgoing LTN. It is assigned to the side SW and added to the traffic counter for the sub NW. “Calling SW → Calling side LS” indicating the call volume received from the calling side SW to the calling side LS by using the call volume assigned to the calling side SW and the “calling side LS → calling side LS call volume ratio table” Of the call volume is added to the “calling side SW → calling side LS call volume table” (S ₇ ). Checks whether processing all the originating LS as originating LS (S _8), the process returns to step S ₅ If there is remaining performs processing as originating LS for all originating LS similar processing.

Next, using the "calling side SW-> calling side LS call volume table", the total number of outgoing calls, which is the sum of the incoming call volume for each called LS, is calculated.
The call volume ratio of the outgoing SW → the incoming call LS, which indicates the ratio of the incoming SW to the incoming call LS with respect to the total W call volume, is calculated, and the “calling side SW → the incoming side LS call volume is calculated. Ratio table ”(S ₉ ).

Next, the traffic counter value for work is initialized (0) (S ₁₀ ), one switch is selected from the calling SW, and it is set as the calling SW (S ₁₁ ). Using the depth search algorithm, assign each switch and link on the route (created using the routing table of the switch) from the source SW to each switch (destination SW) in the destination LTN, and work traffic counter To (S
₁₂ ).

The call volume assigned to the receiving side SW and the "calling side S
W → calling side LS call volume ratio table ”is used to calculate the link call volume between the receiving side SW and the receiving side LS and add it to the work traffic counter (S ₁₃ ). The content of the work traffic counter, which is the calculation result so far, is added to the sub NW traffic counter (S ₁₄ ), and thereafter, the content of the work traffic counter is initialized (set to 0),
Check whether all of the originating SW and the processing of as originating SW (S _15), and returns to step S ₁₀ if there is remaining.

After performing the above calculation for all the originating SWs, it is checked whether all the LTAs other than the originating LTA have been calculated as the destination LTAs (S ₁₆ ), and if there are remaining LTAs, the process returns to step S ₃ . . After all is done, then all L
Checks were treated with TA as originating LTA (S _17), the flow returns to step S ₂ if the remaining. In this way, the traffic allocation is completed, and the number of lines between the exchanges is calculated based on the allocated call volume (sub-NW traffic counter value) (S ₁₈ ).

FIG. 20 shows a flowchart of the conditional depth-first search algorithm. The idea of this algorithm is as follows. In the directed graph G, some nodes are designated as entrance nodes, some other nodes are designated as exit nodes, and the remaining nodes are designated as internal nodes (S ₁ ). It is assumed that the entrance node has only an output branch, and the exit node has only an input branch. The call volume is assigned to the entrance node in advance and is marked as assigned (S ₂ ). Unassi for all remaining nodes and branches
it is assumed that marked that gned (S _3). It is assumed that the ratio of assigning traffic to the branches emerging from each node is fixed.

The conditional depth-first search means assigned and mar
If you cannot visit other than the marked nodes,
The priority search is performed. With conditional depth-first search
Emits one node v in G marked as assigned
Set as the starting point. mark v as visited
(S_Four). At the same time, the call volume assigned to v is changed from v
For each of the outgoing branches, divide according to the traffic ratio.
wear. At this time, at the node w adjacent to v,
If traffic is assigned to all of the branches that apply,
The sum of the call volumes that flow into node w is the call volume assigned to w
And change the mark of node w to assigned (S₇). Su
Step S _FiveAfter assigning the call volume to the branch coming out of the node with
Unassigned nodes with all
Judge whether there is a node assigned to the branch
(S₆), If there is that node, step S₇Move on to
If so, check whether all nodes except the exit node have been visited.
Click (S₈), If there are any remaining steps S_FourBack to the article
Assi adjacent to v using recursive depth-first search
Process the nodes marked gned one by one. Article from v
When you have visited all the nodes that can be visited with a condition, v
The related search ends. Nodes not yet visited (Exit node
Except) remains, one of them is the new starting point
Choose as. Visit all G nodes (excluding exit nodes)
Repeat this until

FIG. 21 shows an example of a directed graph to which the traffic allocation algorithm is applied. The nodes A to G correspond to the TN in the traffic network and the switch in the line network. The branches with arrows connecting these nodes indicate the links between the nodes and the traffic flow direction (set in the routing table). Node A is the entrance node,
The node G is designated as an exit node and the remaining nodes are designated as internal nodes, and traffic is assigned by the conditional depth-first search algorithm.

22 and 23 are schematic flowcharts of the network topology creating unit 22 of the circuit network designing unit 16. First, from the call volume flowing through each TN obtained from the traffic network design unit 15, the exchange model and the number of exchanges corresponding thereto are obtained (S ₁ ). Next, enter the link (line group) setting method. For example, in the case of setting a link from a switch in one TN to a switch in another TN, the reliability is taken into consideration and the number of switches from one switch to be set is input ( S ₂ ).

Next, the T obtained from the traffic network design unit 15
The pair of the calling TN- wearing TN with the N inter-connection matrix connection relationship selects one (S _3). With respect to the two selected TNs, paying attention to the exchanges in each TN, the links are set as follows. First, select one of the one switch in the calling TN (S _4), now the switch in wearing TN, with respect to the originating exchange in question, it is determined whether the counter link is set (S ₅₎ .

In the case where the opposite link is set up, among the exchanges within the destination TN to which the opposite link is set up, for the exchange having a small link counter value indicating the number of accommodated links for each exchange. preferentially setting the link, is described between exchanges connection matrix (S _6). While performing this process until the number of links required by the originating exchange, and judged to set the number of outgoing links required (S _7), if the destination exchange number counter link is set is insufficient In regard to the shortage, among the exchanges within the destination TN in which the opposite link is not set, the exchange having a smaller link counter value is preferentially set with a link and is described in the inter-exchange connection matrix (S ₈ ).

If the opposite link is not set in step S ₅ , the process proceeds to step S ₈ and the link is preferentially set in the exchange having the smaller link counter value among the exchanges within the TN of the destination TN. Repeat until the number of links required by the exchange is reached and enter in the inter-exchange connection matrix. Checks were link setting for all switches in the calling TN (S _9), the same processing returns to step S ₄ any remaining.

Next, it is checked whether or not the links have been set for all the source TN-target TN columns (S ₁₀ ), and if there are remaining links, the process returns to step S ₃ and a connection relation is newly added from the TN connection matrix. One of the originating TN and the destination TN pair in is selected and the same processing is performed. This process connects all Ts
Perform on N. An inter-TN connection matrix for the traffic network shown in FIG. 24A is shown in FIG. 24B, and a circuit network is shown in FIG. 24C.
FIG. 24D shows an example of creating an inter-switch connection matrix of TN1-TN3 (line network) in the case shown in FIG. In the TN-to-TN connection matrix of the traffic network, a link is set between the TN-exchanges having a connection relationship between the TNs and is expressed as an TN-connection matrix. For example, in the figure, 1 is set in TN1 (row) → TN3 (column) from the TN connection matrix of the traffic network, and it can be seen that there is a link in TN1 → TN3. TN1 has switches for SW1 and SW2, and TN3 has switches for SW3, SW4, and SW5.
A connection matrix between 1-> TN3 exchanges is created. That is, SW1, SW2 are rows, and SW3, SW4, SW5
In columns, from SW1 of TN1 to SW5 of TN3, TN
1 is set (the link is set) from SW2 of 1 to SW3 of TN3. Note that, here, it is assumed that a link is set from one exchange in one TN to only one exchange in another TN.

FIG. 25 shows an example of creating a routing table for the exchange. LTN in the traffic network shown in FIG. 25A
25 from the routing table of No. 1 and the inter-switch connection matrix of LTN1 → HTN1 in the circuit network shown in FIG. 25B.
Routing tables for the exchanges SW1 and SW2 in N1 are created as shown in FIGS. 21C and 21D, respectively. LT
The relationship between the switch and the link between A1 and HTN1 is shown in FIG.
This is the case of 5E.

As shown in FIG. 25C, the target column of the routing table of SW1 is copied from the target column of the routing table of LTN1, and LTN1 → HT is entered in the next exchange column.
In the inter-switch connection matrix of N1 (in the routing table of LTN1, the target is LTA2 and the next TN is HTN1, so this connection matrix is used).
Enter SW3 and SW4 (both are called sides) where 1 is set at 1 (calling side). The same will be described below in the routing table.

[0065]

The network construction method according to the present invention has the following effects. By hierarchizing the network into two layers, a traffic network and a circuit network from the upper layer, first designing the traffic network, and automatically handing over the design results to the circuit network design, concealing design parameters between layers As a result, the complexity of network design is reduced and can be expressed concisely, and complex networks can be handled easily.

Also, even if the design parameter in one layer changes, it does not affect other layers, and only that layer needs to be redesigned. For example, even if the exchange type is changed, it is not necessary to redesign the traffic network. For each of the traffic network and circuit network design,
By functionally separating the network creation part and the traffic allocation part, and expressing the interface between them in the general form of the internode connection matrix and routing table, even if the network creation part is changed, the traffic creation part is not affected. Will fall short.

Even if the routing table is defined for each node so that the routing table is modified, the modification only needs to be performed on the corresponding part. By automatically generating the network topology and routing table creation, the design operation can be greatly reduced. By applying a conditional depth-first search algorithm that newly considers traffic allocation in the traffic allocation unit of the network design unit, the amount of calculation can be greatly reduced.

By defining the network in the design target area as a network having a complex structure, it is possible to represent an actual network. In this way, complex networks can be expressed in a simple and easy-to-understand manner, it is possible to flexibly respond to changes in network configuration and design algorithms, and design operations can be significantly reduced. As a result, large-scale networks such as real networks can be obtained. However, it is possible to realize a system that can be efficiently designed in a centralized manner.

[Brief description of drawings]

FIG. 1 is a diagram showing a target range of network design.

FIG. 2 is a diagram showing the concept of network hierarchization according to the present invention.

FIG. 3 is a diagram showing a concept of a traffic network.

FIG. 4 is a diagram showing a concept of a circuit network corresponding to the traffic network of FIG.

FIG. 5 is a diagram showing an example of a network having a composite structure.

FIG. 6 is a block configuration diagram showing an outline of a network configuration device according to an embodiment of the invention of claim 3;

FIG. 7 is a flowchart showing an outline of a processing procedure of a network creating unit 17 of the traffic network designing unit 15.

FIG. 8 is a view showing a sequel to FIG. 7;

FIG. 9 is a diagram showing an image of creating a traffic network.

FIG. 10 is a diagram showing a connection image to a service network (network such as international call and facsimile) in the traffic network.

FIG. 11 is a diagram showing an example of a connection matrix between an element NW and its TN.

FIG. 12 is a diagram showing an example of an element NW and a membership table thereof.

FIG. 13 is a diagram showing an example of elements NW1 and 2, TN connection matrices thereof, and an example of a sub-NW TN connection matrix formed by connecting these NWs.

FIG. 14 is a diagram showing an example of creating a routing table of an element NW with respect to a membership table.

FIG. 15 is a flowchart showing an outline of creating a traffic network.

FIG. 16 is a flowchart showing a processing procedure of traffic allocation in the traffic network.

FIG. 17 is a view showing a sequel to FIG. 16;

FIG. 18 is a flowchart showing a traffic allocation processing procedure in the circuit network.

FIG. 19 is a view showing a sequel to FIG. 18;

FIG. 20 is a flowchart showing a conditional depth-first search algorithm.

FIG. 21 is a diagram showing an example of a directed graph.

FIG. 22 is a flowchart showing a schematic processing procedure of a network topology creation unit of the circuit network design unit.

FIG. 23 is a view showing a sequel to FIG. 22;

FIG. 24 is a diagram showing an example of a traffic network, an inter-TN connection sequence, and a line network corresponding to an inter-switch connection matrix of the line network.

FIG. 25 is a diagram showing a routing table of a line network, a routing table of a corresponding traffic network, and a line network, respectively.

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Hiroharu Saito 1-1-6 Uchisaiwaicho, Chiyoda-ku, Tokyo Inside Nippon Telegraph and Telephone Corporation (56) Reference JP-A-5-207132 (JP, A) JP 5-207068 (JP, A) JP-A-7-44519 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H04M 3/00

Claims (3)

(57) [Claims] 1. In a network configuration method for carrying a call (traffic) such as a telephone, the network is a node focusing on the flow of traffic.
A traffic network consisting of links connecting nodes,
A group of switches that make up a node and a line group that connects the switches
In the process of creating the above-mentioned traffic network, the node
Membership table that indicates whether the node belongs to
An internode connectivity matrix showing how they are connected
Create and use the created membership table
Go first with the node that is the final destination of the call for each
Create a routing table showing the nodes and
Connection matrix between nodes and the routing table for each node
Based on the traffic matrix between subscriber line exchanges,
The traffic volume flowing to the nodes and links, and the traffic volume flowing to the nodes during the process of creating the network .
Based on the traffic, find the exchange model and the number of exchanges, create an inter-exchange connection matrix showing the connections between exchanges, and create a routing table for each exchange from the inter-exchange connection matrix and the routing table for each node above, Based on the connection matrix between exchanges and the routing table for each exchange, the traffic is assigned to the network and the exchanges are switched.
Network configuration wherein that you calculate the number of lines between. 2. A network comprising one or several
It is expressed as a network with a complex structure that consists of a combination of sub-networks that can be designed independently and the element networks that make up the sub-network.
Can be used as the starting point for each
The flow of service traffic is independent of each other, and element networks are networks with a hierarchical structure. One or several of these are connected to form a subnetwork, and each subnetwork and each element network are configured. Method
, It by configuring the network according to claim 1, wherein
A method for configuring a network. 3. A link connecting nodes to each other
Traffic network design means and the nodes
Design of a circuit network consisting of a line group connecting a switchboard and a switchboard
And means, the traffic network design means, attributable to which node node
Means to create a membership table that indicates whether
Node-to-node connection that indicates how the nodes are connected
A means for creating a matrix and the created membership table
With the node that is the final destination of the call for each node
Create a routing table that shows the first node to go through.
Means, inter-node connection matrix, and each node
Based on the routing table of
Traffic matrix to the nodes and links.
Means for allocating the amount of traffic, and the circuit network design means is for traffic flowing to the node.
A method for determining the replacement model and the number of replacements based on
The connection matrix between the exchanges showing the connection of
Interchange from the continue column and the routing table for each node above
A means for creating a routing table for each exchange, the connection matrix between exchanges , and the routing table for each exchange.
Switch to assign traffic to the network based on
And a means for calculating the number of lines between the network constituent devices.