JPH0983546A - Route selecting method/device and communication network design method/device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently allocate a more inexpensive optimum route by selecting the routes of plural bi-directional connections in accordance with an order based on evaluation values calculated on the plural bi-directional connections. SOLUTION: A route selection device 1 consists of an evaluation value calculation part 2 and a route selection part 3. The evaluation values of the bi- directional connections are calculated based on the band of the up connection and the band of the down connection of the bi-directional connection, and the routes of the plural bi-directional connections are selected in accordance with the order based on the evaluation values calculated on the plural bi-directional connections. When the bands of the up connection and the down connection differ, namely, even when the route of an asymmetrical direction connection is selected, the same route can be selected on the up connection and the down connection, which make a bi-directional pair.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a route selection method and a route selection device for selecting an optimum information transfer route in a communication network formed by connecting a plurality of node devices.
The present invention also relates to a communication network designing method and a communication network designing device for designing an optimum communication network capable of accommodating a desired traffic demand.

[0002]

2. Description of the Related Art For example, an ATM (Asynchronous) for connecting a plurality of terminal devices through a plurality of exchanges and transferring fixed length packets called cells between these terminal devices for communication.
In a usTransfer Mode (asynchronous transfer mode) communication network, for example, it is necessary to set a VP network as a path for transferring the cell by establishing a plurality of VPCs (Virtual Path Connections) on the physical network.

In a communication network composed of a plurality of nodes and links, such as this ATM communication network, there have been various conventional solutions to the problem of determining the optimum routes for a plurality of connections (route selection problem). Proposed.

In general, in order to optimally solve this problem, simultaneous equations must be solved by using the simplex method or the like, and finding an optimum solution for a large-scale problem requires the required amount of calculation and memory. In terms of quantity, we have to give up.

Therefore, a method for approximating the route selection problem is required. This approximate solution includes, for example, Ryoichi Sasaki, Tsutomu Nakamura, Michio Suzuki, and Takashi Kagei, "Network Reconfiguration Optimization Techniques for High-speed Digital Circuit Failure" (IEICE Transactions vol. J72-BI, No.4). , pp.272-280, 1989)
There is.

This method is described for route selection for unidirectional connections, but not for bidirectional connections. Therefore, in order to select a path for a bidirectional connection using the above method, it is possible to consider the upstream connection and the downstream connection as independent unidirectional connections and select the path. Then, there is no guarantee that the upstream connection and the downstream connection forming the bidirectional pair will follow the same route.

As a path selection algorithm with a smaller amount of calculation, a method of setting a connection to the path with the smallest cost in the order of increasing the bandwidth can be considered. This method gives good results when selecting a route for a unidirectional connection or a bidirectional connection (symmetric bidirectional connection) in which the bandwidths of the upstream connection and the downstream connection forming a bidirectional pair are equal.

However, when a route is selected for a bidirectional connection (asymmetric bidirectional connection) in which the bandwidths of the upstream connection and the downstream connection forming a bidirectional pair are different, there are two types of bandwidth for one bidirectional connection. This algorithm cannot be applied because it is assigned. Of course, it is possible to select the route by considering the uplink connection and the downlink connection of the bidirectional connection as independent one-way connections, but in that case, there is no guarantee that the uplink connection and the downlink connection forming the bidirectional pair will pass the same path. .

On the other hand, considering the case of accessing the image database, for example, the connection from the terminal to the image database may be slow, but the connection from the image database to the terminal is required to be fast. It is expected that the demand for establishing such asymmetric bidirectional connections will increase in the future.

In an ATM communication network, when a connection setting request is issued along a certain route, VCI (Virtual Channel Identifie) is applied to both the up and down directions on the route.
r), that is, the identifier of each connection is secured. In other words, when setting up a one-way connection,
This is achieved by setting the required bandwidth in the reverse direction to "0", and the VCI in the reverse direction remains secured. Therefore, if different routes are set for the upstream connection and the downstream connection forming a bidirectional pair, the VCIs allocated in the opposite directions of the upstream and downstream connections are wasted.

Further, when the upstream connection and the downstream connection forming a bidirectional pair pass through different paths, OAM (Oprations, A) for each of the upstream and downstream connections.
dministration and Maintenance) A cell loopback connection is required, and a total of four connections are required.

On the other hand, when the paths of the upstream connection and the downstream connection are the same, the upstream (downstream)
Since a downlink (uplink) connection can be used for the connection for loopback of the OAM cell of the connection, two connections may be established. Therefore,
From the viewpoint of connection management as well, it is desirable that the routes of the upstream connection and the downstream connection forming the bidirectional pair are the same.

Although the ATMT communication network has been taken as a specific example here, there is not only the ATM communication network but also a communication network in which it is desirable that the paths of the upstream connection and the downstream connection forming the bidirectional pair are the same, and It is expected to appear in the future.

Next, the approximate solution of the above-mentioned route selection problem ("high speed digital line failure network reconfiguration optimization technique") will be briefly described.

First, the gain is defined. It is conceivable that there are routes that are not desirable routes and routes that are established, and if all the connections cannot be established, the connection to be established may have priority. A variable indicating such a preference of a route and a priority of a connection is defined as a gain, and a gain obtained when the route 1 is assigned to the connection k is represented by C _kl . Then, here, under the condition that the total bandwidth of the connections passing through each link is equal to or less than the capacity of the link, the route of the connection is determined so as to maximize the obtained gain. Therefore, the link j (= 1,
, ..., J) is used as E _j, and the vector P _u is defined as P _u = {E ₁ , E _2, ..., E _J } by using this.

When the route l is assigned to the connection k, the vector P _kl is defined as P _kl = {F _kl1 , F _kl2 , ..., F _klJ } by using the load F _klj newly added to the link j. G _kl is _calculated by equation (1), and the connection and route that maximizes G _kl is searched for.

[0017]

[Equation 1] However, A · B represents the inner product of the vectors A and B, and | A | represents the length of the vector A. When the connection that maximizes G _kl is m and the path is i, the path i is assigned to the connection m, and P _u = P _u + P _mi . After that, the same operation is repeated to allocate the route to the connection.

Conceptually, this method is designed so that the gain obtained is as large as possible and the vector P _u representing the band used by the routed connection to date is as _small as possible. Has determined the connection and its route. This is because the load on the link is distributed as much as possible in terms of the bandwidth used,
It can be said that the load on each link is made uniform.

As a specific example, in the network shown in FIG. 16, the bandwidth from node A to node C is 10 Mbps.
Consider the case of 10 connections. However, it is assumed that the links of the network in FIG. 16 are unidirectional links and have a capacity of 150 Mbps. Also, for simplicity,
The value of C _kl shall be equal to the bandwidth of connection k.
This means that high bandwidth connections are prioritized over low bandwidth connections and that each connection has no preference for a route.

At first, since G _kl cannot be calculated because P _u = 0,

[0021]

[Equation 2] Find the connection and route that maximizes. As a result, 1
Assuming that the main route is set to route A → B → C, P _u = {10
/ 150,10 / 150,0,0,0}, and the second route is stretched over the route A → D → C. After that, by repeating this, five connections are established on the routes A → B → C, and five connections are established on the routes A → D → C. In this example, 10 connections are route A → B →
Although it can be set to C, the connections are distributed over the two routes.

Since the load tends to be distributed in this way, the load of the link may be distributed more than necessary, which may deteriorate the performance of path selection. For example, consider a case where four connections as shown in FIG. 17 are set up in the network of FIG.

Initially, since G _kl cannot be calculated because P _u = 0, a connection and a route that maximize Equation (2) are searched for. As a result, route D → E to connection 4
Is assigned. Next, P _u = {0,0,0,0,5
/ 150} to find the connection and route that maximizes equation (1). As a result, connection 1 is route A →
It will be assigned to B → C. Next, P _u = {10 /
150,10 / 150,0,0,5 / 150} is used to find the connection and route that maximizes equation (1).
As a result, the connection 2 is assigned to the routes A → D → C. Next, P _u = {10/150, 10/1
50, 10/150, 10/150, 5/150} is used to find the connection and the route that maximize the expression (1). However, because the remaining bandwidth of link 3 is insufficient,
There is no route in which connection 4 can be set. In this example, in order to disperse the load of the link, the route of the connection 2 is set to A → D → C instead of A → B → C, and the route of the connection 3 is lost.

Similar to the above-described route selection problem, for example, when designing a packet leased line network, the amount of packets to be transmitted is estimated for each combination of the transmitting node and the receiving node. However, it is necessary to solve the problem of determining the capacity of the link set between the nodes and the route of the packet, that is, the communication network design problem. As a conventional solution example of this communication network design problem, for example, Queueing Systems (vol 2, pp. 270-421, by Leonard Kleinrock,
John Wiley & Sons, 1976).

This method is based on the Flow Deviation method (FD method).
Is called an effective method. However, in this method, the objective function to be minimized must be a function of the amount of traffic passing between the nodes, and the objective function must be a convex function that can be differentiated twice with respect to the amount of traffic passing between the nodes. I have to. Specifically, when minimizing the sum of link costs in a network, the sum of link costs must be expressed as a function of the amount of traffic passing between nodes, and it also represents the sum of link costs. The function must be a convex function that can be differentiated twice with respect to the amount of traffic passing between each node. If this condition is not satisfied, the FD method can only obtain a local optimum solution.

When a link does not have an arbitrary capacity but a fixed capacity for each type of link, the sum of the link costs becomes a discontinuous function with respect to the amount of traffic passing through each node, and twice. Not differentiable. Therefore, it was not possible to obtain an optimal solution, and there was no effective method for obtaining an approximate solution.

[0027]

As described above, in the operation of a communication network composed of a plurality of nodes (switches, terminal devices, etc.) and links (transmission lines) and the design of such a communication network, generally, It is indispensable to find the optimum route, taking into consideration the link capacity, the traffic volume between each node, and the cost.

However, the conventional approximate solution method of the route selection problem is to perform the route selection for the one-way connection. Therefore, when performing the route selection of the asymmetrical bidirectional connection, the bidirectional pair is made up. The first problem is that the route selection problem cannot be solved under the condition that the connection and the downstream connection pass through the same route.

Further, the conventional approximate solution method of the route selection problem is
In order to select the route to distribute the load on the link,
There is a second problem that the result of route selection is deteriorated by excessively distributing the load.

Further, in the conventional network design problem solving method, if the evaluation function is not a convex function that can be differentiated twice with respect to the traffic amount passing through each link, the optimum solution cannot be obtained and an approximate solution is obtained. There was the third problem of having an effective method.

In view of the first problem, the present invention has
Even when the bandwidths of the upstream connection and the downstream connection forming the bidirectional pair are different, that is, when the route selection of the asymmetrical bidirectional connection is performed, the same route can be selected for the upstream connection and the downstream connection forming the bidirectional pair. It is an object of the present invention to provide a route selection method and a route selection device that can be used.

Further, in view of the second problem, the present invention provides a route selection method and a route selection method capable of obtaining a highly accurate approximate solution of a route selection problem without excessively distributing load distribution of links. An object is to provide a selection device.

Furthermore, the present invention has been made in view of the third problem.
Provided is a communication network designing method and a communication network designing device capable of providing an accurate approximate solution of a network designing problem even when the evaluation function of the communication network designing problem is not a convex function that can be differentiated twice.

[0034]

According to a route selection method of the present invention, a plurality of bidirectional connection routes are passed through a communication network formed by connecting a plurality of node devices through the same upstream and downstream connections. A route selection method for selecting the evaluation value of the bidirectional connection based on the bandwidth of the uplink connection and the bandwidth of the downlink connection of the bidirectional connection, and for each of the plurality of bidirectional connections. When the bandwidths of the upstream connection and the downstream connection forming the bidirectional pair are different by selecting the routes of the plurality of bidirectional connections according to the order based on the calculated evaluation value, that is, the routes of the asymmetric bidirectional connection Even when making a selection, a bidirectional pair of upstream connection and downstream connection It is possible to select the same path for emission.

Further, the route selecting device of the present invention selects a plurality of bidirectional connection routes in the communication network formed by connecting a plurality of node devices so that the upstream connection and the downstream connection pass through the same route. A route selection device, comprising: an evaluation value calculation means for calculating an evaluation value of the bidirectional connection based on a bandwidth of an up connection and a bandwidth of a down connection of the bidirectional connection; A route selection means for selecting the routes of the plurality of bidirectional connections in accordance with the order based on the evaluation value calculated for each of the bidirectional connections of 1. Also, select the same route for the upstream connection and the downstream connection that form a bidirectional pair. It becomes possible.

Further, the route selection method of the present invention selects a plurality of bidirectional connection routes in a communication network formed by connecting a plurality of node devices so that the upstream connection and the downstream connection pass through the same route. A route selection method, which calculates an evaluation value of the bidirectional connection based on the parameters output as necessary, the bandwidth of the up connection and the bandwidth of the down connection of the bidirectional connection, and the calculated value According to the order based on the evaluation value, the routes of the plurality of bidirectional connections are respectively selected, the selection result is output as the route selection result corresponding to the parameter, and the route selection result is output based on the output route selection results. When selecting the route of an asymmetric bidirectional connection by determining the optimal route of multiple bidirectional connections Also, it is possible to select the same path for the uplink connections and downlink connections that form a bidirectional pair.

Further, the route selecting device of the present invention selects a plurality of bidirectional connection routes in the communication network formed by connecting a plurality of node devices so that the upstream connection and the downstream connection pass through the same route. A route selection device, based on the parameter output means for outputting different parameters as necessary, the parameters output by the parameter output means, and the bandwidth of the upstream connection and the bandwidth of the downstream connection of the bidirectional connection. , The evaluation value calculation means for calculating the evaluation value of the bidirectional connection and the order based on the evaluation value calculated by the evaluation value calculation means, each of the plurality of bidirectional connection routes is selected, and the selection result is selected. Output means for outputting as a route selection result corresponding to the parameter, and a plurality of output means output by this output means By providing a determining means for determining the optimum route of the plurality of bidirectional connections based on the route selection result, even when asymmetric bidirectional connection routes are selected, an upstream connection and a downstream connection forming a bidirectional pair are provided. It is possible to select the same route for.

The route selection method of the present invention is a route selection method for selecting routes of a plurality of connections in a communication network formed by connecting a plurality of node devices, and searches for routes that the plurality of connections can take. , Reserve a band of the connection to a link on the searched route, and select one of the searched routes for each of the plurality of connections based on the reserved state of the band. By doing so, a highly accurate approximate solution of the route selection problem can be obtained.

Further, the route selection device of the present invention is a route selection device for selecting a route of a plurality of connections to a communication network formed by connecting a plurality of node devices, and a route which the plurality of connections can take. Route search means to search,
Based on the bandwidth reservation means for reserving the bandwidth of the connection to the link on the route searched by the route search means and the reservation state of the bandwidth reserved for the link by the bandwidth reservation means, By providing route selection means for selecting one of the routes searched by the route search means for each, an accurate approximate solution of the route selection problem can be obtained.

The communication network designing method of the present invention includes at least the number of nodes constituting the communication network, identification information of each node, network information including the type and cost of a link that can be set between each node, and at least communication. A communication network designing method for designing a communication network capable of accommodating a traffic demand required by the traffic information, based on traffic information including identification information of a source node and a destination node of a flow and the amount of the communication flow, , The initial topology of the communication network is set based on the network information, the initial topology is changed as necessary, and the set initial topology or the topology obtained by the change, Based on the traffic information, the route of the desired flow is set, the cost of the link on the set route is calculated, and the calculated cost is calculated. Cost, until the condition that a predetermined, while changing the set route, to calculate the cost of the changed link on the route,
When the calculated cost satisfies a predetermined condition, at least the path of the desired flow at that time, by outputting the communication network design information including the type of link set between the nodes, It is possible to give an accurate approximate solution of a network design problem.

Further, the communication network designing method of the present invention includes at least the number of nodes constituting the communication network, identification information of each node, network information including types and costs of links that can be set between the nodes, and A communication network designing method for designing a communication network capable of accommodating the connection, based on at least a requested bandwidth for each connection, connection information including identification information of a source node and a destination node, and based on the connection information , A communication flow is generated by gathering required bandwidths of connections having the same source node and destination node, and based on the generated communication flow and the network information, a route of the communication flow and a link between the nodes are generated. Design a virtual network by setting,
The route of the connection is determined based on the designed virtual network, and the link corresponding to the connection is set between the nodes based on the determined route of the connection. An accurate approximate solution can be given.

Further, the communication network designing apparatus of the present invention includes at least the number of nodes constituting the communication network, identification information of each node, network information including the type and cost of a link that can be set between each node, A communication network designing device for designing a communication network capable of accommodating the traffic demand requested by the traffic information based on the traffic information including at least the identification information of the source node and the destination node of the communication flow and the amount of the communication flow. An initial topology setting means for setting an initial topology of the communication network based on the network information; a topology changing means for changing the initial topology set by the initial topology setting means, if necessary; The initial topology set by the initial topology setting means or the topologies changed by the topology changing means. , A route setting means for setting a route of a desired flow based on the traffic information, a cost calculating means for calculating the cost of a link on the route set by the route setting means, and the cost. Change the route set by the route setting means until the cost calculated by the calculating means satisfies a predetermined condition,
When the cost calculation means controls the cost calculation means to calculate the cost of the link on the changed route and the cost calculated by the cost calculation means satisfies a predetermined condition, at least, By providing an output means for outputting the communication network design information including the path of the desired flow and the type of the link set between the nodes, it is possible to give a highly accurate approximate solution of the network design problem. it can.

Further, the communication network designing apparatus of the present invention includes at least the number of nodes constituting the communication network, identification information of each node, network information including the type and cost of a link that can be set between each node, A communication network designing device for designing a communication network capable of accommodating the connection based on at least a requested bandwidth for each connection, connection information including identification information of a source node and a destination node, based on the connection information. , A flow generation means for generating a communication flow by collecting required bandwidths of connections having the same source node and destination node, and the communication flow based on the communication flow and the network information generated by the flow generation means. And a virtual communication network designing means for designing a virtual network by setting links between the nodes and the virtual network. Connection route determining means for determining the route of the connection based on the virtual network designed by the totalizing means, and the connection between the nodes corresponding to the connection based on the route of the connection determined by the connection route determining means. By providing the link setting means for setting the link, it is possible to provide a highly accurate approximate solution to the network design problem.

[0044]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

First, the first embodiment will be described.
In the following description of the first embodiment, a unidirectional connection is defined as a bidirectional pair with a connection whose direction is opposite and whose bandwidth is "0".
As a result, even if one-way connections and two-way connections are mixed, all can be treated as two-way connections. That is, the unidirectional connection is treated as a bidirectional connection by regarding the unidirectional connection as a bidirectional pair with the connection having the opposite direction and the band of "0". Further, for simplicity, the bidirectional connection is simply described as a connection, and when it is necessary to separately represent two connections forming a bidirectional pair, they are described as an upstream connection and a downstream connection.

FIG. 1 schematically shows the configuration of the route selection device 1 according to the first embodiment.

In FIG. 1, the route selection device 1 is composed of an evaluation value calculation unit 2 and a route selection unit 3.

The evaluation value calculation unit 2 has information necessary for selecting a route for a desired connection, for example,
Connection information including band, originating / destination node identification information, etc. is input, and an evaluation value for the connection is calculated by a predetermined procedure.

Information on a network to be routed, for example, network information including the topology of the network, the capacity of each link, the cost, etc. is input to the route selection unit 3 in advance. Based on this information,
For the modeled network for route selection,
The evaluation value calculation unit 2 performs a process of assigning an optimum route to each connection for which the evaluation value has been calculated.

FIG. 2 schematically shows the configuration of the evaluation value calculation unit 2, which includes a band comparison unit 2a, a weight sum calculation unit 2b, a first weight storage unit 2c, and a second weight storage unit 2d. Composed.

The bandwidth comparing unit 2a compares the bandwidth W _down bandwidth W _{Stay up-,} a downlink connection of the uplink connection is input as the connection information, and outputs a comparison result to the weight sum calculating unit 2b There is. That is,
Max (W _up , W _down ) and min (W _up , W _down ) are calculated. Here, the function max (x, y) is a function that returns the smaller value of x and y, and the function min (x,
y) A function that returns the value of x or y that is not larger.

The weight sum calculation unit 2b is adapted to obtain the evaluation value E from the equation (3) for each connection.

E = a ₁ · max (W _up , W _down ) + a ₂ · min (W _up , W _down ) ... (3) a ₁ and a ₂ are called weights, and first weight storage means, respectively. 2c, a value previously stored in the second weight storage means 2d.

Evaluation value E calculated by the weight sum calculation unit 2b
Is output to the route selection unit 3.

FIG. 3 schematically shows the configuration of the route selection unit 3, which includes an evaluation value storage unit 3a and a connection selection unit 3
b, the minimum cost route calculation unit 3c.

In the evaluation value storage means 3a, the evaluation value calculation unit 2
Evaluation value E calculated for each connection output from
Is temporarily stored.

The connection selection unit 3b selects connections in descending order of the evaluation value based on the evaluation value temporarily stored in the evaluation value storage unit 3a. The selected connection is the minimum cost route calculation unit. 3c is output.

The minimum cost route calculation unit 3c assigns the connection selected by the connection selection means to the route with the smallest cost among the routes having the remaining bandwidth that can accommodate the connection. After that, the connection selector 3b again selects a new connection, the minimum cost route calculator 3c performs a process of assigning a route, and by repeating this,
Routes are assigned to all desired connections.

In order to efficiently select the connection in the connection selection unit 3b, the evaluation values may be sorted in advance according to the size thereof to order the connections.

Next, referring to the flow charts shown in FIGS. 4 and 5, the operation process of the route selection device 1 of FIG. 1 will be described. Note that FIG. 4 mainly shows the processing of the operation of the evaluation value calculation unit 2, and FIG. 5 shows the operation processing of the route selection unit 3.

Further, here, a case will be described in which the route of each connection shown in FIG. 7 is selected for the network of the mesh type topology as shown in FIG.

FIG. 6 schematically shows the structure of a network which is a route selection target.
The nodes A, B, C, D, and E are accommodated, and the nodes are connected by a logical transmission path, that is, a link (indicated by an arrow in FIG. 6). In FIG. 6, links that are opposite to each other are described between the nodes, but they do not have to be physically two transmission lines. For example, it may be a wireless transmission line, or may be a single transmission line such as when performing ping-pong transmission in which the transmission direction is changed by time division.

From node X (X = A to E) to node Y (Y
= A to E) is described as a link XY, and a link X
It is assumed that the capacity Q (XY) and the link cost C (XY) are defined in Y.

Further, it is assumed that the total bandwidth of the connections passing through the link XY cannot exceed Q (XY). Further, the source node is X and the destination node is Z (Z = A
To E), the bidirectional connection i is, for example, route X →
The cost required for assigning Y → Z (which will be referred to as a route XYZ) can be defined by Equation (4) as D _i (XYZ).

D _i (XYZ) = {C (XY) + C (YZ)} W _down + {C (ZY) + C (YX)} W _up (4) However, W _down goes from node X to node Z A band for a downlink connection, and W _up is a band for an uplink connection from the node Z to the node X.

The cost D _i (XYZ) can also be defined by the equation (5) without using the connection bandwidths W _down and W _up .

D _i (XYZ) = {C (XY) + C (YZ)} + {C (ZY) + C (YX)} (5) In the description here, the formula (5) is used to calculate the cost D. I will use it. Further, for simplification of explanation, the capacity is set to 150 [Mbps] for all links, and the link cost is set to "1". Of course, the capacity and the link cost may take different values for each link. Also,
The link cost may change depending on the remaining capacity of the link.

In FIG. 7, the band of the connection from the source node to the destination node is described as the downstream band, and the band of the connection from the destination node to the source node is described as the upstream band.

As shown in FIG. 7, the values of the weights a ₁ and a ₂ stored in the first weight storage unit 2c and the second weight storage unit 2d are set to "1" and "0", respectively. A process of selecting a route of a connection (connection number i = 1 to 5) will be specifically described below with reference to FIGS. 4 and 5.

First, as connection information, for example, a connection number, a source node, a destination node, a downstream band, and an upstream band are input to the evaluation value calculation unit 2, the connection is sequentially selected (steps in FIG. 4). S2), the band comparison unit 2a compares the upstream band and the downstream band (step S).
3).

The comparison result shows that the up band W _up is the down band W.
_{If it is down} or more, the process proceeds to step S4, and if not, the process proceeds to step S5 and the weighted sum calculation unit 2b calculates the evaluation value E
Is calculated. That is, the upstream band W _up is the downstream band W _down
In the above case, the equation (3) becomes E = a ₁ · W _up + a ₂ · W _down (6), and when the downlink band is larger, the equation (3) becomes E = a ₁ · W _down + A ₂ · W _up (7), and in step S4, the evaluation value E is calculated using equation (6).
Is calculated, and in step S5, the evaluation value E is calculated using the equation (7).

For example, in the case of connection 1, since the downlink band is larger than the uplink band, the process proceeds to step S5,
From the equation (7), the evaluation value E is calculated as 150, the process proceeds to step S5 in the case of the connection 2 as well, and the evaluation value E is similarly calculated as 150 in the case of the connection 3, and the process proceeds to step S4 in the case of the connection 3. The evaluation value E is calculated as 140. In the case of connection 4, since the upstream band is larger than the downstream band, the process proceeds to step S4, and the evaluation value E is calculated as 130 from the equation (6). In the case of the connection 5, for example, the process proceeds to step S4, and the evaluation value is calculated as 10. With regard to the connections 3 and 5, since the sizes of the upstream band and the downstream band are equal, it is not necessary to provide a weight difference between the upstream band and the downstream band. Therefore, the same value is obtained in both step S4 and step S5. Calculate the evaluation value.

The connection i calculated in this way
When the evaluation value of (i = 1 to 5) is expressed as Ei, E1 = 1
50, E2 = 150, E3 = 140, E4 = 130, E
5 = 10.

When the evaluation values are calculated for all the connections (step S1), the evaluation values of all the connections are temporarily stored in the evaluation value storage unit 3a of the route selection unit 3.

The route selecting means 3 sequentially selects the connection i from the one having the largest evaluation value while incrementing i from "1" by "1", and selects a route having a residual band capable of accommodating the connection. The process of assigning the route with the lowest cost is performed. Here, since all the link costs are "1", the route with the smallest number of links may be selected to select the route with the smallest cost.

First, in the connection selection section 3b,
Of the evaluation values stored in the evaluation value storage means 3a, the connection 1 having the largest evaluation value (i = 1) is selected (steps S5 to S7), and in step S8,
Of the routes having the remaining bandwidth that can accommodate the connection 1, the route with the lowest cost, that is, the route AB in FIG. 6 is assigned as the route of the connection 1 (steps S8 to S10).

Next, i is incremented by "1" (step S11), and the connection 2 having the second largest (i = 2) evaluation value among the evaluation values stored in the evaluation value storage means 3a. Select (Step S5 to Step S
7) In step S8, the path with the smallest cost is searched from among the paths having the remaining bandwidths that can accommodate the connection 2 (step S8). At this time, since the link AB does not have a residual band enough to accommodate the upstream connection of the connection 2, the route ACB is selected and assigned to the connection 2 (steps S9 to S10).

Next, i is incremented by "1" (step S11), and the connection 3 having the third largest evaluation value (i = 3) among the evaluation values stored in the evaluation value storage means 3a. Select (Step S5 to Step S
7) In step S8, the route with the lowest cost is searched from among the routes having the remaining bandwidths that can accommodate the connection 3 (step S8). At this time, since the link AC does not have a residual band for accommodating the connection of the connection 3, the route with the lowest cost, AEDC, is selected as the downlink connection of the connection 3 and assigned. (Step S9 to Step S10).

Next, i is incremented by "1" (step S11), and the connection 4 having the evaluation value of the fourth largest value (i = 4) among the evaluation values stored in the evaluation value storage means 3a. Select (Step S5 to Step S
7) In step S8, the path with the lowest cost is searched from among the paths having the remaining bandwidths that can accommodate the connection 4 (step S8). At this time, since the link CD has already been allocated the connection 3 and does not have a residual band for accommodating the connection 4, the connection 4 has the highest cost among the routes not using the link CD. Although a small route is selected, in this case, since there is no route having a band for accommodating the connection 4, it cannot be assigned (step S9).

Next, i is incremented by "1" (step S11), and the connection 5 having the fifth largest evaluation value (i = 5) among the evaluation values stored in the evaluation value storage means 3a. Select (Step S5 to Step S
7) In step S8, the route with the smallest cost is searched from among the routes having the remaining bandwidths that can accommodate the connection 5 (step S8). At this time, although the connection 3 is already assigned to the route CD, there is a residual band for accommodating the connection 5, so the route CD is assigned as the route to the connection 5 (steps S9 to S10).

Next, when i is incremented by "1" (step S11), since the value is larger than the number of connections (i = 5) given in advance, the processing ends here, and is assigned to each connection. The route thus obtained is output (steps S6 and S12).

Here, in the connection selection unit 3b, the connections 1 and 2 having the same evaluation value are selected in the ascending order of connection numbers and the route is selected, but the source node number, the destination node number, and the connection The order for selecting the route of the connection may be determined using information such as priority.

Now, consider the case where "0.5" and "0.5" are used as the values of the weights a ₁ and a ₂ , respectively. As described above (according to the flowchart in FIG. 4), when the evaluation values are calculated for the connections 1 to 5 in FIG. 7, E1 = 80, E2 = 75, E3 = 140, E
4 = 75 and E5 = 10. In this case, in step S7 shown in the flowchart of FIG. 5, the order selected by the connection selection unit 3b is the connection 3, 1, 2,
The order is 4, 5.

When the processing is performed according to the flowchart of FIG. 5, the path AC is used for the connection 3, the path AB is used for the connection 1, and the path AED is used for the connection 2.
B, the route CD is assigned to the connection 4, and the route CD is assigned to the connection 5 as routes for the respective connections.

Next, consider the case where "0" and "1" are used as the values of the weights a ₁ and a ₂ , respectively. As described above (according to the flowchart of FIG. 4), when the evaluation value is calculated for the connections 1 to 5 of FIG. 7, E1 = 1
0, E2 = 0, E3 = 140, E4 = 20, E5 = 10
Becomes In this case, in step S7 shown in the flowchart of FIG. 5, the order selected by the connection selection unit 3b is the order of the connections 3, 4, 1, 5, 2.

When the processing is performed according to the flow chart of FIG. 5, the connection 3 has the path AC, the connection 4 has the path CD, the connection 1 has the path AB, the connection 5 has the path CD, and the connection 2 has the path A.
The EDB will be assigned as a route for each connection.

As described above, depending on the values of the weights a ₁ and a ₂ , it can be seen that the optimum route having a smaller cost is efficiently allocated to all the given connections 1 to 5.

Next, the second embodiment will be described.

FIG. 8 schematically shows the configuration of the route selection device 5 according to the second embodiment. In FIG. 8, the route selection device 5 includes a parameter change unit 6, an evaluation value calculation unit 7, a route selection unit 8, and a best result storage unit 9.

The parameter changing unit 6 holds in advance the values of various parameters required for the process of selecting the optimum route for a desired connection, or changes the values by a predetermined algorithm to change various parameters. It manages a plurality of values of the parameter for each and outputs it to the evaluation value calculation unit 7 as needed. The parameters include weights a ₁ and a ₂ , for example.

The evaluation value calculation unit 7 is inputted with information necessary for selecting a route for a desired connection, for example, connection information including band, destination node identification information and the like, and a predetermined procedure is followed. , The evaluation value for the connection is calculated.

The route selection unit 8 is preliminarily input with information about the network of the route selection target, for example, network information including the topology of the network, the capacity of each link, the cost, etc. Based on this information, the model is selected. With respect to the network that is the target of the route selection, the processing for assigning the optimum route to each connection for which the evaluation value is calculated by the evaluation value calculation unit 2 is performed. A specific configuration example of the route selection unit 8 is similar to that of the route selection unit 3 in the first embodiment (see FIG. 3).

The route selection result obtained by performing a predetermined process in the evaluation value calculation unit 7 and the route selection unit 9 while changing the parameter value is the optimum route selection result obtained in the route selection processes up to the previous time. The best result storage section 9 stores only the more optimal route selection result as compared with the other route selection results.

FIG. 9 schematically shows the configuration of the evaluation value calculation unit 7, which includes a band comparison unit 7a, a weight sum calculation unit 7b, a first weight calculation unit 7c, and a second weight calculation unit 7d. Composed.

The operation of the band comparison unit 7a is the same as the operation process of the band comparison unit 2a in the first embodiment.

Various parameters necessary for calculating the evaluation value in the weight sum calculation unit 2b are output from the parameter changing unit 6 as necessary, and are stored in a predetermined storage unit. For example, the weights a ₁ and a ₂ are stored in the first weight calculator 7c and the second weight calculator 7d, respectively.

Based on the parameter values output from the parameter changing section 6, the weight sum calculating section 7b calculates an evaluation value for each connection. The details are the same as the processing operation of the weighted sum calculation unit 2b in the first embodiment.

Next, the operation processing of the route selection device 5 of FIG. 8 will be described with reference to the flowchart shown in FIG. Note that, here, the portions different from the first embodiment will be mainly described.

First, the parameter changing unit 6 sets the initial values of the weights a ₁ and a ₂ as parameters to the evaluation value calculating unit 7
(Step S20).

The evaluation value calculator 7 calculates an evaluation value for each connection given from the parameter value. The details are the same as the operation processing of the evaluation value calculation unit 2 in the first embodiment (step S21).

Next, the route selection unit 8 performs the route selection process while sequentially selecting the connection from the largest evaluation value. The details are the same as the operation processing of the route selection unit 3 in the first embodiment (step S).
22). At the stage of the end of the first route processing, since the other route selection result for each connection is not yet stored in the best result storage unit 9, the process immediately proceeds from step S23 to step S25 and the route selection result is set to the best result. It is stored in the storage unit 9.

The parameter changing unit 6 determines the parameter value to be used next and outputs it to the evaluation value calculating unit 7 (steps S26 to S27).

The evaluation value calculation unit 7 recalculates the evaluation value again for the same connection as the previous time using the new parameters, and the route selection unit 8 makes the connection in order from the highest evaluation value. While selecting, the route selection process is performed (steps S21 to S22).

Next, this route selection result is compared with the route selection result for each connection already stored in the best result storage unit 9, and the route selection result this time is stored in the best result storage unit 9. When the result is more desirable than the selected route selection result, the stored content of the best result storage unit 9 is updated (steps S23 to S25).

By repeating this, by finally outputting the route selection result stored in the best result storage unit 9 (step S28), the parameter values, that is, the values of the weights a ₁ and a ₂ are set to the same value. It is possible to obtain better results than in the case of using one.

In step S24 of FIG. 10, the route evaluation function is used to compare the newly obtained route selection result with the contents stored in the best result storage unit 9 to determine which is more desirable. .

The route evaluation function calculates (a) the number of connections that could not be assigned a route, (b) the maximum value of the link usage rate, and (c) the product of the link usage amount and the link cost. In consideration of the sum of links, etc., the smaller the value of each of these, the more preferable the route selection result is, that is, the optimum result.

Further, the best result storage unit 9 may store the route selection result itself as a result to be stored, but if the storage capacity is limited, a parameter giving the best result (here, (For example, the values of weights a ₁ and a ₂ )
May be stored, and when the final route selection result is output, the route selection process may be performed using the parameters stored in the best result storage unit 9 and the result may be output.

As a concrete example, in the network shown in FIG. 6, the route selection of each connection shown in FIG. 7 will be performed according to the flowchart of FIG. here,
The value a ₁ of the first weight as a parameter, the value a ₂ of the second weight and a ₂ = 1-a _1, the evaluation value E of the connection to be obtained from the equation (3). Furthermore, as the route evaluation function, a function based on the number of connections for which routes cannot be assigned is used. Parameter value a
₁ is set to "1", "0.5", or "0".

First, the parameter changing unit 6 outputs the value "1" of the weight a ₁ as a parameter to the evaluation value calculating unit 7 (step S20). At this time, the value of the second weight a ₂ becomes “0”, and if the route selection process is performed in that case,
Similarly to the above, the route cannot be assigned to the connection 4, and the route evaluation function becomes "1" (steps S21 to S22).

Since nothing is stored in the best result storage unit 9, the route selection result is stored (steps S23 to S25).

Next, the parameter changing unit 6 outputs the value "0.5" of the weight a ₁ as a parameter to the evaluation value calculating unit 7 (step S20). At this time, the second weight a ₂
Is 0.5, and if a route is selected in this case, routes can be assigned to all the connections, and the route evaluation function becomes 0 (steps S21 to S22). . At that time, the best result storage unit 9
Since the new route selection result is more desirable than the route selection result stored in (4), the contents of the best result storage unit 9 are updated (steps S23 to S25).

Further, the parameter changing unit 6 outputs the value "0" of the weight a ₁ as a parameter to the evaluation value calculating unit 7 (step S20). At this time, the value of the second weight a ₂ becomes “1”, and if the route is selected in this case, the routes can be assigned to all the connections as described above, and the route evaluation function becomes “0”. (Step S21
-Step S22). In this case, since the same route selection result as the result stored in the best result storage unit 9 is obtained,
The best result storage unit 9 is not updated (step S23).
-Step S25).

The route selection result finally output by the above processing is the route selection result when a ₁ = 0.5 and a ₂ = 0.5 (step S 28).

In the above description, the evaluation value calculating unit 7 receives the values of the weights a ₁ and a ₂ as parameters from the parameter changing unit 6 and calculates the evaluation value from the equation (3), but this is not the only case. Instead, it is possible to receive a random number seed from the parameter changing unit 6 and use the random number value as the connection evaluation value.

Further, the maximum value of the random number and the seed of the random number may be received as the parameter, and the random number value may be used as the evaluation value of the connection.

Furthermore, the values of the first weight value a ₁ and the second weight value a ₂ and the seeds of random numbers are received as parameters,
The evaluation value E of the connection may be obtained from E = a ₁ · max (W _up , W _down ) + a ₂ · min (W _up , W _down ) + r (8).

However, r is the value of the random number generated from the seed of the random number. In this case, it is desirable that the random number value r be sufficiently small.

By using not only the bandwidth of the connection but also the random number value when calculating the evaluation value of the connection, the following two effects are obtained.

The first effect is that different evaluation values can be given to connections having the same band. For example,
An example will be described in which the route of each connection as shown in FIG. 11 is selected in the network as shown in FIG.

When the evaluation value of the connection shown in FIG. 11 is obtained from the equation (3) without using the random number value r, the same evaluation value is obtained for all the connections. Therefore, the route selection unit 8 performs the route selection process in ascending order of connection numbers, for example.

In this case, the routes BAE in FIG. 6 are assigned to the connections 1 to 3 in FIG. 11 and the route ACDE is assigned to the connections 4 to 6, resulting in a costly route selection result.

When the evaluation value of each connection is obtained from the equation (8) using the random number value r, there is a possibility that the connections 4 to 6 will have a larger evaluation value than the connections 1 to 3.
Therefore, there is a possibility that the connections 4 to 6 will be set before the connections 1 to 3. Thereby, a better route selection result is obtained.

The second obtained by using the random value r
The effect of is that when a route is selected using a plurality of parameter values and the best route selection result is made the final result, a random number value r
If is not used, the same route selection result is likely to be obtained for each parameter value. On the other hand, the random number r
With, different route selection results can be obtained for each parameter value, and better final results are obtained.

As described above, the use of the random number r can be expected to improve the route selection performance.

Further, the random number value r shows the effect when there are many connections to be routed which have the same band, but in other cases, the random number r does not significantly affect the evaluation value calculation using the band. Therefore, it is desirable to set the random value r to a small value.

As described above, according to the first and second embodiments, a numerical value called an evaluation value is obtained for the bidirectional connection, and the bidirectional connection route is selected using this numerical value. Therefore, optimal route selection can be efficiently performed under the condition that the upstream connection and the downstream connection belonging to the same bidirectional connection pass through the same route.

Further, according to the first embodiment, the evaluation value calculation unit 2 of the route selection device 1 determines the bandwidth of the upstream connection and the bandwidth of the downstream connection forming the bidirectional pair for each bidirectional connection. One evaluation value is calculated by using the evaluation value, and the route selection unit 3 determines a route based on this evaluation value as if the upstream connection and the downstream connection forming a bidirectional pair are one connection.
The same route can be assigned to the upstream connection and the downstream connection forming a bidirectional pair.

Further, the evaluation value calculation unit 2 compares, for each bidirectional connection, the band of the upstream connection and the band of the downstream connection forming the bidirectional pair.
An evaluation value of the scalar amount is obtained by adding a value obtained by multiplying the value of the band determined to be large by the first weight and a value obtained by multiplying the value of the band determined to be small by the second weight, The route selection unit 3 determines the order in which routes of bidirectional connections are determined in descending order of evaluation value.
At this time, since the evaluation value is a scalar quantity, it is possible to easily compare the evaluation values.

Furthermore, the route selecting unit 3 can determine the route of the bidirectional connection in the ascending order of the evaluation value by the shortest route (that is, the route of the minimum cost), and therefore the processing can be performed at high speed.

Further, according to the second embodiment, a plurality of parameter values are prepared in the parameter changing unit 6 of the route selecting device 5, and the evaluation value calculating unit 7 uses each value for each parameter value. The evaluation value of the two-way connection is calculated, the route is selected by the route selection unit 8 based on this evaluation value, and the route is repeatedly selected using the values of the plurality of parameters. As a result, the best one is found. By storing the value of the parameter that gives the best result in the result storage unit 9, for example, a route more suitable than the case of using one parameter value as in the first embodiment. The selection result is obtained.

Of course, the best result storage unit 9 does not store the value of the parameter giving the best result, but stores the route of each bidirectional connection when the route is selected using the value of the parameter giving the best result. It may be stored.

Next, the third embodiment will be described.
Here, a case where a route is selected for a one-way connection will be described.

FIG. 12 shows the configuration of the route selection device 100 according to the third embodiment. Route selection device 1
00 includes a route search unit 101, a band reservation unit 102, and a route determination unit 103.

The route search unit 101 includes network information including information sufficient for modeling a network as a route selection target, such as network topology and link capacity, a source / destination node of each connection, a band, an allowable maximum number of hops, Connection information including information such as the pros and cons of a route is input, all the routes that meet the request are obtained for each connection in the target network, and the routes are output to the bandwidth reservation unit 102 as the route information. There is.

For example, when the bandwidth of the connection is larger than the link capacity, the route search unit 101 ignores the route passing through the link. Further, when the connection has a required value for the delay in information transmission, a route having a delay larger than the requested delay amount is ignored. In this way, the route search unit 1 determines routes that should not be assigned or may not be assigned to each connection.
By ignoring 01, the amount of calculation required for the subsequent processing can be reduced.

The band reserving unit 102 uses the route information obtained by the route searching unit 101 to reserve the band of each connection for a link on a route that the connection can take. When two or more routes share the same link among the routes that a connection can take, it is possible to prohibit more reservations for the link that reserved the bandwidth once, but the bandwidth of the same connection can be repeatedly used. You may repeat the reservation.

The bandwidth reservation method of the bandwidth reservation unit 102 will be described with reference to FIG. In FIG. 13, the node devices A to E and links 1 to 5 as transmission lines between the nodes.
Shows a network configured by connecting with each other.
Link 1 is provided from node A to node B, link 2 is provided from node B to node C, link 3 is provided from node C to node D, and link 4 is provided from node B to node E. A link 5 is provided from the node E to the node D.

For example, consider a case where the source node A establishes one connection of the destination node D in the network of FIG. The route that this connection can take is route A →
There are two routes, B → C → D and route A → B → E → D,
Link 1 is shared by two routes. At this time, the band reserving unit 102 may doubly reserve the band of the connection for the link 1, but once the band is reserved, the subsequent reservation is prohibited and the band for one connection may be reserved. Good.

The route determination unit 103 determines the route of each connection using the bandwidth reservation information obtained by the bandwidth reservation unit 102 and outputs the result (route selection result).

FIG. 14 shows a configuration example of the route determination unit 103 of the route selection device 100 according to the third embodiment. The route determination unit 103 uses the first overflow band calculation unit 11
0, an evaluation value calculation unit 111, and a route selection unit 112.

The first overflow bandwidth calculation unit 110 uses the bandwidth reservation information output from the bandwidth reservation unit 102 and the network information such as the link capacity to determine the first overflow bandwidth which is the amount by which the reserved bandwidth overflows from each link. Is to ask.

The evaluation value calculation unit 111 calculates the first overflow band calculated by the first overflow band calculation unit 110, network information such as network topology and link capacity, and the origin / destination node / band / allowable maximum of each connection. Number of hops
An evaluation value for a route for setting each connection is obtained by using connection information such as the preference of the route.

For example, the evaluation value calculation unit 111 defines the cost of each link by multiplying the amount of the first overflow band of each link by the constant defined for each link, and the cost of the route is calculated on the route. Calculated as the sum of the link costs of
The evaluation value is obtained by dividing the difference between the delay amount required by the connection and the delay amount guaranteed by the route by the cost of the route. That is, assuming that the evaluation value for setting the connection k to the route 1 is V _kl , V _kl is obtained as follows.

[0145]

[Equation 3]The route selection unit 112 uses the evaluation value V_klIn descending order of a (>
1) Allocate a route to a book connection. That is,
The path selection unit 112 uses the connection that maximizes Expression (9).
Let m be the route and i be the route.
Assignment, then maximize equation (9) except connection m
Connection m'and route i'to
Assign path i'to m '. go through a connections
After allocating the route, the reserved bandwidth of the set connection
Removes the link that is not on the path,
Recalculate the range values. Route all connections to the same
It is decided by repeating the same operation.

Further, the route selection unit 112 may select a ( > 1) connections in descending order of bandwidth and allocate a route having the maximum evaluation value to each connection. Further, the route selection unit 112 may select a ( > 1) connections in descending order of priority and select the route having the largest evaluation value for each connection. Further, the evaluation value calculation unit 111 calculates the cost by using a predetermined constant instead of the value of the first overflow band for the link in which the value of the first overflow band is “0”. Good.

In the third embodiment, the evaluation value is calculated and the connection paths are determined in descending order of the evaluation value. However, instead of the evaluation value, the cost may be calculated and the connection paths may be determined in the ascending order. It is possible.

FIG. 15 is a flow chart for explaining the operation processing of the route selection device 100 according to the third embodiment.

In FIG. 15, first, the route search unit 101.
Searches all the routes that each connection can take (step S50) and outputs it to the bandwidth reservation unit 102 as route information. The bandwidth reservation unit 102 reserves the bandwidth of each connection for a link on its possible route (step S5).
1), the first overflow bandwidth calculation unit 110 includes the bandwidth reservation unit 10
Using the bandwidth reservation information output from 2 and network information such as link capacity, the first overflow bandwidth, which is the amount by which the reserved bandwidth overflows from each link, is obtained (step S52), and the evaluation value calculation means 111 uses the first overflow bandwidth. Based on the first overflow band and the like obtained by the overflow band calculation unit 110, the evaluation value is calculated from the equation (9) (step S53), and the route selection unit 112
Routes are assigned to a ( > 1) connections in descending order of evaluation value V _kl (step S54). If there is a connection whose route has not been decided (step S55), the reserved band of the set connection is removed from the link not on the route, and the value of the first overflow band is recalculated (step S56).

Regarding the network as shown in FIG. 16,
The route selection processing shown in the flowchart of FIG. 15 will be specifically described by taking the case of selecting the route of each connection shown in FIG. 17 as an example.

FIG. 16 shows a network constructed by connecting the node devices A to E and the nodes by links 1 to 5 as transmission lines. Link 1 is provided from node A to node B, link 2 is provided from node B to node C, link 3 is provided from node A to node D, and link 4 is provided from node D to node C. A link 5 is provided from the node D to the node E.

For simplicity of explanation, it is assumed that the cost of the link is defined by the amount of the first overflow band, and the requirement regarding the delay of the connection is not taken into consideration. Further, instead of selecting the connection and its route in the descending order of the evaluation value, the connection and its route are selected in the ascending order of cost.
The capacity of each link is set to 150 Mbps.

First, each connection i shown in FIG.
Obtain possible routes for (i = 1 to 4) (step S5)
0). The routes that can be taken by the connections 1 and 2 in which the source node is A and the destination node is C are route A → B → C and route A → D →
The route that can be taken by the connection 3 whose source node is C and whose destination node is A and whose destination node is E is route A → D → E, and the route that can be taken by the connection 4 whose source node is D and whose destination node is E is route D → E.

Next, the bandwidth of the connection is reserved for each link (step S51). The resulting bandwidth reservation status is shown in FIG. That is, with regard to the connection 1 in FIG. 17, the possible routes are the route ABC and the route ADC.
bps is reserved. With regard to the connection 1 in FIG. 17, the possible routes are the route ABC and the route ADC, so that 10 Mbps is reserved for each of the links 1 to 4. With regard to the connection 3 in FIG. 17, the possible route is the route ADE, so 145 Mbps is reserved for the link 3 and the link 5, respectively. With regard to the connection 4 in FIG. 17, the possible route is the route DE, so 5 Mbps is reserved for the link 5.

Next, the first overflow band in each link is calculated (step S52). From the reservation status of each link shown in FIG. 18, links 1 to 5 are
20Mbps, 20Mbps, 165Mbps, 20M
It can be seen that the bandwidths of bps and 150 Mbps are reserved. Now, here, the amount of the overflow band of the link m (m = 1 to 5) is represented by ξ _m , and the first overflow vector O =
When {ξ ₁ , ξ ₂ , ..., ξ ₅ } is defined, since the capacity of each link is 150 Mbps, O = {0,0,15,0,0} as apparent from FIG. . That is, in this example, the overflow band is a value obtained by subtracting the link capacity from the total value of the reserved bands.

However, the combination of reserved bandwidths that can actually be accommodated in the link 3 is {connection 1, connection 2} or {connection 3}. for that reason,
First in the sense of the sum of bands that cannot be accommodated on a link
The overflow bandwidth of 145 Mbps or 20 Mbps (=
10 + 10). In other words, the knapsack problem is solved from the link capacity and each reserved band, and the sum of the reserved bands that did not enter the link
It may be an overflow band.

Next, an evaluation value for each connection is obtained based on this first overflow band (step S5).
3). Here, the link cost is defined by the amount of the first overflow band, and the connection and its route are selected in the ascending order of cost.

That is, in step S54, a connections are set to the route in ascending order of cost.
Here, if the route A → B → C is set in the connection 1 in which the first overflow band is recalculated in step S56 every time one connection is established with a = 1, the cost of the route is " If the route A → D → C is set, the cost of the route is “15”. If the route A → B → C is set for the connection 2, the cost of the route is “0”, and if the route A → D → C is set, the cost of the route is “15”. If the route A → D → E is set to the connection 3, the cost of the route is “15”, and the route D → D to the connection 4 is set.
If E is set, the cost of the route is “0”.

The path with the lowest cost is searched from among these. There are three routes with cost "0", so choose one of them,
It is assumed that the route A->B-> C is set in the connection 1. Then, in step S56, the first overflow band is recalculated. At this time, the bandwidth of the connection 1 reserved for the routes A → D → C is removed, and the link capacity on the routes A → B → C is reduced by the bandwidth of the connection 1. As a result, the first overflow band O becomes O = {0,0,5,0,0}. Similarly, the route A → B → C is assigned to the connection 2. Then, the first overflow band is recalculated, and O =
Get {0,0,0,0,0}. By repeating the same process, the route A → D → E is set in the connection 3 and the route D → E is set in the connection 4.

Here, the costs of all the routes of all the connections are once calculated, and the route with the minimum cost is selected. However, the connection with the maximum bandwidth is first selected, and the selected connection is set as the minimum cost route. You may do it.

Next, a fourth embodiment will be described.

In the route selection device 1 according to the fourth embodiment, the configuration of the route determination unit 103 is different from that of the third embodiment.

FIG. 19 shows another configuration example of the route determination unit 103 of FIG. 12, in which the second overflow band calculation unit 12 is used.
0, an evaluation value calculation unit 121, and a route selection unit 122.

The second overflow bandwidth calculation unit 120 assumes that a connection has been set up on a certain path, and if the reserved bandwidth of the connection is removed from a link other than the path, the amount of the reserved bandwidth overflows from the link. A second overflow band is determined for all routes of all connections.

The evaluation value calculation unit 121 defines the cost for all routes of all connections by the weighted sum of the second overflow bandwidth and the constant defined for each link. The evaluation value calculation unit 121 obtains an evaluation value by the reciprocal of this cost.

The route selection unit 122 determines at least one connection as a route in descending order of evaluation value.
The reserved bandwidth of the connection whose route is determined is removed from the links other than the route. After that, this operation is repeated until the routes of all the connections are determined.

The evaluation value calculation unit 121 defines a gain obtained when a certain connection passes through a route instead of obtaining the evaluation value by the reciprocal of the cost, and obtains the evaluation value by dividing this gain by the cost. May be. When the value of the second overflow band is “0” for all links, the evaluation value calculation unit 121 may use a constant given in advance as the cost of the route.

FIG. 20 is a flow chart for explaining the operation processing of the route selection device 100 according to the fourth embodiment.

First, the route search unit 101 searches for all possible routes of each connection (step S60),
The result is output to the bandwidth reservation unit 102 as route information.

The bandwidth reservation unit 102 reserves the bandwidth of each connection for a link on the possible route (step S
61), the second overflow bandwidth calculation unit 120 assumes that a connection has been set up on a certain path, and if the reserved bandwidth of the connection is removed from a link other than the path, the amount of overflow of the reserved bandwidth from the link A certain second overflow band is obtained for all routes of all connections (step S62).

The evaluation value calculation unit 121 calculates an evaluation value using the second overflow band (step S63), and the route selection unit 122 selects a connection route in descending order of the calculated evaluation value. (Step S64) Until the routes of all the connections are determined (step S65), the bandwidth reservation information is changed (step S66), and the processes of steps S62 to S65 are repeated.

Regarding the network as shown in FIG. 16,
The route selection process shown in the flowchart of FIG. 20 will be specifically described by taking the case of selecting the route of each connection shown in FIG. 17 as an example.

First, each connection i shown in FIG.
A possible route for (i = 1 to 4) is calculated (step S6).
0), the bandwidth of the connection is reserved for each link (step S61). The resulting bandwidth reservation status is shown in FIG. Further, FIG. 21 shows possible routes searched at this time. That is, in FIG.
A route number 1 is attached to the route A → B → C, and a route A → D →
The route number 2 is attached to C, the route number 3 is attached to the routes A → D → E, and the route number 4 is attached to the routes D → E.

Next, the second overflow band calculation unit 121
The overflow band of is calculated (step S62). Here, the second overflow band on the assumption that the connection k is set to the route of the route number l will be described as O _kl . Assuming that the connection 1 of FIG. 17 is set as the route of the route number 1 of FIG. 21, the reserved band of the connection 1 is removed from the band on the route of the route number 2 and the second overflow band is obtained. O ₁₁ = { 0,0,5,0,0}.

Assuming that the connection 1 is set as the route of the route number 2, the reserved band of the connection 1 is removed from the band on the route of the route number 1, and the second overflow band is obtained. O ₁₂ = {0,0 , 15, 0, 0}.

Similarly for connection 2, O ₂₁ =
{0,0,5,0,0}, O ₂₂ = {0,0,15,0,
0 °.

Assuming that the route with the route number 3 is set in the connection 3, there is no other route that the connection 3 can take, so that there is no reserved band to be removed, and the second overflow band is O ₃₃ = {0,0. , 15, 0, 0}.

Similarly for the connection 4, when the second overflow band is obtained, O ₄₄ = {0,0,15,0,0}.

Next, the evaluation value V _{kl is calculated} from the second overflow band O _kl.
Is calculated by the following equation (step S63).

[0180]

(Equation 4) However, C _kl represents the gain obtained when connecting the connection k to the route l. Here, for simplicity, C _kl = (bandwidth of connection k). Of course, C _kl may take different values for each connection and each route.

When the evaluation values calculated using the equation (10) are listed, V ₁₁ = 10/5, V ₁₂ = 10/15, V ₂₁ =
10/5, V ₂₂ = 10/15, V ₃₃ = 145/15, V
₄₄ = 5/15.

Next, in the process of step S64 in the flowchart of FIG. 20, assuming that a = 2, connection 1 is the route of route number 1 and connection 3 is the route number 3
The connection will be set to the route.

Next, the bandwidth reservation information is changed (step S66). That is, the reserved bandwidth of the connection 1 is removed from the links (link 3, link 4) other than the route of the route number 1.

Then, returning to step S62, the second overflow band is obtained for the connection whose route is not determined. Then, O ₂₁ = {0,0,0,0,0}, O ₂₂ =
{0,0,5,0,0}, O ₄₄ = {0,0,5,0,
0 °. When O _kl = {0,0,0,0,0}, the evaluation value cannot be obtained using the equation (10),
The evaluation value is determined by V _kl = C _kl (step S6)
3). As a result, V ₂₁ = 10, V ₂₂ = _10/5 , V ₄₄ =
It becomes 5/5, and connection 2 is the route with route number 2.
The connection 4 is set to the route with the route number 4 (step S64).

In this example, the second overflow bandwidth is calculated for all routes of all connections, and the evaluation value is calculated. However, in order to reduce the amount of calculation, at least one connection is arranged in descending order of gain. You may select a connection and calculate the evaluation value only for the selected connection.

Next, the fifth embodiment will be described.

In the route selection device 1 according to the fifth embodiment, the configuration of the route determination unit 103 is different from that of the third embodiment.

FIG. 22 shows still another configuration example of the route determination unit 103 of FIG. 12, in which the first overflow band calculation unit 130, the second overflow band calculation unit 131, and the evaluation value calculation unit 1
32 and a route selection unit 133.

The first overflow bandwidth calculation unit 130 uses the bandwidth reservation information and the network information such as the link capacity to determine the first overflow bandwidth, which is the amount by which the reserved bandwidth overflows from each link.

The second overflow bandwidth calculating unit 131 removes the reserved bandwidth of the connection from the link which is not on the path, assuming that the connection has passed through a certain path for all the paths of all the connections. In this case, the second overflow band, which is the amount by which the reserved band overflows from the link, is obtained.

The evaluation value calculation section 132 is adapted to calculate an evaluation value for each connection and its route from the first overflow band, the second overflow band, the network information and the connection information.

The route selection unit 133 has an evaluation value calculation unit 132.
The route is set in the connection in accordance with the evaluation value calculated in.

FIG. 23 is a flow chart for explaining the operation processing of the route selection device 100 according to the fifth embodiment.

First, the route search unit 101 searches all routes that each connection can take (step S70), and outputs the result to the band reservation unit 102 as route information.

The bandwidth reservation unit 102 reserves the bandwidth of each connection for a link on the possible route (step S
71), the first overflow bandwidth calculation unit 130 is
02 using the bandwidth reservation information and the network information such as the link capacity to obtain the first overflow bandwidth, which is the amount by which the reserved bandwidth overflows from each link (step S72), and the second overflow bandwidth calculation unit. The 131 selects one connection for which the route has not been determined and has not yet obtained the second overflow band (step S73), and obtains the second overflow band for each route through which the connection can pass (step S73).
74).

When the second overflow bandwidth has been calculated for all routes whose routes have not been decided yet (step S75),
The evaluation value calculation unit 132 sets the evaluation values for all the routes of all the connections whose routes have not been determined to the first overflow band and the second overflow band.
And the overflow band of (step S76).

The route selection unit 133 obtains the connection and route for which the calculated evaluation value is the maximum and sets the connection to the route (step S77). When there is a connection whose route has not been determined, step S7 is executed again.
Returning to step 2, the processes of steps S72 to S78 are repeated until the routes are set for all the connections.

In the network as shown in FIG. 16, the route selection processing shown in the flowchart of FIG. 23 will be specifically described by taking the case of selecting the route of each connection shown in FIG. 17 as an example.

First, each connection i shown in FIG.
A possible route for (i = 1 to 4) is calculated (step S7).
0), the bandwidth of the connection is reserved for each link (step S71). The resulting bandwidth reservation status is shown in FIG.

At this time, when the first overflow band is obtained, O =
It becomes {0,0,15,0,0} (step S72).

Next, the second overflow band is obtained (step S73). For simplicity of description, the possible paths searched are shown in FIG. Further, the second overflow band on the assumption that the connection k is set to the route 1 will be described as O _kl .

Assuming that the connection 1 is set as the route of the route number 1, the reserved band of the connection 1 on the route of the route number 2 is removed, and the second overflow band is obtained, O ₁₁ =
It becomes {0,0,5,0,0}.

[0203] Assume that setting the connection 1 in the path of the path number 2, remove the reserved bandwidth of the connection 1 on the path of the path number 1, when obtaining the second overflow band O ₁₂ =
It becomes {0,0,15,0,0}.

Similarly for connection 2, O ₂₁ =
{0,0,5,0,0}, O ₂₂ = {0,0,15,0,
0 °.

Assuming that the route with the route number 3 is set for the connection 3, there is no other route that the connection 3 can take, so that there is no reserved band to be removed, and the second overflow band is O ₃₃ = {0,0. , 15, 0, 0}.

Similarly for the connection 4, if the second overflow band is obtained, O ₄₄ = {0,0,15,0,0}.

Next, in step S76, evaluation values for all routes of each connection are obtained. If you can write an evaluation value when you set the connection k in the path l and V _kl, V _{kl is, V kl = C kl (O} · (O-O kl)) ... can be calculated (11). However, O · O ′ represents the inner product of the vector O and the vector O ′, and C _kl is the route k through the connection k.
This is the gain obtained when setting to. Also, the evaluation value V _kl
May be calculated as V _kl = (O · (O−O _kl )) / K _kl (12) using the cost K _kl required to determine the connection k as the route 1.

Here, the cost K _kl is defined as the product of the bandwidth of the connection k and the number of hops of the route, and the evaluation value is calculated by the equation (12). Then, V ₁₁ = (15 × 10) / (2 × 10) = 15/2 V ₁₂ = 0 / (2 × 10) = 0

Similarly for the connections 2, 3 and 4, V ₂₁ = (15 × 10) / (2 × 10) = 15/2 V ₂₂ = 0 / (2 × 10) = 0 V ₃₃ = 0 / (2 × 145) = 0 V ₄₄ = _0/5 = 0.

Next, proceeding to step S77, when the maximum evaluation value is searched for, it is found that V ₁₁ = V ₂₁ = 15/2. Therefore, the route with the route number 1 is set in the connection 1.

Further, in step S72, the first overflow band is recalculated. That is, the reserved bandwidth of the connection 1 reserved for the link on the route of the route number 2 is removed, and the capacity of the link on the route 1 is reduced by the bandwidth of the connection 1. After that, when the amount of reserved bandwidth overflowing from each link is obtained, O = {0,0,5,0,0}.

When the second overflow band is obtained in the same manner, O
_{21 = {0,0,0,0,0}, O 22} = {0,0,5,
0,0}, O ₃₃ = {0,0,5,0,0}, O ₄₄ =
It becomes {0, 0, 5, 0, 0} (step S73 to step S74).

When the evaluation value is calculated, V ₂₁ = 5/20, V ₂₂
= 0 and V ₄₄ = 0 (step S76), and the maximum evaluation value is V ₂₁ , so the route of route number 1 is set for connection 2.

Thereafter, by repeating the same process, the route with the route number 3 is set to the connection 3 and the route with the route number 4 is set to the connection 4.

Here, the first overflow bandwidth is calculated every time one connection is established. However, in order to reduce the calculation amount, the first overflow bandwidth is maintained until a (> 1) connections are established. The first overflow band may be recalculated each time a connections are established without recalculating the band. Also, here, the evaluation value is calculated for all routes of all connections, but in order to reduce the calculation amount, the connection that obtains the evaluation value from information such as connection priority and bandwidth is limited. You may.

For example, consider the network shown in FIG. FIG. 24 shows a case where a link 6 is further provided between the nodes A and D and two links (link 3 and link 6) are provided between the nodes in the network shown in FIG. There is. At this time, the link 3 and the link 6 are logically regarded as one link, and the bandwidth is reserved. Then, when obtaining the first or second overflow bandwidth, the overflow bandwidth can be obtained by subtracting the sum of the link capacities from the sum of the bandwidth of the connection for which the route has already been determined and the reserved bandwidth. Alternatively, the overflow bandwidth may be obtained by solving the knapsack problem under the condition that the bandwidth of the connection whose route has already been determined is always accommodated in some link.

As described above, according to the third to fifth embodiments, the route searching unit 101 searches for all the routes that each connection can take, and the band reserving unit 102 routes the band of each connection. The route determination unit 103 reserves the link on the route obtained by the search unit 101,
By determining the route of the connection by using this reservation status as information for route selection, it becomes possible to solve the route selection problem with high approximation accuracy. That is,
For each connection, reserve the bandwidth for all possible routes of the connection, find the amount of reserved bandwidth overflowing from each link called overflow bandwidth, and determine the connection path to reduce this overflow bandwidth as much as possible. Without excessively balancing the link load, so you can
An accurate approximate solution of the route selection problem can be obtained.

Further, according to the third embodiment, the route determining unit 1 determines the amount of the first overflow band, which is the band that cannot be accommodated in each link among the bands reserved for each link.
By using 03, it is possible to know the amount of band that may overflow from each link, and by determining the route of the connection so as to reduce this first overflow band, route selection can be performed with high approximation accuracy. It becomes possible to solve the problem.

Further, according to the fourth embodiment, it is assumed that a certain connection has passed through a certain route, and if the reserved bandwidth of the connection is removed from the link not on the certain route, the connection is accommodated in each link. By using the second overflow band, which is the amount of the band that may not be possible, by the route determination unit 103, it is possible to know the amount of the overflow band when a certain connection passes through a certain route, and therefore the most overflow band. You can know the connection and its route that can reduce This makes it possible to solve the route selection problem with high approximation accuracy.

Also, according to the fifth embodiment, the first
By using the amount of the overflow band and the amount of the second overflow band by the route determination unit 103, it is possible to know the amount of the overflow band that decreases when a certain connection goes through a certain route. The route of the connection can be determined so that the amount of the overflow band is "0". Therefore, it becomes possible to solve the route selection problem with high approximation accuracy.

Up to this point, the case has been described in which the optimum route of the connection is selected in consideration of the capacity of the link set between the nodes, the cost, etc., for the given network.

Next, for example, when the traffic information and the network topology information are given in advance, an optimal link is established between the desired nodes in consideration of the bandwidth, cost, etc. based on them. A communication network designing method for designing a network and a communication network designing apparatus using the communication network designing method will be described.

First, the sixth embodiment will be described.

FIG. 25 shows the configuration of the communication network designing apparatus 200 according to the sixth embodiment.

In FIG. 25, the communication network designing apparatus 200
Is an initial topology setting unit 201 and a topology changing unit 2
02, an initial route setting unit 203, a route changing unit 204, and an edge cost calculating unit 205.

The initial topology setting unit 201 uses the network information such as the number of nodes and the types of links that can be set between the nodes, and the traffic information such as the originator / destination node of the communication flow, the bandwidth, and the badness of the communication path. , The initial topology of the communication network is calculated.

The topology changing unit 202 finds a new topology by using the topology found by the initial topology setting unit 201 or the topology found by the topology changing unit 202 previously.

The initial route setting unit 203 determines the route of the requested flow on the topology determined by the initial topology setting unit 201 or the topology changing unit 202.

The edge cost calculation unit 205 is adapted to obtain the cost of the edge from the amount of flow flowing on the edge for each edge.

The route changing unit 204 is adapted to change the flow route by using the edge cost calculated by the edge cost calculating unit 205. When the flow path is changed, the amount of flow flowing through each edge also changes, so the edge cost calculation unit 205 calculates the edge cost again. After that, the route changing unit 204
Changes the flow route using the newly obtained edge cost. By repeating this, the flow path is determined.

After the flow route is determined, the topology changing unit 202 changes the currently used topology and sends the new topology to the initial route setting unit 203.
In the same way, the flow route is assigned to this new topology. By thus repeatedly changing the topology and setting the route, the communication network can be designed.

Next, the operation processing of the communication network designing apparatus 200 of FIG. 25 will be described with reference to the flowchart shown in FIG.

First, the initial topology setting unit 201 finds the initial topology (step S100), and the initial route setting unit 203 finds the initial route of the flow in the initial topology (step S101).

Then, in the edge cost calculation unit 205,
The edge cost and the net cost are obtained (steps S102 and S103), and the condition is determined in step S104. At this stage, since only the initial route of the flow and the initial cost of the network have been obtained, unconditionally step S105.
Proceed to.

In step S105, the route changing means 50
The flow path is changed by 0, and then step S102,
In step S103, the edge cost calculation unit 205 calculates the edge cost and the net cost.

Then, the process proceeds to step S104, and if the network cost is lower than the last time, step S105.
Go to and repeat the change of the flow route. If the network cost has increased from the previous time, the process proceeds to step S106, and the flow route is returned to the previous route.

Next, the process proceeds to the condition judgment of step S107, but at this stage, only the network cost using the initial topology has been obtained, so unconditionally step S10.
8, the process of steps S101 to S106, that is, the flow route and the network cost are obtained using the new topology provided by the topology changing unit 202. If the newly obtained network cost is smaller than the network cost when the previous topology is used (step S107), the process proceeds to step S108, and the topology is changed to obtain the flow route as described above. If the newly obtained network cost is higher than the network cost using the previous topology (step S10)
7) Proceed to step S109, and output the communication network design information obtained by using the previous topology.

As the contents of the communication network design information, for example,
There are edges (communications network topology) that are provided between nodes given in advance, types of links, flow paths, and the like.

Next, the procedure for designing the communication network will be specifically described with reference to the flowchart shown in FIG.

The initial topology setting unit 201 obtains the initial topology from the network information and the traffic information as shown in step S100 of FIG.

For example, if the communication network to be designed is node A,
Assuming that the network information is composed of B, C, and D, the network information includes, for example, as shown in FIG. 27, the capacity of the link that can be set between these nodes and the cost required therefor. That is, in the network information shown in FIG. 27, the number of nodes is “4”, and the types of links that can be set between nodes are links AB, AC, AD, BA, and B.
It includes C, BD, CA, CB, CD, DA, DB, DC, the capacity for each of these links, and the cost for that capacity.

[0242] Further, as the traffic information, for example,
It is assumed that a demand of flowing a flow of 50 Mbps (flow 1) from the node A to the node B and a flow of 50 Mbps (flow 2) from the node A to the node D is given.

FIG. 28 shows a specific example of the initial topology obtained by the initial topology setting unit 201 based on such network information and traffic information.

In FIG. 28, each node A, B, C, D
A line drawn between is called an edge, and a link can be set between nodes only when an edge is set between nodes. In FIG. 28, the full mesh topology is used as the initial topology, but it is not necessary to use the full mesh topology as the initial topology when there is an edge that is not desired to be set from the network information or traffic information.

Next, proceeding to step S101 in FIG.
The initial route setting unit 203 obtains the initial route of the flow for the initial topology shown in FIG. Then, using the amount of flow flowing through the edge, the edge cost calculation unit 205
Calculates the edge cost and the network cost (step S102,
Step S103), using these, the route changing unit 204
Changes the flow route (step S105).

The edge cost calculation unit 205 recalculates the edge cost based on the route of the new flow (step S
102 to step S103), the route changing unit 204 changes the route of the flow using the recalculated edge cost (steps S104 to S105).

By repeating this route change, the flow route is determined. A concrete example of this route determination process is shown in FIG. Note that in FIG. 29, the numbers in parentheses indicate edge costs.

First, as shown in FIG. 29A, a cost is assigned to each edge of the initial topology. Here, the link with the smallest capacity is selected from the links that can be set for each edge, and the cost per unit capacity of the link is taken as the edge cost. Then, the cost of the edge is used to find the minimum cost route, which is used as the initial route. as a result,
As the initial flow route, route A → B is given to flow 1, and route A → D is given to flow 2.

Next, in step S102, the edge cost calculation unit 205 obtains a new edge cost. In FIG. 29B, for an edge where the total amount of passing flows is larger than “0”, a link whose link capacity is larger than the total amount of flows and has the smallest cost is selected and its cost is selected. As in the case of determining the initial route, the cost per unit capacity of the link having the minimum capacity is given to the edge of which the total flow amount is “0”. That is,
In FIG. 29 (b), the capacitance 6 is added to the edge AB (flow 1).
A link with a cost of "12 (10,000 yen)" is given at 0 Mbps, and a link with a capacity of 60 Mbps and a cost of "12 (10,000 yen)" is given to the edge AD (flow 2).

Next, in step S103, the route changing unit 204 determines the network cost. In FIG. 29B, the sum of the costs of the edges through which the flow passes is calculated, and as a result, the network cost becomes “24”.

Next, in step S104, the network costs are compared between the case of using the previous route and the case of using the new route, but this time there is no previous route, so there is no condition. Proceeding to step S105, the route changing unit 2
04 changes the flow route.

For example, in FIG. 29C, the minimum cost route is assigned to the flow by using the edge cost. Specifically, by obtaining the minimum cost route using the edge cost of FIG. 29B, the route A → C → B is obtained as the route after the change of the flow 1, and the route after the change of the flow 2 is obtained. The route A → C → D is required. FIG. 29C shows a state in which the routes of the flow 1 and the flow 2 are changed.

In FIG. 29 (c), the flow 1 and the flow 2 are all changed to a new route, but Leonard Kleinrock
Queueing Systems (vol 2, pp. 270-42
1, John Wiley & Sons, 1976), a part of the flow may be changed to a new route as in the FD method.
In addition, as in FIXED ROUTING by Kleinrock, change flow 1 to a new route, and if the network cost does not decrease, return the route of flow 1 to the original route, and then change the route of flow 2 to a new route. However, if the network cost does not decrease, the route of the flow 2 may be restored so that the route of the flow 2 is restored.

Next, in step S102, the edge cost is obtained again. In FIG. 29 (c), the capacity 100 Mbp is added to the route A → C common to the routes of the flow 1 and the flow 2.
s, the link of cost "16" is given, and the route C of the flow 1
A link having a capacity of 60 Mbps and a cost of "3" is given to B, and a link having a capacity of 60 Mbps and a cost of "3" is given to the route C → D of the flow 2.

The process proceeds to step S103, and the network cost is calculated as "22", which is smaller than the previous time (step S104). Step S1
Go to 05 and change the flow route.

Then, since the flow route becomes as shown in FIG. 29B, the network cost increases, and this time, in step S1.
From 04, the process proceeds to step S106 to return the flow path to the state before the change. That is, the flow path is returned to the state shown in FIG.

Next, the condition is determined in step S107, but since the initial topology is used, the process unconditionally advances to step S108. In step S108, the topology changing unit 202 changes the topology, and using the new topology, the initial route setting unit 203, the edge cost calculating unit 205, and the route changing unit 204 perform the processes of the above-described steps S101 to S106. Done,
Determine the flow path.

In step S107, the route changing unit 204
Is compared to the network cost when using the previous topology,
If the network cost using the current topology has decreased, the process proceeds to step S108, and if not, step S1.
In step 09, the communication network design information such as the type of link set between the route nodes of the flow when the previous topology is used is output.

In the description of the route determination process of FIG. 29, the edge cost is calculated from the flow flowing through each edge, and the route of the flow is changed to the minimum cost route. Instead, the edge cost calculation unit 205 selects the edge cost 1 which is the edge cost obtained from the flows flowing through each edge, and the edge cost when it is assumed that the selected flow is added to each link. The edge cost 2 is calculated. The route changing unit 204 obtains the increased cost, which is the difference between the edge cost 2 and the edge cost 1, for each link, and changes the route of the selected flow to the route having the smallest sum of the increased costs of the links on the route. Try. The flow route may be changed by performing this operation for all the flows.

Also in the other embodiments of the communication network designing apparatus to be described later, even if the destination of the flow route is changed, the edge cost is obtained from the current flow amount and the route with the smallest sum of the edge costs is selected. Alternatively, it is possible to obtain an edge cost that increases as a flow for which a route change is attempted is newly added to each link and select a route having the smallest sum of the increased costs.

Next, the seventh embodiment will be described.

FIG. 30 shows the configuration of the communication network designing apparatus 300 according to the seventh embodiment.

In FIG. 30, the communication network designing apparatus 300
Is an initial topology setting unit 301 and a topology changing unit 3
02, an initial route setting unit 303, a route changing unit 304, an edge cost calculating unit 305, and a best result storage unit 306.

The initial topology setting unit 301 uses the network information such as the number of nodes and the types of links that can be set between the nodes, and the traffic information such as the originator / destination node of the communication flow, the bandwidth, and the good / bad of the communication route. , The initial topology of the communication network is calculated.

The topology changing unit 302 obtains a new topology using the topology obtained by the initial topology setting unit 301 or the topology obtained previously by the topology changing unit 202.

The initial route setting unit 303 is adapted to obtain the route of the requested flow on the topology obtained by the initial topology setting unit 301 or the topology changing means 302.

The edge cost calculation unit 305 is adapted to obtain the cost of an edge from the amount of flow flowing on the edge for each edge.

Next, the operation processing of the communication network designing apparatus 300 of FIG. 30 will be described with reference to the flowchart shown in FIG.

First, in step S120, for example,
The initial value “0” is stored as the number of times h the topology is previously stored in the topology storage unit 310 of the topology changing unit 302. Then, proceeding to step S121, the initial topology setting unit 301 obtains the initial topology, and then proceeds to step S122, where the initial route setting unit 303 obtains the initial route of the flow targeting the initial topology.

Then, step S123 to step S1.
At 24, the edge cost calculation unit 305 calculates the edge cost and the net cost from the flow flowing through each edge.

Next, in step S125, it is determined whether or not the current network cost is lower than the previous network cost. This time, the initial route of the flow and the initial cost of the network are obtained. Therefore, the processing unconditionally advances to step S126.

In step S126, the route changing unit 304
Changes the route of the flow, and proceeds to step S123 and step S124 again, and the edge cost calculation unit 305 obtains the edge cost and the network cost. Then, step S12
In step 5, if the network cost is lower than the previous network cost, the process proceeds to step S126, and if the cost is increased, the process proceeds to step S127.

In step S127, the flow path is returned to the state before the change, and the flow advances to step S128 to make a condition determination. That is, when the number of times h the topology has been changed is smaller than the specified number K, the process proceeds to step S129, 1 is added to the number of times the topology has been changed, and the process proceeds to step S130. At this time, in step S130, since the network information has not been stored in the best result storage unit 306 in the past, the process unconditionally proceeds to step S131, the communication network design information of this time is stored in the best result storage unit 306, and In step S132, the topology changing unit 302 changes the topology, and the process returns to step S122. After that, the same thing is repeated.

If it is determined in step S128 that the number of times the topology has been changed, h, is equal to or greater than the specified number, K, then step S128.
Proceeding to 133, the communication network design information stored in the best result storage unit 306 is output, and the processing ends.

The communication network design apparatus 300 according to the seventh embodiment is different from the sixth embodiment in that the topology change is repeated and the network cost is higher than the network cost when the previous topology is used. Even in such a case, the topology is continuously changed, the topology is changed a certain number of times (K), and then the communication network design information stored in the best result storage unit 306 is output.

FIG. 32 shows a configuration example of the topology changing unit 302, which is composed of a topology storage unit 310, an edge deleting unit 311, and a topology selecting unit 312.

The initial topology information obtained by the initial topology setting unit 301 is output, and the topology storage unit 3
It is supposed to be stored in 10.

The edge deletion unit 311 uses the traffic information such as the origination node / destination node of the communication flow, the band, the preference of the route, and the route allocation result of the past flow to store the topology stored in the topology storage unit 310. A plurality of new topologies are obtained by removing k (> 0) edges from the edges. The topology information obtained as a result is the initial route setting unit 303.
Is output to

The topology selecting unit 312 uses the result of the route determined by the route changing unit 304 for the plurality of topologies obtained by the edge deleting unit 311, and selects the topology from which the best result can be obtained. It is designed to be selected. Then, the selected topology is stored again in the topology storage unit 312.

The edge deletion unit 311 obtains a new topology by deleting k edges from the topology newly stored in the topology storage unit 310. Thereafter, by repeating this, the communication network is designed.

Next, the operation processing of the topology changing unit 302 having the configuration shown in FIG. 32 will be described in detail with reference to the flowchart shown in FIG.

First, in step S140, it is assumed that the topology storage unit 310 stores information on the initial topology obtained by the initial topology setting unit 301. That is, the target of this initial topology is already shown in FIG.
It is assumed that the processing of step S122 to step S125 of No. 1 has been performed.

Next, in step S141, the edge deletion unit 311 creates a new topology by removing b edges of the topology stored in the topology storage unit 310, and outputs this. Using the output topology, steps S122 to S12 in FIG.
5 is performed, and as a result, when the communication network design information is received from the route changing unit 304, the edge deleting unit 311 again.
Generates a new topology by removing b edges from the topology stored in the topology storage unit 310, and outputs this (step S142).
At this time, it goes without saying that the edge deleting unit 311 generates a topology different from the topology generated in the past. In this way, the edge deletion unit 311 creates new topologies. That is, assuming that the topology storage unit 310 stores the initial topology as shown in FIG. 28, the edge deletion unit 311 generates the initial topology by removing, for example, b = 2 edges. A specific example of the new topology is shown in FIG.

FIGS. 34 (a) to 34 (o) show the case where two of the six edges drawn between the nodes A, B, C and D are deleted as the initial topology. Shows all possible new topologies.

Here, the edge deleting unit 311 has generated all the topologies obtained by deleting the b edges, but for example, two or more routes between arbitrary nodes are required, and this limitation is imposed. You may prohibit the deletion pattern of the edge that destroys. Also, to reduce the amount of calculation,
It is also possible to prohibit the generation of a topology that is unlikely to reduce the network cost, by using traffic information such as favorable or bad for the flow route and communication network design information.

If the edge deleting unit 311 has generated all the allowed topologies (step S143), the process proceeds to step S144, where the topology selecting unit 312 deletes the edges using the topology currently stored in the topology storing unit 310. The topology that gives the minimum network cost is selected from the topologies generated by the unit 311.

If the network cost using the topology selected in step S144 is lower than the network cost using the topology stored in topology storage unit 310 (step S145), step The topology selected in S144 is stored in the topology storage unit 3
10 (step S146), step S141
Proceed to and continue to generate new topology.

On the other hand, the network cost when the topology selected in step S145 is used is the topology storage unit 3
If the network cost is not lower than that when the topology stored in 10 is used (step S14)
5) The generation of the new topology is completed. Alternatively,
As in step S128 of FIG. 31, the topology may be generated a predetermined number of times even if the network cost does not decrease.

At this time, b may be a fixed value or may be determined by a non-decreasing function of the number of edges of the topology stored in the topology storage unit 310.

The topology changing unit 3 shown in FIG.
The configuration example of 02 is also effective as a configuration example of the topology changing unit 202 in the sixth embodiment described above. In this case, in step S107 of FIG. 26, the route changing unit 2
When determining whether or not 04 changes the topology, the network cost when the topology stored in the topology storage unit 310 is used and the edge deletion unit 311
It goes without saying to compare the network costs when using the topology that gives the best result among the plurality of topologies obtained by removing b (> 0) edges from that topology.

Next, the eighth embodiment will be described.
Here, a communication network designing method and a communication network designing device when it is permitted to set two or more links between nodes will be described.

As a method of calculating the cost of an edge when it is permitted to set two or more links between nodes, for example, there are the following methods. That is, if the total amount of flows passing through the edge e is written as y, a link combination in which the sum of the link capacities is y or more and the total cost of the links is the minimum is selected, and the cost of the selected links is calculated. The edge cost calculation unit 305 obtains the sum as the cost of the edge e, so that the communication network design apparatus 300
Can be used as a communication network design device when it is permitted to set two or more links between nodes.

Next, a method of selecting links to be set when it is permitted to set two or more links between nodes will be described. In the following description, the total amount of flows passing through the edge e is y, and there are n types of links that can be set at that edge, and the capacity of the i-th link is a _i and the cost is c _i. Then, the number of i-th links to be set on the edge is represented by x _i . Then, it is understood that the following problem (P (y)) should be solved to select the link with the minimum cost.

The problem (P (y)) is to find a solution that minimizes z expressed by the equation (11) under the constraint condition of the equation (12).

[0295]

(Equation 5)

[0296]

(Equation 6) Note that y, c _j , and a _j are non-negative integers.
Even if such an assumption is made, when y, c _j , and a _j are real numbers with finite digits, they can be converted into integers by multiplying them by an appropriate integer, so there is no practical problem.

The optimum value of the problem (P (y)) is written as G (y). That is,

[0298]

(Equation 7) Is defined.

Next, the optimal solution X = (x of the problem (P (y))
₁ , x ₂ , ..., X _n )

[0300]

(Equation 8) Is defined. Then, if x _j > 0, X ′ = (x ₁ ,
..., x _j-1 , x _j , x _{j + 1} , ..., x _n ) is the problem (P
(Y−a _j )) is the optimal solution. Expression (15) holds for all j for which x _j > 0.

[0301]

[Equation 9] Therefore,

[0302]

(Equation 10) In other words, when T (y) = φ,

[0303]

[Equation 11] Becomes Since G (0) = 0 by definition, G (y) can be recursively obtained using the equation (17). Also,
When T (y) ≠ φ,

[0304]

(Equation 12) Can be obtained by

Furthermore, G (y) has the following properties. That is, when a _max = max _j {a _j },

[0306]

(Equation 13) If y > y ^* , the following equation holds for y.

[0307]

[Equation 14] The problem (P (y)) can be more easily solved by using the periodicity of G (y).

By summarizing the above, an algorithm (procedure) for solving the problem P (y) will be described. (Procedure 1) If y = 0, G (y) = 0 and x = 0 are set, and the process ends. (Procedure 2) a _max = max _j a _j , y ^* = 1, G (i)
= ∞ (i = 1, 2, ..., Y), d (i) = n (i = 1,
2, ..., Y), x _j = 0 (j = 1, 2, ..., N) (procedure 3) i = 1 (procedure 4) If i = y ^* + a _max, proceed to procedure 9, if i> y If so, go to procedure 11 (procedure 5) Obtain S (i) and T (i), and obtain G (i) using equation (18). At this time, j that satisfies the expression (18) is represented by j ^* . (Procedure 6) d (i) = j ^* (Procedure 7) If j ^* ≠, y ^* = i + 1 (Procedure 8) Go to procedure 4 with i = i + 1 (Procedure 9) k = “(y−y ^* )” / a ₁ ", y '= y-
ka ₁ (procedure 10) G (y) = kc ₁ + G (y ′), i =
y ′ (procedure 11) If i = 0, end (procedure 12) j = d _i , x _j = x _j +1 (procedure 13) i = i−a _j (procedure 14) Go to procedure 11, where “x” is , X and the maximum integer less than or equal to x.

By the above method, the link to be stretched over the edge can be determined from the total amount of flows passing through the edge. For example, in a network including nodes A, B, C and D, if two types of links of capacity and cost as shown in FIG. 35 can be set between these nodes, a plurality of links should be provided between each node. If it can be set, the relationship between the flow amount flowing at each edge and the edge cost is as shown in FIG.

FIG. 35 shows a table group showing the capacity and the cost that can be set for each link of the network composed of the nodes A, B, C and D.
Shows a table group showing, for each edge of the network, a specific example of the relationship between the amount of flow flowing to the edge and the cost.

For example, suppose that the links that can be set between the nodes A and B shown in FIG. 35 are links with a capacity of 20 Mbps and a cost of 40,000 yen and links with a capacity of 100 Mbps and a cost of 160,000 yen. From the relationship between the flow amount and the edge cost, the flow amount is 0
At ~ 20 Mbps, one link with a capacity of 20 Mbps is put on, and it can be seen that the cost is 40,000 yen.
Further, when the flow amount is 20 to 40 Mbps, two links having a capacity of 20 Mbps are put, and thus the cost is 80,000 yen. Also, the flow amount is 40-60
It can be seen that at Mbps, three links with a capacity of 20 Mbps are installed, and therefore the cost is 120,000 yen. Further, when the flow amount is 60 to 100 Mbps, one link having a capacity of 100 Mbps is provided, and thus the cost is 1.
It turns out that it is 60,000 yen.

Note that the edge cost calculation unit 305 uses, for example, FIG. 3 instead of performing steps 1 to 14 described above.
A table group indicating the relationship between the amount of flowing flow at each edge and the edge cost such as 6 may be stored in advance.

Next, the ninth embodiment will be described.

FIG. 37 shows the configuration of the communication network designing apparatus 400 according to the ninth embodiment. The flow calculating section 40 is shown in FIG.
1, virtual network design unit 402, connection route determination unit 40
3 and a link setting unit 404.

The flow calculation unit 401 receives, as connection information, a demand for a connection to be set, obtains the total bandwidth of the connections having the same originator node and destination node, and virtually generates a flow.

The virtual network design unit 402 has a flow calculation unit 40.
This is to design a network for setting the flow determined by 1.

The connection route determination unit 403 determines the connection route from the flow routes designed by the virtual network design unit 402. The link setting unit 404 determines the link to be set between the nodes based on the determined connection route.

For example, given three nodes A, B and C, the bandwidth from node A to node B is 70 Mbps.
Connection from node C to node A, one connection with bandwidths of 50 Mbps and 70 Mbps from node C, and bandwidth from node C to node B with 80 Mbps and 7
Consider a case where there is a demand for setting one connection for 0 Mbps and one connection for 10 Mbps. The connection route should not be branched into two or more.

The flow calculation unit 401 generates a flow based on such connection information. As a result, the bandwidth required to be set from node A to node B is 14
The flow of 0 Mbps, the bandwidth requesting to be set from the node C to the node A is 120 Mbps, and the bandwidth requesting to be set from the node C to the node B is 1.
A 60 Mbps flow is generated.

At this time, it is assumed that the types of links that can be set between the nodes are given as shown in FIG. That is, between node A and node B, node B and node C
Between nodes C and A,
It is assumed that a bps link can be set. Also, the cost is assumed to be the same for each node.

At this time, information regarding the capacity and cost of this link (network information) and the flow calculation unit 401.
Using the flow information generated by the virtual network designing unit 40.
2 determines the link to be set up between the flow route and the node. FIG. 39 shows the flow path obtained as a result.

Using the route of this flow, the connection route determining unit 403 determines the route of the connection.
In FIG. 39, a part of the flow for requesting the setting from the node C to the node B passes through the node A. When the flow route is branched in this way, the connection set for that route is determined according to the flow bandwidth of each route. In FIG. 39, the route of the flow set from the node C to the node B is branched into two routes of 150 Mbps and 10 Mbps. The bandwidth of the connection that is the basis of this flow is 80 Mbps, 70 Mbps, 10 M
Since it is bps, the connection path is determined so as to be as close as possible to the flow bandwidth by using this connection bandwidth. As a result, the connection of 10 Mbps is set to the route C → A → B as shown in FIG.

As a result, the connection route is as shown in FIG.
It is determined as 0. That is, two connection paths with a bandwidth of 70 Mbps from node A to node B,
Bandwidth from node C to node A is 50 Mbps and 70 Mbps
Two connection paths of ps, two connection paths of bandwidths 80 Mbps and 70 Mbps from node C to node B, and a connection path of 10 Mbps from node C to node B via node A are determined. .

Then, since the connection passing through the nodes is already determined, the link accommodating the connection passing between the nodes can be determined.

In this example, the final network accommodating the connection and the virtual network set the same link, but if necessary, the designed link is different between the virtual network and the finally designed network. May be.

In the description of the configuration of the communication network designing apparatus 400 shown in FIG. 37, the description has been made on the assumption that the connection group to be set is given in advance as the connection information. However, if the connection setting request occurs stochastically. Alternatively, a connection group to be set may be obtained such that the connection setting rejection probability is equal to or less than a predetermined threshold.

Also, the communication network designing device 40 shown in FIG.
The virtual network designing unit 402 of 0 is also effective when using the communication network designing devices 200 and 300 as shown in FIGS. In this case, the communication flow information obtained by the flow calculation unit 401 of the communication network designing device 400 of FIG. 37 may be used as the traffic information.

As described above, according to the sixth to ninth embodiments, it is possible to obtain a highly accurate approximate solution to the communication network design problem by repeating the topology change and the flow route change. If possible, and if the connection to be set is given, convert the connection to a flow, solve the communication network design problem for the virtual flow, and determine the route of the connection based on that solution. By doing so, the network can be designed with a small amount of calculation.

Further, according to the communication network designing method and the communication network designing apparatus using the method according to the sixth embodiment, the initial route setting unit 203 makes the initial topology given by the initial topology setting unit 201 flow. Settings,
Next, the edge cost calculation unit 205 obtains the cost of the edge from the amount of flow flowing on each edge, and the route change unit 204 obtains a new route of the flow based on the result, and a part or all of the flow is The cost of the edge that changes with the amount of flow that flows on each edge that changes by moving to a new path is recalculated, and the same operation is repeated until the network cost does not decrease. The initial topology is changed, the route of the flow is determined using this changed topology, and the change of the topology and the determination of the route of the flow are repeated for the new topology, and the network cost is reduced by changing the topology. By repeating until
It is possible to obtain an accurate approximate solution of the network design problem.

Further, according to the communication network designing method and the communication network designing apparatus using the method according to the seventh embodiment, the initial route setting unit 303 executes the flow in the initial topology given by the initial topology setting unit 301. Settings,
Next, the edge cost calculation unit 305 obtains the cost of the edge from the amount of flow flowing on each edge, and the route change unit 304 obtains a new route of the flow based on the result, and a part or all of the flow is The cost of an edge that changes with the amount of flow that flows on each edge that changes by moving to a new path is recalculated, and the same operation is repeated until the flow path is no longer changed. Change the initial topology, use this changed topology to determine the flow path, and then repeat the change of the topology and the flow path for the new topology. By storing the result in the best result storage unit 306, it is possible to obtain a highly accurate approximate solution of the network design problem.

A plurality of topologies are obtained by changing the combination of edges to be removed when the edge deleting means removes a given number of edges from the topologies stored in the topology storage unit 310, and the plurality of topologies are obtained. , The topology selection unit 312 selects the topology that gives the best result, stores it in the topology storage unit 310 again, and repeats this operation to obtain an accurate approximate solution of the network design problem. Can be obtained.

Further, according to the communication network designing method and the communication network designing apparatus using the method according to the eighth embodiment, the edge cost calculation unit 305 determines the edge cost of a link that can accommodate a flow. By selecting the combination of the links with the lowest cost among the combinations, it is possible to obtain an accurate approximate solution of the network design problem.

Further, according to the ninth embodiment, the network design problem in the case where a plurality of connections to be set is given and the path of one connection is not allowed to branch into two or more is Since it is a combination problem, it is very difficult to solve it. However, the flow calculation unit 401 collects the connections having the same source node and destination node, obtains the flow, and the flow route is branched into two or more. By adopting this, the virtual network design unit 402 can easily solve the network design problem. A connection route is determined using the connection route determination unit 403 based on the flow route thus obtained. After that, the link setting unit 404 obtains a link to be set between the nodes based on the connection route. By this operation, it becomes possible to solve the network design problem of setting the connection with a small amount of calculation.

[0334]

As described above, according to the present invention,
A route selection method and a route selection device capable of efficiently selecting an optimal route under the condition that an upstream connection and a downstream connection belonging to the same bidirectional connection pass through the same route.

Further, according to the present invention, there is provided a route selection method and a route selection device capable of providing a highly accurate approximate solution to the route selection problem of selecting a plurality of routes on a network given in advance. Can be provided.

Further, according to the present invention, it is possible to obtain an accurate approximate solution of the communication network design problem of designing the communication network from the capacity and cost of the link that can be set between the nodes and the traffic demand with a small amount of calculation. A communication network designing method and a communication network designing device can be provided.

[Brief description of drawings]

FIG. 1 is a block diagram schematically showing the configuration of a route selection device according to a first embodiment of the present invention.

FIG. 2 is a block diagram schematically showing the configuration of an evaluation value calculation unit in FIG.

FIG. 3 is a block diagram schematically showing the configuration of a route selection unit in FIG.

FIG. 4 is a flowchart illustrating a processing operation of an evaluation value calculation unit.

FIG. 5 is a flowchart for explaining a processing operation of a route selection unit.

FIG. 6 is a diagram showing a configuration example of a network for which a route is selected.

FIG. 7 is a diagram showing a specific example of the contents of connection information necessary for selecting a route.

FIG. 8 is a block diagram schematically showing a configuration of a route selection device according to a second exemplary embodiment of the present invention.

9 is a block diagram schematically showing the configuration of an evaluation value calculation unit in FIG.

10 is a flowchart for explaining the processing operation of the route selection device in FIG.

FIG. 11 is a diagram showing a specific example of the contents of connection information used for specifically explaining the processing operation of the route selection device of FIG.

FIG. 12 is a block diagram schematically showing the configuration of a route selection device according to a third exemplary embodiment of the present invention.

FIG. 13 is a diagram showing a configuration example of a network for which a route is selected.

FIG. 14 is a block diagram schematically showing the configuration of a route determination unit shown in FIG.

15 is a flowchart for explaining the processing operation of the route selection device in FIG.

FIG. 16 is a diagram showing another example of the configuration of a route selection target network.

FIG. 17 is a diagram showing another specific example of the content of the connection information necessary for selecting a route.

FIG. 18 is a diagram showing a specific example of the reservation status of the bandwidth of each connection.

FIG. 19 is a block diagram schematically showing a configuration of a route determination unit according to a fourth embodiment of the present invention.

FIG. 20 is a flowchart for explaining an operation process of the route selection device according to the fourth embodiment.

FIG. 21 is a diagram showing an example of correspondence between routes searched by a route search unit and route numbers.

FIG. 22 is a block diagram schematically showing the configuration of a route determination unit according to a fifth embodiment of the present invention.

FIG. 23 is a flowchart for explaining operation processing of the route selection device according to the fifth embodiment.

FIG. 24 is a diagram for explaining a route selection method when a plurality of links are provided between nodes.

FIG. 25 is a block diagram schematically showing the configuration of a communication network designing device according to a sixth embodiment of the present invention.

FIG. 26 is a flowchart for explaining an operation process of the communication network design device according to the sixth embodiment.

FIG. 27 is a diagram showing a specific example of the contents of network information necessary for designing a communication network, which includes information on types of links that can be set between nodes, capacity of the links, and cost.

FIG. 28 is a diagram showing a specific example of the initial topology obtained by the initial topology setting unit.

FIG. 29 is a diagram for explaining a process of determining a flow path.

FIG. 30 is a block diagram schematically showing the configuration of a communication network designing device according to a seventh embodiment of the present invention.

FIG. 31 is a flowchart for explaining an operation process of the communication network designing device according to the seventh embodiment.

FIG. 32 is a block diagram schematically showing the configuration of a topology changing unit.

FIG. 33 is a flowchart for explaining operation processing of the topology changing unit in FIG. 32.

FIG. 34 is a diagram showing a specific example of a new topology created by deleting a predetermined number of edges by the edge deleting unit.

FIG. 35 is a diagram showing another specific example of contents of network information necessary for designing a communication network, which includes information on types of links that can be set between nodes, capacity of the links, and cost. .

FIG. 36 is a diagram showing a specific example of the correspondence between the amount of flows passing through an edge and the edge cost when two or more links can be set between nodes.

FIG. 37 is a block diagram schematically showing the configuration of a communication network designing device according to a ninth embodiment of the present invention.

38 is a diagram for explaining the processing operation of the communication network designing device in FIG. 37, showing an example of links that can be set between the nodes.

FIG. 39 is a diagram for explaining the processing operation of the communication network designing device of FIG. 37, and is a diagram showing a specific example of a flow route obtained as a result of designing a virtual network by the virtual network designing unit.

FIG. 40 is a diagram for explaining the processing operation of the communication network designing device of FIG. 37, and is a diagram showing a specific example of the route of each connection determined by the connection route determining unit.

[Explanation of symbols]

1 ... Route selection device, 2 ... Evaluation value calculation unit, 3 ... Route selection unit, 5 ... Route selection device, 6 ... Parameter change unit, 7 ... Evaluation value calculation unit, 8 ... Route selection unit, 9 ... Best result storage unit 1
00 ... Route selecting device, 101 ... Route searching unit, 102 ... Band reservation unit, 103 ... Route determining unit, 200 ... Communication network designing device, 201 ... Initial topology setting unit, 202 ... Topology changing unit, 203 ... Initial route setting unit , 204 ... Route changing unit, 205 ... Edge cost calculating unit, 300 ... Communication network designing device, 301 ... Initial topology setting unit, 302 ... Topology changing unit, 303 ... Initial route setting unit, 304 ... Route changing unit, 305 ... Edge Cost calculation unit, 306 ... Best result storage unit.

Claims (20)

[Claims] 1. A route selection method for selecting a plurality of bidirectional connection routes in a communication network formed by connecting a plurality of node devices so that the upstream connection and the downstream connection pass through the same route. Based on the bandwidth of the upstream connection and the bandwidth of the downstream connection of the bidirectional connection, calculate the evaluation value of the bidirectional connection, according to the order based on the evaluation value calculated for each of the plurality of bidirectional connections, A route selection method characterized by selecting a plurality of bidirectional connection routes. 2. When the bandwidth of the uplink connection of the bidirectional connection is larger than or equal to the bandwidth of the downlink connection, the value of the bandwidth of the uplink connection is multiplied by a value of a first weight, and the downlink An evaluation value of the bidirectional connection is calculated by adding the value of the bandwidth of the connection multiplied by the value of the second weight, and when the bandwidth of the upstream connection is smaller than the bandwidth of the downstream connection, The bidirectional connection is obtained by adding a value obtained by multiplying the value of the band of the uplink connection by the value of the second weight and a value obtained by multiplying the value of the band of the downlink connection by the value of the first weight. The route selection method according to claim 1, wherein the evaluation value of is calculated. 3. A route selection device for selecting a plurality of bidirectional connection routes in a communication network formed by connecting a plurality of node devices so that the upstream connection and the downstream connection pass through the same route. An evaluation value calculation means for calculating an evaluation value of the bidirectional connection based on the bandwidth of the upstream connection and the bandwidth of the downstream connection of the bidirectional connection, and the evaluation value calculation means for each of the plurality of bidirectional connections. And a route selecting unit that selects routes of the plurality of bidirectional connections according to the ranking based on the evaluated value. 4. The evaluation value calculating means assigns a first weight value to the value of the bandwidth of the upstream connection when the bandwidth of the upstream connection of the bidirectional connection is greater than or equal to the bandwidth of the downstream connection. The evaluation value of the bidirectional connection is calculated by adding the product of the multiplication and the product of the value of the bandwidth of the downlink connection multiplied by the value of the second weight, and the bandwidth of the uplink connection is the bandwidth of the downlink connection. When smaller than the band,
The bidirectional connection is obtained by adding a value obtained by multiplying the value of the band of the uplink connection by the value of the second weight and a value obtained by multiplying the value of the band of the downlink connection by the value of the first weight. The route selection device according to claim 3, wherein the evaluation value of is calculated. 5. A route selection method for selecting a plurality of bidirectional connection routes in a communication network formed by connecting a plurality of node devices so that the upstream connection and the downstream connection pass through the same route. Based on each of the different parameters output according to, the bandwidth of the upstream connection and the bandwidth of the downstream connection of the bidirectional connection, calculate the evaluation value of the bidirectional connection, based on the calculated evaluation value According to the order, the routes of the plurality of bidirectional connections are respectively selected, the selection result is output as a route selection result corresponding to the parameter, and the plurality of bidirectional connections are output based on the output route selection results. A route selection method characterized by determining the optimum route of 6. A route selection device for selecting a plurality of bidirectional connection routes in a communication network formed by connecting a plurality of node devices so that the upstream connection and the downstream connection pass through the same route, which is required. Parameter output means for outputting different parameters according to, the parameter output by the parameter output means, and the evaluation value of the bidirectional connection based on the bandwidth of the upstream connection and the bandwidth of the downstream connection of the bidirectional connection The evaluation value calculation means for calculating the route and the order based on the evaluation value calculated by the evaluation value calculation means, each of the plurality of bidirectional connection routes is selected, and the selection result is the route selection result corresponding to the parameter. Based on the output means for outputting as a plurality of route selection results output by the output means Routing apparatus being characterized in that comprising determination means for determining an optimal path in the number of two-way connections, the. 7. A route selection method for selecting a route of a plurality of connections in a communication network formed by connecting a plurality of node devices, wherein a route that can be taken by the plurality of connections is searched and the searched route is searched. The bandwidth of the connection is reserved for the upper link, and one of the searched routes is selected for each of the plurality of connections based on the reserved state of the bandwidth. Route selection method. 8. The overflow band that cannot be accommodated in the link is calculated from the bands reserved for the link, and an evaluation value is calculated for a route that the connection can take based on the calculated overflow band. Then, according to the order based on the calculated evaluation value, selecting one of the searched routes for at least one connection of the plurality of connections means that the route for all connections is The route selection method according to claim 7, wherein the process is repeated until the route is selected. 9. When one of the searched routes is assigned to the connection and the reserved band of the connection is removed from a link not on the route, the link is reserved in the reserved band. An overflow band that cannot be accommodated in the storage device, calculates an evaluation value for a route that can be taken by the plurality of connections based on the calculated overflow band, and calculates the evaluation value according to the order based on the calculated evaluation value. 8. The method of selecting one of the searched routes for at least one of the connections of the above is repeated until the route is selected for all of the connections. Route selection method. 10. A first overflow band that cannot be accommodated in the link among the bands reserved for the link is calculated, and it is assumed that one of the searched routes is assigned to the connection, and the route is calculated. When the reserved band of the connection is removed from a link that is not on the upper side, a second overflow band that cannot be accommodated in the link in the reserved band is calculated, and the first overflow band and the second overflow band are calculated. First, an evaluation value is calculated for a route that the plurality of connections can take, and the searched for at least one connection among the plurality of connections according to the order based on the calculated evaluation value. 8. The route selection method according to claim 7, wherein selecting one of the routes is repeated until routes are selected for all the connections. 11. A route selection device for selecting a route of a plurality of connections in a communication network formed by connecting a plurality of node devices, and route search means for searching a route that the plurality of connections can take. Bandwidth reservation means for reserving the bandwidth of the connection to the link on the route searched by the route search means, and each of the plurality of connections based on the reservation state of the bandwidth reserved for the link by the bandwidth reservation means On the other hand, a route selecting device comprising: a route selecting unit that selects one of the routes searched by the route searching unit. 12. The route selecting means calculates an overflow band that cannot be accommodated in the link among the bands reserved for the link by the bandwidth reservation means, and the overflow band calculating means calculates the overflow band. Based on the overflow bandwidth
Of the plurality of connections, according to an evaluation value calculation unit that calculates an evaluation value for a route that can be taken by the connection searched by the route search unit, and an order based on the evaluation value calculated by the evaluation value calculation unit. For at least one connection, selecting means for selecting one of the routes searched by the route searching means, and updating the reserved state of the bandwidth until routes are selected for all connections. The route selection device according to claim 11, further comprising: a control unit that controls the overflow band calculation unit to calculate the overflow band based on the updated reserved state of the band. 13. The route selecting unit assumes that one of the routes searched by the route searching unit is assigned to the connection, and removes the reserved bandwidth of the connection from a link that is not on the route. And an overflow bandwidth calculation means for calculating an overflow bandwidth that cannot be accommodated in the link among the reserved bandwidths, and based on the overflow bandwidth calculated by this overflow bandwidth calculation means, the connection searched for by the route search means. The evaluation value calculation means for calculating an evaluation value for a possible route, and the route search for at least one of the plurality of connections according to the order based on the evaluation value calculated by the evaluation value calculation means. Selecting means for selecting one of the routes searched for by the means, and the bandwidth reservation until the routes are selected for all connections. The control means for updating the status and controlling the overflow bandwidth calculation means to calculate the overflow bandwidth based on the updated bandwidth reservation status. Route selector. 14. The route selecting means calculates a first overflow band which cannot be accommodated in the link among the bands reserved for the link by the band reservation means, and the connection in the connection. If it is assumed that one of the routes searched by the route search means is assigned and the reserved band of the connection is removed from the link not on the route, the second band cannot be accommodated in the reserved band.
Based on the second overflow band calculated by the second calculator, the first overflow band calculated by the first calculator, and the second overflow band calculated by the second calculator. Of the plurality of connections, according to an evaluation value calculation unit that calculates an evaluation value for a route that can be taken by the connection searched by the route search unit, and an order based on the evaluation value calculated by the evaluation value calculation unit. For at least one connection, selecting means for selecting one of the routes searched by the route searching means, and updating the reserved state of the bandwidth until routes are selected for all connections. A control for controlling the first calculating means and the second calculating means to calculate the first overflow band and the second overflow band, respectively, based on the updated reserved state of the band. Means and Routing device according to claim 11, characterized in that it comprises a. 15. Network information including at least the number of nodes forming a communication network, identification information of each node, type and cost of a link that can be set between the nodes, and
A communication network design method for designing a communication network capable of accommodating a traffic demand required by the traffic information, based on traffic information including at least identification information of a source node and a destination node of the communication flow and the amount of the communication flow. Then, based on the network information, the initial topology of the communication network is set, the initial topology is changed as necessary, and the set initial topology or the topology obtained by the change Then, based on the traffic information, the route of the desired flow is set, the cost of the link on the set route is calculated, and until the calculated cost satisfies a predetermined condition, While changing the set route, the cost of the link on the changed route is calculated, and the calculated cost is
When the predetermined conditions are met, at least
A communication network designing method, wherein communication network designing information including a path of the desired flow and a type of a link set between the nodes at that time is output. 16. From the set initial topology,
Of a plurality of topologies that are created by removing at least one edge from among the plurality of edges that connect the nodes that make up that topology and changing the combination of the removed edges, The cost of the link on the route of the desired flow that has been set stores a condition that satisfies a predetermined condition, and the route of the desired flow that is set for the stored topology, between the nodes. The communication network designing method according to claim 15, wherein the communication network designing information including the set link type is output. 17. At least the number of nodes forming a communication network, the identification information of each node, network information including the type and cost of a link that can be set between each node, at least the required bandwidth for each connection, A communication network design method for designing a communication network capable of accommodating the connection based on connection information including identification information of a node and a destination node, wherein the source node and the destination node are the same based on the connection information. A communication flow is generated by putting together requested bandwidths of connections, and a virtual network is designed by setting a path of the communication flow and a link between the nodes based on the generated communication flow and the network information. , The route of the connection is determined based on the designed virtual network, and the route of the determined connection is determined. Communication network design method and sets the link corresponding to the connection between the nodes and. 18. A network information including at least the number of nodes constituting a communication network, identification information of each node, type and cost of a link that can be set between the nodes, and
A communication network designing device for designing a communication network capable of accommodating the traffic demand requested by the traffic information based on the traffic information including at least the identification information of the source node and the destination node of the communication flow and the amount of the communication flow. An initial topology setting unit that sets an initial topology of the communication network based on the network information; and a topology changing unit that changes the initial topology set by the initial topology setting unit, if necessary, Route setting means for setting a desired flow route based on the traffic information for the initial topology set by the initial topology setting means or the topology changed by the topology changing means, and this route setting Cost calculation means for calculating the cost of the link on the route set by the means, The route set by the route setting unit is changed until the cost calculated by the calculating unit satisfies a predetermined condition, and the cost of the link on the changed route is calculated by the cost calculating unit. When the cost calculated by the cost calculating means and the control means for controlling so as to satisfy a predetermined condition, at least the desired flow path at that time and the type of the link set between the nodes are set. A communication network designing device comprising: an output unit that outputs communication network designing information. 19. The topology changing means connects a storage means for storing the topology set by the initial topology setting means, and a node for configuring the topology from the topology stored in the storage means. Of the plurality of edges, at least one of the edges is removed, and among the plurality of topologies generated by changing the combination of the edges removed by the edge removal means, the route setting means is provided for the topology. 19. The communication network design according to claim 18, further comprising: a control unit that controls the storage unit to store a topology satisfying a predetermined condition in the path of the desired flow set in 1. apparatus. 20. At least the number of nodes forming a communication network, identification information of each node, network information including the type and cost of a link that can be set between each node, at least a required bandwidth for each connection, and an originating address. A communication network designing device for designing a communication network capable of accommodating the connection based on connection information including identification information of a node and a destination node, wherein the source node and the destination node are the same based on the connection information. A flow generating unit that generates a communication flow by collecting the requested bandwidths of the connections, and based on the communication flow and the network information generated by the flow generating unit, a path of the communication flow and a link between the nodes are determined. A virtual communication network design means for designing a virtual network by setting and a virtual network designed by this virtual network design means Based on the connection route determining means for determining the route of the connection, a link setting means for setting a link corresponding to the connection between nodes based on the route of the connection determined by the connection route determining means, A communication network designing device comprising: