JP5376571B2 - Control device, control method, node, and transmission method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve efficiency in the use of a network by effectively distributing traffic. <P>SOLUTION: A network 12 is configured by connecting a plurality of nodes 21 via a link 20 and transmits traffic input to the outgoing node 21 through the intermediate node 21 to the incoming node 21. On the basis of the configuration of the network and capacity of available bands of the link 20, a traffic volume ratio calculation device 11 calculates the ratio of distributing traffic with each node 21 as the intermediate node by combination of the outgoing node 21 and the incoming node 21. The present invention may be applicable to e.g., a device for controlling nodes. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to a control device, a control method, a node, and a transmission method, and more particularly, to a control device, a control method, a node, and a transmission method that can effectively distribute traffic and improve network use efficiency. .

  In a network, by adopting an appropriate route control method, it is possible to improve resource use efficiency and increase throughput (see, for example, Non-Patent Document 1). In other words, with an appropriate route control method, traffic can be optimally assigned to each route, and more traffic can be accommodated in the network. In addition, with an appropriate route control method, it is possible to suppress network congestion and increase durability against traffic fluctuation.

  One method for improving the effect of routing is to minimize the maximum value of the utilization rate of all the links in the network (hereinafter referred to as network congestion rate). By minimizing the network congestion rate, the traffic that can be accommodated in the network can be increased.

  By the way, various studies have been conducted on route control. As a result, for example, Non-Patent Document 2 formulates a general traffic engineering problem, that is, a traffic demand, as a problem in which a route can be freely selected and distributed between the arrival and departure nodes.

  Non-Patent Document 3 describes that elaborate route control can be achieved using, for example, MPLS-TE (Multi-Protocol Label Switching Traffic-Engineering) technology.

  However, in a network that has already been constructed, a shortest path-based routing protocol such as OSPF (Open Shortest Path First) or IS-IS (Intermediate System to Intermediate System) is mainly adopted. Therefore, in order for a node such as an IP (Internet Protocol) router that has already been introduced to adopt MPLS-TE, the node needs to be upgraded. Therefore, it is considered necessary to make effective use of existing routing protocols as much as possible.

  Therefore, as described in Non-Patent Documents 4 to 6, a route control method for optimizing a link weight of a network based on OSPF has been studied. In this route control method, the link weight is calculated as the link distance in calculating the shortest route. When the traffic demand changes, the optimum link weight is recalculated, and the recalculated link weight is reset to the node on the network. Then, the route is recalculated according to the latest link weight. Therefore, if the route changes frequently, instability of the route on the network is caused, which causes a packet loss and a loop.

  In addition, a distributed route control method that distributes traffic and improves the use efficiency of the network has been studied. For example, Non-Patent Document 7 proposes TPR (two-phase routing) over shortest paths. In TPR, the traffic flow sent from the source node is distributed to the intermediate nodes on the network, and goes from the intermediate node to the destination node. In TPR, the number of paths to be distributed can be increased, and thereby the network congestion rate can be reduced.

  In the TPR described in Non-Patent Document 7, there is a restriction that the route selection from the source node to the intermediate node and the route selection from the intermediate node to the destination node use an existing routing protocol. Yes. On the other hand, in the TPR described in Non-Patent Document 8, there is no restriction that an existing routing protocol is used.

  The operation of TPR will be described below. Before that, a network model that employs TPR is defined as follows.

The network is represented by an effective graph G (V, E). Here, V represents a set of nodes, and E represents a set of links. A link from the node iεV to the node jεV is represented as link (i, j) εE. The usable bandwidth capacity of link (i, j) εE is represented as c _ij, and the usage rate of link (i, j) εE is represented as L _ij . A traffic matrix representing traffic demand d _pq from node p to node q is represented as T = {d _pq }.

  The network congestion rate is represented by r. Note that the network can accept up to traffic multiplied by 1 / r with respect to the current traffic, and minimizing the network congestion rate r maximizes the acceptable traffic. Therefore, one of the objective functions of distributed path control is to minimize the network congestion rate r.

が成り立つ。この式は、ホースモデルにおけるノードｐの制約条件を表す。 Furthermore, in the TPR described in Non-Patent Documents 7 and 8, since a hose model in which T = {d _pq } is known or can be estimated is assumed as a network model, the same assumption is made here. Therefore, α _p is defined as the total amount of traffic transmitted from the node p to the network. Therefore, the formula

Holds. This expression represents the constraint condition of the node p in the hose model.

が成り立つ。この式は、ホースモデルにおけるノードｑの制約条件を表す。 Also, β _q is defined as the total amount of traffic flowing into the node q from the network. Therefore, the formula

Holds. This expression represents the constraint condition of the node q in the hose model.

As a network model, a hose model in which α _p and β _q are known or can be estimated may be assumed.

  Next, the operation of the TPR when the network model is defined as described above will be described.

が成立する。中間ノードｍに分散されたトラヒックは、着ノードｑに転送される。 In TPR, traffic is not directly transferred from node p to node q, but traffic is distributed to intermediate nodes mεV. Then, for every departure and arrival node pair (p, q), the flow of d _pq is the intermediate node M∈V, are dispersed at a ratio of k _m. Here, the ratio _{k m, 0 ≦ k m ≦} 1 and,

Is established. The traffic distributed to the intermediate node m is transferred to the destination node q.

For example, when the source node is node 1 and the destination node is node 3, as shown in FIG. 1, the traffic from node 1 as the source node to node 3 as the destination node is the node 2 as the intermediate node. , Nodes 3 and 4 are distributed at a rate of k ₂ , k ₃ and k ₄ , respectively.

Therefore, the proportion of traffic distributed from the source node 1 to the destination node 3 via the intermediate node 2 is k ₂ . The proportion of traffic distributed from the source node 1 to the intermediate node and the destination node 3 is k ₃ . The proportion of traffic distributed from the source node 1 to the destination node 3 via the intermediate node 4 is k ₄ .

Also, when the originating node is node 1 and the destination node is node 2, as shown in FIG. 1, the traffic from node 1 as the source node to node 2 as the destination node is the node as the intermediate node. 2, node 3 and node 4 are distributed at a ratio of k ₂ , k ₃ , and k ₄ , respectively.

Therefore, the node 1 is the source node, the percentage of traffic that is distributed to a node 2 which is an intermediate node and the destination node becomes k _2. The proportion of traffic distributed from the source node 1 to the destination node 2 via the intermediate node 3 is k ₃ . The proportion of traffic distributed from the source node 1 to the destination node 2 via the intermediate node ₄ is k ₄ .

That is, in TPR, traffic is distributed to the nodes 2, 3 and 4 as intermediate nodes at the ratios k ₂ , k ₃ and k ₄ regardless of the arrival / departure node pair (p, q).

  For example, the shortest route method is used as a route selection method from the node p to the intermediate node m and a route selection method from the intermediate node m to the node q. This can be realized by setting an IP-in-IP or GRE (Generic Routing Encapsulation) tunnel between the node p and the intermediate node m.

Next, a method of calculating the ratio k _m to minimize network congestion ratio r in TPR.

First, consider the traffic transferred from the node p to the node q via the intermediate node m. Here, when the traffic between the node p and the intermediate node m is represented by _bpm , the traffic between the node p and the node m is composed of two components. The first component is the traffic that is distributed with node p as the originating node and intermediate node m as the intermediate node, and the second component forwards with node p as the intermediate node and intermediate node m as the destination node Traffic.

Therefore, if the traffic as the first component is defined as b _pm ⁽¹⁾ and the traffic as the second component is defined as b _pm ⁽²⁾ , the traffic b _pm is expressed by the following equation (1). expressed.

b _pm = b _pm ⁽¹⁾ ₊ b _pm ⁽²⁾
... (1)

Further, using the constraint condition of the node p in the hose model described above, the traffic b _pm ⁽¹⁾ is given by the following formula (2), and the traffic b _pm ⁽²⁾ is given by the following formula (3).

Therefore, the traffic b _pm is expressed by the following equation (4) using the total traffic amounts α _p and β _q .

_{b pm = k m α p +} k p β m
... (4)

Here, if the link link (i, j) is on the path between the node p and the intermediate node m, a function that takes the value 1 and otherwise takes the value 0 is defined as F _pm ^ij. The usage rate L _ij of i, j) is given by the following equation (5).

However, (psi ₎ ^mij is defined by following Formula (6).

  The difference between the traffic flowing into node j and the traffic flowing out from node j is given by the following equation (7).

However, ω _m ^j is defined by the following equation (8).

Thus, in TPR, the optimal path problem of determining the ratio k _m, is formulated by the following formula (9a) to (9f).

Equation (9a) shows an objective function that minimizes the network congestion rate r. Equation (9b), with respect to nodes of the network, indicating that all of the sum of the proportion k _m is 1. Expression (9c) indicates that the total sum of traffic passing through the link link (i, j) is less than or equal to c _ij × r in all links. Equation (9d) shows the flow conservation law.

By solving the above manner formulated by the formula (9a) to (9f), the ratio k _m to minimize network congestion ratio r in TPR is calculated. Since the equations (9a) to (9f) are linear programming problems (LP (Linear Programming)), they can be solved by a standard LP solver.

R. Zhang-Shen and N. McKeown, “Designing a Fault-Tolerant Network Using Valiant Load-Balancing”, IEEE Infocom 2008, April 2008 Y. Wang and Z. Wang, “Explicit routing algorithms for internet traffic engineering,” IEEE International Conference on Computer Communications and Networks (ICCCN), 1999 D. Awduche et al., “Requirements for Traffic Engineering Over MPLS”, RFC 2702, Sept. 1999 Y. Wang and M. R. Ito, “Dynamics of load sensitive adaptive routing”, IEEE International Conference on Communications (ICC), 2005 B. Fortz and M. Thorup, “Optimizing OSPF / IS-IS weights in a changing world”, IEEE Journal on Selected Areas in Communications, vol. 20, no. 4, pp. 756-767, 2002 J. Chu and C. Lea, “Optimal Link Weights for Maximizing QoS Traffic”, IEEE ICC 2007, pp. 610-615, 2007 M. Antic and A. Smiljanic, “Oblivious Routing Scheme Using Load Balancing Over Shortest Paths”, IEEE ICC 2008, 2008 M. Kodialam, T. V. Lakshman, J. B. Orlin, and S. Sengupta, “Pre-Conguring IP-over-Optical Networks to Handle Router Failures and Unpredictable Traffic”, IEEE Infocom 2006, April 2006

As described above, in the TPR, the ratio k to distribute traffic to intermediate node m is the departure node pair independent, a k _m that depends only on the intermediate node m. Therefore, there is a possibility that the traffic cannot be dispersed effectively.

  The present invention has been made in view of such a situation, and is intended to effectively distribute traffic and improve the use efficiency of a network.

The first aspect of the present invention is configured by connecting a plurality of nodes via a link, and by TPR (Two-Phase Routing), among the plurality of nodes between the originating node and the destination node. in certain networks via intermediate nodes Ru done communication traffic, the inputted traffic to the calling node, the calling node recognizes the information for specifying the destination node from the input packet, wherein The intermediate node is determined according to a ratio of distributing traffic according to a destination node, the intermediate node information is added to the packet, and the packet is forwarded toward the intermediate node. when transmitting the packet to the destination node via, in the control unit for controlling the communication of the traffic, the net Construction of over click, and, based on the amount of available bandwidth of the link, for each combination of the destination node and the calling node, calculating means for calculating a ratio of dispersing the traffic of each node as the intermediate node And a supply unit that supplies, to each of the plurality of nodes, a ratio of distributing the traffic for each destination node when that node becomes a source node .

The first aspect of the present invention is configured by connecting a plurality of nodes via a link, and by TPR (Two-Phase Routing), among the plurality of nodes between the originating node and the destination node. in certain networks via intermediate nodes Ru done communication traffic, the inputted traffic to the calling node, the calling node recognizes the information for specifying the destination node from the input packet, wherein The intermediate node is determined according to a ratio of distributing traffic according to a destination node, the intermediate node information is added to the packet, and the packet is forwarded toward the intermediate node. when transmitting the packet to the destination node through the control method of the control unit for controlling the communication of the traffic, Configuration of the serial network and calculating step based on the amount of available bandwidth of the link, for each combination of the destination node and the calling node, which calculates a ratio of dispersing the traffic of each node as the intermediate node And a supply step of supplying, to each of the plurality of nodes, a ratio of distributing the traffic for each destination node when that node becomes a source node .

According to a second aspect of the present invention, there is provided a node constituting a network configured by connecting a plurality of nodes via a link, and the traffic inputted by the TPR (Two-Phase Routing) A destination node recognition means for recognizing information identifying the destination node from an input packet in a node that transmits to a destination node in the network via a predetermined intermediate node among a plurality of nodes in the network ; The traffic is transmitted based on a ratio of distributing the traffic using each node in the network as the intermediate node obtained for each combination of the node as the source node to which the traffic is input and the destination node. determining means for determining an intermediate node, the information of the intermediate node determined by the determining means It was added to the serial packet, a node and a transmission means for transmitting the packet information of the intermediate node is added to the intermediate node.

According to a second aspect of the present invention, there is provided a node constituting a network configured by connecting a plurality of nodes via a link, and the traffic inputted by the TPR (Two-Phase Routing) In a node transmission method for transmitting to a destination node in the network via a predetermined intermediate node among a plurality of nodes in the network, the destination node recognition for recognizing information identifying the destination node from an input packet And based on a ratio of the traffic distributed using each node in the network as the intermediate node, obtained for each combination of the node as the source node to which the traffic is input and the destination node. A determination step of determining an intermediate node to transmit the message, and a determination by the processing of the determination step The information of the intermediate nodes added to the packet, a transmission method and a transmission step of transmitting the packet in which information of the intermediate node is added to the intermediate node.

In the first aspect of the present invention, a plurality of nodes are connected to each other via a link, and the TPR (Two-Phase Routing) includes a plurality of nodes between a source node and a destination node. in a given network via an intermediate node communication traffic Ru performed for the inputted traffic to the calling node, the calling node recognizes the information for specifying the destination node from the input packet, Determining the intermediate node according to a ratio of distributing traffic according to the destination node, adding the information of the intermediate node to the packet, and forwarding the packet toward the intermediate node; when transmitting the packet to the destination node via the configuration of the network, and, usable of the link Based on the volume of such zone, for each combination of the destination node and the calling node, the percentage calculated dispersing the traffic of each node as the intermediate node, for each of said plurality of nodes, whose node A ratio of distributing the traffic for each destination node when becoming a source node is supplied.

In the second aspect of the present invention , input traffic is transmitted by TPR (Two-Phase Routing) among a plurality of nodes in a network configured by connecting a plurality of nodes via links . When transmitting to a destination node in the network via a predetermined intermediate node, information identifying the destination node is recognized from an input packet, and the node as the source node to which the traffic is input An intermediate node that transmits the traffic is determined based on a ratio of distributing the traffic with each node in the network as the intermediate node, obtained for each combination of destination nodes , and information on the intermediate node is the packet. And the packet to which the information of the intermediate node is added is transmitted to the intermediate node .

  As described above, according to the present invention, it is possible to effectively distribute traffic and improve the use efficiency of the network.

<First embodiment>
[Configuration example of one embodiment of network system]
FIG. 2 is a diagram showing a configuration example of an embodiment of a network system to which the present invention is applied.

  The network system 10 in FIG. 2 includes a traffic volume ratio calculation device 11 and a network 12 including nodes 21-1 to 21-4 connected via links 20-1 to 20-5. The network system 10 transmits the traffic input to the source node to the destination node via the intermediate node using a method improved by TPR (herein, FTPR (Fine Two Phase Routing)).

  In the following description, the links 20-1 to 20-5 are collectively referred to as the link 20 when it is not necessary to distinguish the links 20-1 to 20-5. Similarly, the nodes 21-1 to 21-4 are also referred to as nodes 21. Further, the nodes 21 that are the source node, the intermediate node, and the destination node are also referred to as the source node 21, the intermediate node 21, and the destination node 21, respectively.

  The traffic amount ratio calculation device 11 is connected to each node 21 of the network 12. Based on the configuration of the network 12 and the usable bandwidth capacity of the link 20, the traffic amount ratio calculation device 11 calculates a ratio for distributing the traffic input to the source node 21 to the intermediate node 21 for each source node pair. To do. The traffic amount ratio calculation device 11 supplies the ratio when the node 21 is the originating node to the node 21.

  The nodes 21-1 to 21-4 have traffic distribution units 22-1 to 22-4, respectively. Hereinafter, when there is no need to particularly distinguish the traffic distribution units 22-1 to 22-4, they are collectively referred to as the traffic distribution unit 22.

  Based on the ratio supplied from the traffic volume ratio calculation device 11, the traffic distribution unit 22 of the node 21 distributes and transfers the traffic input using the node 21 as the originating node to the intermediate node 21.

  Specifically, based on the ratio supplied from the traffic volume ratio calculation device 11, the traffic distribution unit 22 transmits an intermediate node 21 that transmits traffic that is input using the node 21 including the traffic distribution unit 22 as an originating node. decide. Then, the traffic distribution unit 22 transmits the traffic to the intermediate node 21.

  Next, traffic distribution in the network system 10 of FIG. 2 will be described with reference to FIG.

  In the network system 10, since the ratio of traffic distributed to the intermediate node 21 is calculated for each pair of outgoing and incoming nodes, as shown in FIG. 3A, the outgoing node is the node 21-1, and the incoming node is the node 21-3. 3B and the case where the source node is the node 21-1 and the destination node is the node 21-2, as shown in FIG. 3B, the proportion of traffic distributed to the intermediate node 21 is different.

Specifically, for example, as shown in FIG. 3A, traffic from the source node 21-1 to the destination node 21-3 is respectively transmitted to the intermediate node 21-2, the intermediate node 21-3, and the intermediate node 21-4. , K ₂ ¹³ , k ₃ ¹³ , k ₄ ¹³ .

Therefore, the proportion of traffic distributed from the source node 21-1 to the destination node 21-3 via the intermediate node 21-2 is k ₂ ¹³ . The proportion of traffic distributed from the source node 21-1 to the intermediate node and the destination node 21-3 is k ₃ ¹³ . The proportion of traffic distributed from the source node 21-1 to the destination node 21-3 via the intermediate node 21-4 is k ₄ ¹³ .

Incidentally, a source node is the node 21-p, and the ratio to be distributed to the nodes 21-m when the destination node is the node 21-q represents the k _m ^pq.

On the other hand, when the destination node is the node 21-3, as shown in FIG. 3B, the traffic from the source node 21-1 to the destination node 21-3 is intermediate node 21-2, intermediate node 21-3, The intermediate nodes 21-4 are distributed at a ratio of k ₂ ¹² , k ₃ ¹² , and k ₄ ¹² different from k ₂ ¹³ , k ₃ ¹³ , and k ₄ ¹³ , respectively.

Therefore, the proportion of traffic distributed from the source node 21-1 to the destination node 21-2 is k ₂ ¹² . The proportion of traffic distributed from the source node 21-1 to the destination node 21-2 via the intermediate node 21-3 is k ₃ ¹² . The proportion of traffic distributed from the source node 21-1 to the destination node 21-2 via the intermediate node 21-4 is k ₄ ¹² .

  As described above, in the network system 10, the traffic is distributed to the intermediate node 21-2, the intermediate node 21-3, and the intermediate node 21-4 at a different rate depending on the arrival / departure node pair. Thus, traffic can be effectively distributed at a more optimal rate. As a result, the network congestion rate can be reduced.

[First example of calculation method of variance ratio]
Next, a method of calculating the ratio k _m ^pq.

  First, the network model of the network system 10 is defined as follows.

The network 12 is expressed by an effective graph G (V, E). Here, V represents a set of nodes 21 and E represents a set of links 20. The link 20 from the node iεV to the node jεV is expressed as link (i, j) εE. The usable bandwidth capacity of link (i, j) εE is represented as c _ij, and the usage rate of link (i, j) εE is represented as L _ij . A traffic matrix representing traffic demand d _pq from node p to node q is represented as T = {d _pq }. The network congestion rate is represented by r. As a network model, a pipe model is assumed.

When the network model of the network system 10 is defined as described above, the traffic b _pm ⁽¹⁾ distributed with the node p as the originating node and the intermediate node m as the intermediate node is given by the following equation (10).

The traffic b _pm ⁽²⁾ transferred with the node p as an intermediate node and the intermediate node m as a destination node is given by the following equation (11).

Here, if the link link (i, j) is on the path between the node p and the intermediate node m, a function that takes the value 1 and otherwise takes the value 0 is defined as F _pm ^ij. The usage rate L _ij of i, j) is given by the following equation (12).

However, ψ _pqm ^ij is defined by the following equation (13).

  Further, the difference between the traffic flowing into the node j and the traffic flowing out from the node j is given by the following equation (14).

However, ω _pqm ^j is defined by the following equation (15).

Thus, in the network system 10, the optimal path problem of determining the ratio k _m ^pq, is formulated by the following equation (16a) to (16f).

Expression (16a) represents an objective function for minimizing the network congestion rate r. Equation (16b), with respect to node 21 indicates that all of the sum of the proportion k _m ^pq is 1, the ratio k _m ^pq can take all real numbers satisfying the equation (16b).

Further, Expression (16c) indicates that the total sum of traffic passing through the link link (i, j) is less than or equal to c _ij × r in all the links 20. Equation (16d) shows the flow conservation law.

By solving the above manner formalized formula (16a) to (16f), the ratio k _m ^pq to minimize network congestion ratio r is calculated. Since equations (16a) to (16f) are linear programming problems, they can be solved with a standard LP solver.

Accordingly, the traffic quantity ratio calculation unit 11, by solving the equation (16a) to (16f), and calculates the ratio k _m ^pq.

[Detailed Configuration Example of Traffic Ratio Calculation Device]
FIG. 4 is a block diagram illustrating a detailed configuration example of the traffic amount ratio calculation apparatus 11 of FIG.

  In FIG. 4, the traffic volume ratio calculation device 11 includes a traffic volume input unit 31, a network configuration input unit 32, and a traffic volume ratio calculation unit 33.

A traffic demand d _pq is input to the traffic amount input unit 31 as information related to the traffic of each node 21 from the outside. For example, the traffic demand d _pq may be actually detected from each node 21 or may be an expected traffic amount input by an operator. The traffic volume input unit 31 acquires the input traffic demand d _pq and supplies it to the traffic volume ratio calculation unit 33.

Information representing the configuration of the network 12 such as the capacity c _ij of the usable bandwidth of the link link (i, j) and the effective graph G (V, E) from the outside by an operator or the like (hereinafter referred to as network configuration input unit 32). Network information). The network configuration input unit 32 acquires the input capacity c _ij and network information and supplies them to the traffic amount ratio calculation unit 33.

Based on the traffic demand d _pq supplied from the traffic amount input unit 31, the capacity c _ij supplied from the network configuration input unit 32, and the network information, the traffic amount ratio calculation unit 33 performs the above-described equations (16a) to (16a) to according (16f), and calculates the ratio k _m ^pq for each departure node pair. The traffic amount ratio calculation unit 33 supplies the nodes 21 with ratios k _m ^pq when each is used as a source node.

[Description of processing of traffic volume ratio calculation device]
Next, the ratio calculation process by the traffic amount ratio calculation apparatus 11 in FIG. 4 will be described with reference to the flowchart in FIG.

In step S _< b _> 1, the traffic amount input unit 31 acquires the input traffic demand d _pq and supplies it to the traffic amount ratio calculation unit 33.

In step S <b _> 2, the network configuration input unit 32 acquires the input capacity c _ij and network information and supplies them to the traffic amount ratio calculation unit 33.

In step S3, the traffic volume ratio calculation unit 33 calculates the above-described equation based on the traffic demand d _pq supplied from the traffic volume input unit 31, the capacity c _ij supplied from the network configuration input unit 32, and the network information. according (16a) to (16f), and calculates the ratio k _m ^pq for each departure node pair. Then, the traffic amount ratio calculation unit 33 supplies the nodes 21 with the ratios k _m ^pq when each is set as the originating node, and the processing ends.

[Detailed configuration example of traffic distribution unit]
FIG. 6 shows a detailed configuration example of the traffic distribution unit 22 of FIG.

  In FIG. 6, the traffic distribution unit 22 includes a traffic amount ratio input unit 51, a route table storage unit 52, a packet input unit 53, a packet header analysis unit 54, a next node determination unit 55, and a switch unit 56.

Traffic rate input unit 51, the traffic volume ratio calculating device 11 obtains the proportion k _m ^pq of the own node 21 and source node. Then, the traffic rate input unit 51 supplies the ratio k _m ^pq the route table storage unit 52 for storage. Route table storage unit 52 stores the ratio k _m ^pq.

  The packet input unit 53 acquires a packet whose own node 21 is a source node from among packets input as traffic, and supplies the packet to the packet header analysis unit 54.

  The packet header analysis unit 54 analyzes the header of the packet supplied from the packet input unit 53 and recognizes the destination node 21 of the packet. Then, the packet header analysis unit 54 supplies the packet supplied from the packet input unit 53 and information for specifying the corresponding destination node 21 to the next node determination unit 55.

The next node determination unit 55 reads all the ratios k _m ^pq corresponding to the ^destination node 21 from the route table storage unit 52 based on the information specifying the destination node 21 supplied from the packet header analysis unit 54. Then, the next node determining unit 55, an intermediate node m determined according to the ratio k _m ^pq, determining a path (IP tunnel) to the intermediate node m. The next node determination unit 55 adds route specification information for specifying the route to the determined intermediate node m, information for specifying the intermediate node m, and the like to the packet as headers and supplies the packet to the switch unit 56 together with the route specification information.

  The switch unit 56 transfers the packet via a route corresponding to the route specifying information. As a result, the packet is transferred to the intermediate node m and then transferred to the destination node 21.

[Description of processing of traffic distribution unit]
Next, the traffic distribution process by the traffic distribution unit 22 of FIG. 6 will be described with reference to the flowchart of FIG.

  In step S <b> 11, the packet input unit 53 acquires a packet whose own node 21 is a source node from among packets input as traffic, and supplies the packet to the packet header analysis unit 54.

  In step S12, the packet header analysis unit 54 analyzes the header of the packet supplied from the packet input unit 53, and recognizes the destination node 21 of the packet. Then, the packet header analysis unit 54 supplies the packet supplied from the packet input unit 53 and information for specifying the corresponding destination node 21 to the next node determination unit 55.

In step S13, the next node determination unit 55 determines all the ratios k _m ^pq corresponding to the ^destination node from the route table storage unit 52 based on the information specifying the destination node 21 supplied from the packet header analysis unit 54. read out.

In step S14, the next node determining unit 55, an intermediate node m determined according to the ratio k _m ^pq read in step S13, determines a route (IP tunnel) to the intermediate node m. Then, the next node determining unit 55 adds the route specifying information for specifying the route to the determined intermediate node m, the intermediate node m and the like as a header to the packet, and supplies the packet to the switch unit 56 together with the route specifying information.

  In step S15, the switch unit 56 transfers the packet to the intermediate node m determined in step S14 via the route corresponding to the route specifying information, and the process ends. This packet is then transferred to the destination node 21.

[Experiment showing the effect]
Next, the effect of FTPR will be described with reference to FIGS.

  First, FIGS. 8 and 9 show the configurations of the sample networks # 1 to # 6 used in the experiment showing the effect in the network system 10. FIG. In FIG. 8, black circles represent nodes, and lines represent links.

  As shown in FIGS. 8 and 9, sample network # 1 has 6 nodes (No. of nodes), 24 links (No.of links), and 4.00 average node degree (Ave. node degree). It is a network. The average node order is an average value of the number of links connected to one node.

  Sample network # 2 is a network having 12 nodes, 36 links, and an average node order of 3.00. Sample network # 3 is a network having 12 nodes, 48 links, and an average node order of 4.00. Sample network # 4 is a network having 15 nodes, 56 links, and an average node order of 3.73.

  Sample network # 5 is a network having 20 nodes, 68 links, and an average node order of 3.40. Sample network # 6 is a network having 35 nodes, 100 links, and an average node order of 2.86.

  Next, with reference to FIG. 10, the results of experiments performed using the sample networks # 1 to # 6 as described above will be described.

In this experiment, 100 random values that are uniformly distributed between 80 and 120 are set as the capacity c _ij of the usable bandwidth of the link link (i, j). Also, as the traffic demand d _pq , 100 random values that are uniformly distributed from 0 to 100 were set.

Furthermore, in FTPR and TPR, the link weight is set to be the reciprocal of the capacity c _ij , and the route from the source node to the intermediate node and the route from the intermediate node to the destination node are the distances of the link. Is set to be the shortest route.

  In experiments conducted under the above conditions, when traffic is distributed and transferred using TPR, traffic is distributed and transferred using FTPR, and path control is performed using MPLS-TE. In this case, the network congestion rate r is as shown in FIG.

However, in order to compare the network congestion rate r in the three cases, the network congestion rate r in the case of _FTPR and the network congestion rate r in the case of _FTPR and MPLS-TE _MPLS-TE is the network congestion rate r _{TPR in} the case of _TPR. Is standardized as 1.0. Further, the network congestion rates r _FTPR , r _MPLS-TE , and r _TPR shown in FIG. 10 are average values of values under 100 types of capacity c _ij and traffic demand d _pq set at random.

As shown in FIG. 10, the network congestion rate r _FTPR is much smaller than the network congestion rate r _TPR . Specifically, the range of the network congestion rate r _FTPR is 0.46 to 0.59, and it can be seen that FTPR can reduce the network congestion rate by 40% or more compared to the case of TPR.

Further, as shown in FIG. 10, the network congestion rate r _FTPR is equal to the network congestion rate r _MPLS-TE . Therefore, in order to examine the difference between the network congestion rate r _FTPR and r _MPLS-TE in detail, the deviation δ between the network congestion rate r _FTPR and r _MPLS-TE is defined by the following equation (17), and sample network # 1 In all of # 6 to # 6, ε of deviation δ was 10 ⁻⁶ or less.

  That is, ε is the upper limit value of the accuracy of the solution, particularly for the LP problem of each method. Therefore, it can be seen that FTPR can provide routing performance equivalent to MPLS-TE without using sophisticated routing such as MPLS-TE.

[Second example of calculation method of variance ratio]
In the above description, the ratio k _m ^pq the dispersion was set to take a real value, the value of the ratio k _m ^pq may be to take a 0 or 1. That is, the intermediate node 21 may be fixed to one for the departure / arrival node pair. In this case, the traffic distribution unit 22 can be more easily implemented.

In this case, the ratio k _m ^pq optimal path problem of determining the is formulated by the following equation (18a) to (18f).

That is, the optimal path problem of determining the ratio k _m ^pq the above equation (16e) is changed to (18e).

[Third example of calculation method of variance ratio]
Further, in the above description, assuming a pipe model as a network model, to calculate the ratio k _m ^pq of the dispersion, assuming a hose model as a network model, it may be calculated the ratio k _m ^pq of variance .

In this case, the ratio k _m ^pq optimal path problem of determining the is formulated by the following equation (19a) to (19 g).

In equations (19a) to (19g), α _p represents the total amount of traffic transmitted from the node p, and β _q represents the total amount of traffic flowing into the node q.

  In the equations (19a) to (19g), ψ and τ are defined by the following equations (20) and (21), respectively.

  Further, in the equations (19a) to (19g), π, λ, and ξ are defined by the following equation (22).

As shown in equation (19a) to (19 g), assuming the hose model as network model, in the calculation of the ratio k _m ^pq, instead of traffic demand d _pq, the total amount of traffic alpha _p, beta _q is used . Therefore, the traffic volume ratio input unit 51 receives not the traffic demand d _pq but the total traffic volumes α _p and β _q .

The traffic demand d _pq exists as the square of the number of outgoing and incoming node pairs, that is, the number of nodes 21, but the total traffic amounts α _p and β _q exist as many as the number of outgoing and incoming nodes, that is, the number of nodes 21, respectively. . Therefore, when the hose model is assumed as the network model, the number of parameters input to the traffic amount ratio input unit 51 is smaller than when the pipe model is assumed. However, the path control performance is slightly reduced.

In the above description, the path between the nodes 21 is not distributed and the value of F _pm ^ij is 1 or 0. However, the path between the nodes 21 may be distributed.

Further, it is not necessary for all nodes 21 in the network 12 to distribute traffic by FTPR and transfer it. If there is a node 21 in the network 12 that distributes traffic by FTPR and does not transfer, the node 21 is set as a node 21-p ′, and the following conditional expressions (23) are expressed by the expressions (16a) to (16f), (18a) to (18f), or by adding the equation (19a) to (19 g), it is possible to calculate the ratio k _m ^pq.

Furthermore, whether each node 21 in the network 12, or calculates the ratio k _m ^pq using equation (16a) to (16f), and calculates the ratio k _m ^pq using equation (18a) to (18f) May be determined.

  In the present specification, the term “system” represents the entire apparatus constituted by a plurality of apparatuses.

  The embodiments of the present invention are not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the present invention.

It is a figure explaining operation | movement of TPR. It is a figure which shows the structural example of one embodiment of the network system to which this invention is applied. FIG. 3 is a diagram illustrating traffic distribution in the network system of FIG. 2. It is a block diagram which shows the detailed structural example of the traffic ratio calculation apparatus of FIG. 5 is a flowchart for explaining a ratio calculation process by the traffic amount ratio calculation apparatus in FIG. 4. It is a figure which shows the detailed structural example of the traffic distribution part of FIG. 7 is a flowchart for explaining traffic distribution processing by a traffic distribution unit in FIG. 6. It is a figure which shows the structure of a sample network. It is a figure which shows the table | surface which put together the structure of the sample network. It is a figure which shows an experimental result.

Explanation of symbols

  11 traffic volume ratio calculation device, 12 network, 20-1 to 20-5 link, 21-1 to 21-4 node, 22-1 to 22-4 traffic distribution unit, 55th node determination unit, 56 switch unit

Claims (7)

It is configured by connecting multiple nodes via links, and traffic is transmitted from the source node to the destination node via a predetermined intermediate node among the multiple nodes by TPR (Two-Phase Routing). in a network communication Ru is performed in, for inputted traffic to the calling node, the calling node recognizes the information for specifying the destination node from the input packet, to distribute the traffic according to the destination node the intermediate node determined according percentage, by adding information of the intermediate node to the packet by transferring the packet to the intermediate node, the packet to the destination node via the intermediate node When transmitting, in the control device for controlling the traffic communication,
Calculation means for calculating a ratio of distributing the traffic with each node as the intermediate node for each combination of the source node and the destination node, based on the configuration of the network and the available bandwidth capacity of the link When,
A control device comprising: supply means for supplying to each of the plurality of nodes a ratio for distributing the traffic for each destination node when that node becomes a source node . The control device according to claim 1, wherein the calculation unit calculates the ratio based on a configuration of the network, a capacity of a usable bandwidth of the link, and information on the traffic. The control device according to claim 2, wherein the calculation unit calculates the ratio by assuming a pipe model as a model of the network. The control device according to claim 1, wherein the calculation unit calculates a ratio assuming a hose model as a model of the network. It is configured by connecting multiple nodes via links, and traffic is transmitted from the source node to the destination node via a predetermined intermediate node among the multiple nodes by TPR (Two-Phase Routing). in a network communication Ru is performed in, for inputted traffic to the calling node, the calling node recognizes the information for specifying the destination node from the input packet, to distribute the traffic according to the destination node the intermediate node determined according percentage, by adding information of the intermediate node to the packet by transferring the packet to the intermediate node, the packet to the destination node via the intermediate node In the control method of the control device for controlling the traffic communication when transmitting,
A calculation step of calculating a ratio of distributing the traffic with each node as the intermediate node for each combination of the source node and the destination node, based on the configuration of the network and the available bandwidth capacity of the link. When,
And a supply step of supplying, to each of the plurality of nodes, a ratio of distributing the traffic for each destination node when that node becomes a source node . A node constituting a network configured by connecting a plurality of nodes via a link, and the traffic inputted by TPR (Two-Phase Routing) among the plurality of nodes in the network In a node that transmits to a destination node in the network via a predetermined intermediate node,
A destination node recognition means for recognizing information identifying the destination node from an input packet;
The traffic is transmitted based on a ratio of distributing the traffic using each node in the network as the intermediate node obtained for each combination of the node as the source node to which the traffic is input and the destination node. A determination means for determining an intermediate node;
A node comprising: transmission means for adding the information on the intermediate node determined by the determination means to the packet and transmitting the packet to which the information on the intermediate node is added to the intermediate node. A node constituting a network configured by connecting a plurality of nodes via a link, and the traffic inputted by TPR (Two-Phase Routing) among the plurality of nodes in the network In a node transmission method for transmitting to a destination node in the network via a predetermined intermediate node,
A destination node recognition step for recognizing information identifying the destination node from the input packet;
The traffic is transmitted based on a ratio of distributing the traffic using each node in the network as the intermediate node obtained for each combination of the node as the source node to which the traffic is input and the destination node. A decision step for determining intermediate nodes;
A transmission method comprising: adding information on the intermediate node determined by the processing of the determination step to the packet, and transmitting the packet to which the information on the intermediate node is added to the intermediate node .

Images