JP4737325B2 - Information processing apparatus and information processing method - Google Patents

Description

  The present invention relates to an information processing apparatus and an information processing method. More specifically, the present invention relates to a technique for estimating the amount of traffic flowing through a network link.

  There is a technique for estimating a link transmission time of a network and estimating a time required for a packet to pass through an arbitrary route (see, for example, Patent Document 1 or Patent Document 2). According to these techniques, the host sends a search packet, and performs a process of obtaining a round trip time until the packet returns to the host via a plurality of routers in the network. The host can determine round trip times for various routes in the network by repeatedly executing this process. In addition, the host can establish simultaneous linear equations using the transmission time in each link as a variable from the round trip time, and obtain an approximate solution by the learned neural network. The approximate solution is calculated as the transit time of a packet on an arbitrary route. If these techniques are used, since the transmission time of each link of the network is required, it is possible to perform routing with a small transmission delay by estimating a link with a large traffic transmission amount and avoiding the link.

JP 2000-32038 A JP 2000-49863 A

  However, when routing is performed using these techniques, at least the following four points are problematic.

  First, there is a problem that a load on the network may be increased by causing a host to send a plurality (or a large number) of search packets in the network.

  Second, there is a problem that it takes time for the host to prepare data for calculating the transmission time of the link. This is because it is necessary for the host to send a large number of search packets in order to determine the round trip time of the route and to wait until the search packets return. This problem becomes more pronounced as the network becomes larger and more complex.

  Thirdly, the larger the network, the larger the error from the actual value of the approximate solution obtained by the simultaneous linear equations. When the host causes a plurality of search packets to flow through the network, the load on each router in the network increases, and a particularly large load is applied to a router with a large number of connected links. If the load on the router increases and it takes time to process the packet by the router, the round-trip time of the packet cannot be said to be strictly composed of the link transmission time. In addition to this, the transmission time of the link changes from time to time. For example, the larger the network, the more search packets are required. Therefore, the time at which each search packet is processed is different.

  Fourth, there is a problem that it is impossible to respond quickly to changes in the network structure. When the network structure changes due to increase / decrease in the number of routers or links, change of links, etc., the host has to re-determine routes necessary and sufficient for establishing simultaneous linear equations.

  Therefore, the present invention has been made in view of the above problems, and the object of the present invention is to reduce the load on the network mainly when estimating the amount of traffic flowing through the links constituting the network. It is to provide a technology that can.

  In order to solve the above problem, according to one aspect of the present invention, a plurality of nodes are connected to each other by a link and connected to a network, and two nodes are arbitrarily selected from the plurality of nodes. An information processing apparatus that estimates the amount of traffic that is the amount of packets that flow through the link between each pair of nodes obtained in the same manner, and a link information acquisition unit that acquires link information that indicates a connection relationship between the nodes by the link, A processing amount calculation unit for calculating the amount of packets flowing through each node as each traffic processing amount corresponding to the node, and the packet corresponding to the packet amount flowing between each node pair based on the link information and each traffic processing amount A matrix having a ratio of the amount of packets flowing through the node in each element, and each contribution matrix corresponding to the node And an estimation unit that estimates the traffic amount by allocating each traffic processing amount corresponding to each node to each node pair based on each contribution matrix corresponding to each node. An information processing apparatus is provided.

  The estimation unit may use a value proportional to each element value of the contribution matrix as each traffic processing amount allocated to each node pair.

  The estimation unit may allocate the traffic processing amount corresponding to the node to each node pair in order from the node having the largest sum of the element values of the contribution matrix.

  The estimation unit may allocate the traffic processing amount corresponding to the node to each node pair in order from the node having the largest number of non-zero elements in the contribution matrix.

  The estimation unit calculates a traffic processing amount that is a packet amount flowing through each node from the estimated traffic amount, determines a magnitude relationship between the calculated traffic processing amount and the traffic processing amount calculated by the processing amount calculation unit, When it is determined that the calculated traffic processing amount is larger, the traffic processing amount allocated to the node pair including the node is decreased, and when it is determined that the calculated traffic processing amount is smaller, the node is included. The traffic volume may be allocated again by increasing the traffic processing volume allocated to the node pair.

  As described above, according to the present invention, it is possible to reduce the load on the network mainly when estimating the amount of traffic flowing through the links constituting the network.

It is a figure for demonstrating the structure of a network. It is a figure which shows the function structure of the information processing apparatus which concerns on this embodiment. It is a figure which shows the structural example of a network and information processing apparatus. It is a flowchart which shows the flow of a process from acquisition of the routing information which the information processing apparatus performs to calculation of a contribution matrix. It is a figure which shows the example of data of the traffic processing amount for every router. It is a figure which shows the example of data of a routing table. It is a flowchart which shows the flow of the traffic amount estimation process which information processing apparatus performs.

  Exemplary embodiments of the present invention will be described below in detail with reference to the accompanying drawings. In addition, in this specification and drawing, about the component which has the substantially same function structure, duplication description is abbreviate | omitted by attaching | subjecting the same code | symbol.

<First Embodiment>
[Description of overview]
FIG. 1 is a diagram for explaining a network configuration. As shown in FIG. 1, in the network 200 ′, each node (nodes N1 to N5) cooperates to transmit a packet from a transmission source to a transmission destination. A packet is used as a unit of data transmitted from a transmission source to a transmission destination. In addition, here, the network 200 ′ is configured by five nodes, but the number of nodes is not particularly limited. As an example of a node, a router can be used. However, the node is not particularly limited to a router, and a device that constitutes the network 200 ′ and transmits and receives packets is used as appropriate. Can do.

  A node pair indicates a set of arbitrary nodes, and does not mean only nodes that are directly connected by a link. For example, not only the set of the node N1 and the node N2, the set of the node N3 and the node N4, but also the set of the node N1 and the node N4, the set of the node N3 and the node N5, and the like are also referred to as node pairs. When a router is used as an example of a node, the above node pair is particularly referred to as a router pair. In this embodiment, the traffic amount for each node pair is estimated using the traffic processing amount of the node and the contribution matrix. The traffic processing amount of a node is the amount of packets processed by the node, for example, and means the amount of packets per unit time processed by the node, for example. The contribution matrix will be described later. The traffic amount for each node pair is, for example, the amount of packets transmitted / received between the node pairs, and means the amount of packets per unit time transmitted / received between the node pairs, for example.

  However, a number of estimation amount candidates are calculated only from the information of the traffic processing amount and the contribution matrix of this node. Therefore, candidates for the estimated amount are narrowed down by estimating the traffic amount so as to satisfy the condition of “allocating the traffic processing amount of the node so that the traffic amount for each node pair is as equal as possible”. This condition is based on the idea that the larger the network 200 ′, the more uniform the traffic amount for each node pair.

[Explanation of contribution matrix]
The contribution matrix is a matrix created for each router. For example, when the contribution matrix of the node k is referred to, the {i, j} element is k for the traffic amount starting from the node i and ending at the node j. It shows the ratio of traffic volume passing through. If all packets starting from node i and ending at node j pass through node k, the {i, j} element is “1”, and node k is not involved (the packet starting from node i and ending at node j is not In the case of not passing through the node k), it becomes “0”. If the routing method in the network 200 is determined, the entire contribution matrix in the network 200 is also uniquely determined. For example, the contribution matrix delta ⁵ of node N5 shown in FIG. 1 is shown as follows.

Here, for route selection, the shortest route method is adopted, and when there are a plurality of routes, equal distribution is made for each route at a branch point. Further, the traffic volume between the node N2 and the node N4 is divided into a route (a route that passes through the node N2, the node N3, and the node N4 in this order) and a route (a route that passes through the node N2, the node N5, and the node N4 in this order). Since it is divided into two, the {2, 4} and {4, 2} elements are each “0.5”. On the other hand, since the packet transmitted / received between the node N1 and the node N4 always passes through the node N5, the {1, 4} and {4, 1} elements are “1”. The {i, j} element is represented as Δ ⁵ _{i, j} .

Further, if a matrix having {i, j} elements that start from the node i and end at the node j is represented as T _{i, j} , the traffic processing amount t _k passing through the node _k is as follows: It is shown in the form of the sum of products.

  In the following, as an example of a network 200 ′ composed of nodes, a case where a network 200 composed of routers (for example, see FIG. 2 or FIG. 3) is used will be described.

[Description of configuration]
FIG. 2 is a diagram illustrating a functional configuration of the information processing apparatus according to the present embodiment. With reference to FIG. 2, the functional configuration of the information processing apparatus according to the present embodiment will be described.

  As illustrated in FIG. 2, the information processing apparatus 100 is connected to a network 200. The network 200 is configured by connecting a plurality of routers to each other through links (see FIG. 3). The information processing apparatus 100 estimates a traffic amount that is a packet amount flowing through a link between each router pair obtained by arbitrarily selecting two routers from a plurality of routers. The information processing apparatus 100 includes a storage unit 110, a control unit 120, and a communication unit 130. As the packet amount, the number of packets may be used, or the data amount of the packet may be used.

  The storage unit 110 is configured by a storage device such as an HDD (Hard Disk Drive), for example. The storage unit 110 has a function of storing a program executed by the control unit 120 and a function of storing data used by the program.

  The control unit 120 is configured by, for example, a CPU (Central Processing Unit) or the like, and the CPU stores a program stored in the storage unit 110 in a RAM (Random Access Memory) or the like, and the expanded program is executed. That function is realized. The control unit 120 includes a route information acquisition control unit 121, a link information acquisition unit 122, a contribution matrix calculation unit 123, an estimation unit 124, a processing amount calculation unit 125, and the like.

  The communication unit 130 is configured by a communication device or the like, for example, and can communicate with a router in the network 200. The communication unit 130 includes a route information transmission request transmission unit 131, a route information reception unit 132, a traffic data transmission request transmission unit 133, a traffic data reception unit 134, and the like.

  The traffic data receiving unit 134 periodically receives the traffic processing amount from all the routers in the network 200 in the MIB (Management Information Base) information of SNMP (Simple Network Management Protocol). The traffic data reception unit 134 outputs the received traffic processing amount to the processing amount calculation unit 125.

  The processing amount calculation unit 125 calculates the amount of packets that have flowed through each router within the same period from the traffic processing amount output from the traffic data reception unit 134 as each traffic processing amount corresponding to the router. The start time of the same period is preferably the same, but the length of the period may be any length, for example, 10 seconds, 5 minutes, 1 hour. The processing amount calculation unit 125 outputs the calculated traffic processing amount to the estimation unit 124.

  The route information receiving unit 132 receives the routing table (route information) from all routers in the network 200 and outputs the received routing table to the link information acquiring unit 122.

  The link information acquisition unit 122 acquires link information indicating a connection relationship by links between routers in the network 200 based on the routing table output from the route information reception unit 132, and uses the acquired link information as a contribution matrix. Output to the calculation unit 123. When the necessary link information cannot be acquired, the link information acquisition unit 122 outputs identification information for identifying a router that requests transmission of the routing table to the route information acquisition control unit 121.

  The route information acquisition control unit 121 performs control for acquiring a routing table from a router in the network 200. For example, the route information acquisition control unit 121 instructs the route information transmission request transmission unit 131 to transmit a route information transmission request to the router identified by the identification information output from the link information acquisition unit 122.

  The route information transmission request transmission unit 131 transmits a route information transmission request in accordance with an instruction from the route information acquisition control unit 121. The router that has received the route information transmission request transmits a routing table, and the route information reception unit 132 receives the transmitted routing table.

  Based on the link information and each traffic processing amount, the contribution matrix calculation unit 123 sets a matrix having, in each router, a ratio of the amount of packets flowing through the router to the amount of packets flowing between each pair of routers. Calculate as each corresponding contribution matrix. The contribution matrix calculation unit 123 outputs the calculated contribution matrix to the estimation unit 124.

  The estimation unit 124 estimates the traffic amount by allocating each traffic processing amount corresponding to each router to each router pair based on each contribution matrix corresponding to each router. The estimation unit 124 preferably allocates the traffic volume of the estimated router pair so that the traffic volume of each router pair is as uniform as possible (* 1). This is based on the idea that the traffic volume for each router pair becomes uniform as the scale of the network 200 increases. In order to realize this, for example, the estimation unit 124 can use a value proportional to each element value of the contribution matrix as each traffic processing amount allocated to each router pair.

This can be received against the condition (* 1). However, when traffic between router pairs {i, j} (packets flowing between router pairs {i, j}) is routed through a plurality of paths, the contribution matrix Δ ^k _i for the router k on the branched path _{, J} is less than 1. Therefore, in the allocation of the traffic processing amount for the router k, the involvement of the router k with respect to the traffic between the router pairs {i, j} is smaller than the involvement of the router k with respect to the traffic between the router pairs corresponding to the other components “1”. Less. This is a result of allocating the traffic processing amount evenly as much as possible. Since the traffic processing amount for the router of another route is allocated to the router pair {i, j}, it is allocated to the router pair {i, j}. It doesn't mean that traffic volume is low.

  Further, the estimation unit 124 may allocate the traffic processing amount corresponding to the router to each router pair in order from the router having the largest sum of the element values of the contribution matrix. Or the estimation part 124 is good also as allocating the traffic throughput corresponding to the router to each router pair in an order from the router with the largest non-zero element number (S) of a contribution matrix. Although it is conceivable that the traffic amount for each router pair becomes uniform as the scale of the network 200 increases, the traffic amount in which the routers with small S are router pairs among the traffic amounts flowing through the entire network 200 is relatively small. This is based on the assumption (* 2). By placing the allocation to routers with small S inferior, the effect of reducing the amount of traffic with routers having small S as router pairs can be obtained.

  The estimation unit 124 calculates a traffic processing amount that is a packet amount flowing through each router from the estimated traffic amount, and determines a magnitude relationship between the calculated traffic processing amount and the traffic processing amount calculated by the processing amount calculation unit 125. May be. When the estimation unit 124 determines that the calculated traffic processing amount is larger, the estimation unit 124 decreases the traffic processing amount allocated to the router pair including the router, and when it determines that the calculated traffic processing amount is smaller. The traffic processing amount allocated to the router pair including the router may be increased. In this way, the traffic amount may be allocated again. Then, the difference between the calculated traffic processing amount and the traffic processing amount calculated by the processing amount calculation unit 125 can be reduced, and the difference is further reduced if the traffic amount allocation is repeated again. be able to.

  The estimation unit 124 outputs the estimated traffic amount as estimation data. If there is a traffic processing amount that is not output from the processing amount calculation unit 125, the estimation unit 124 instructs the traffic data transmission request transmission unit 133 to transmit a traffic data transmission request to the router corresponding to the traffic processing amount. Can be output.

  The traffic data transmission request transmission unit 133 transmits a traffic data transmission request to the router in accordance with the instruction output from the estimation unit 124. The router that has received the traffic data transmission request transmits the traffic processing amount, and the transmitted traffic processing amount is received by the traffic data receiving unit 134.

  FIG. 3 is a diagram illustrating a configuration example of a network and an information processing apparatus. FIG. 4 is a flowchart showing a flow of processing from acquisition of route information to calculation of a contribution matrix executed by the information processing apparatus. FIG. 5 is a diagram illustrating an example of traffic processing amount data for each router. FIG. 6 is a diagram illustrating an example of data in the routing table.

[Description of operation]
With reference to FIGS. 2-7, the operation example of the traffic amount estimation process which the information processing apparatus 100 performs is demonstrated. For convenience of explanation, it is assumed that the number of routers in the network 200 is N. Also, as an example of route information, a routing table held by a router is used. FIG. 4 is a flowchart showing a process flow from acquisition of route information to calculation of a contribution matrix. First, with reference to FIG. 4, the process from the acquisition of the route information to the calculation of the contribution matrix by the information processing apparatus 100 will be described.

  The following steps S101 to S104 relate to the calculation of the contribution matrix Δ of each router in the network 200. First, the route information receiving unit 132 acquires a routing table possessed by each router (step S101). The link information acquisition unit 122 acquires link information of each router based on the routing table received by the route information reception unit 132 (step S102). This link information indicates to which router each router has a link, that is, to which router each router can transmit a packet.

  The route information acquisition control unit 121 determines whether data is prepared, that is, whether link information of each router is acquired (step S103). If the path information acquisition control unit 121 determines that the data is not complete ("No" in step S103), the process returns to step S101. When returning to step S101, the route information transmission request transmission unit 131 transmits a route information transmission request packet to the router having the routing table required by the link information acquisition unit 122, and the route information reception unit 132 In response to the route information transmission request packet, the route information is received from the router.

When the route information acquisition control unit 121 determines that the data is available (“Yes” in step S103), the contribution matrix calculation unit 123 determines that the contribution matrix {Δ ^α : α∈1,. ., N} is calculated. Steps S101 to S103 are executed whenever there is a change in the routing table, for example.

  FIG. 7 is a flowchart showing the flow of processing from acquisition of traffic data and a contribution matrix to estimation of traffic volume. First, with reference to FIG. 7, the processing from the acquisition of traffic data and the contribution matrix to the estimation of the traffic amount by the information processing apparatus 100 will be described.

The estimation unit 124 acquires the traffic processing amount calculated by the processing amount calculation unit 125 (step S201). The traffic processing amount here means the amount of packets processed by each router in the same period, that is, the traffic processing amount {t _α : α∈1,..., N} of each router in the same period. is there. The traffic processing amount in the same period of each router is used as the traffic processing amount. However, the same period does not have to be exactly the same period, and may be substantially the same period. Subsequently, the estimation unit 124 acquires the contribution matrix Δ calculated by the contribution matrix calculation unit 123 (step S202).

  The estimation unit 124 determines whether or not the data acquired in step S201 and step S202 is complete (step S203). If the estimation unit 124 determines that the data is not complete (“No” in step S203), the estimation unit 124 returns to step S201. When the estimation unit 124 determines that the traffic processing amount is not uniform, the traffic data transmission request transmission unit 133 transmits the traffic data to the router having the traffic data that is the basis of the traffic processing amount that is not uniform. Send a request. The traffic data receiving unit 134 receives traffic data as a response to the traffic data transmission request and outputs the traffic data to the processing amount calculation unit 125. The processing amount calculation unit 125 calculates the traffic processing amount from the traffic data and outputs it to the estimation unit 124.

  When the estimation unit 124 determines that the contribution matrix is not complete, the estimation unit 124 waits for the contribution matrix calculation unit 123 to output the contribution matrix.

If the estimation unit 124 determines that the data is complete (“Yes” in step S203), the estimation unit 124 estimates the traffic amount of the router pair in step S204 and subsequent steps. Estimation unit 124 is used to allocate the traffic throughput of each router, delta ^alpha copies: providing a ^{{Δ fα 1 ≦ α ≦ N} } for each router. Element values of delta ^f.alpha becomes the factor used in allocating the traffic throughput of the router α traffic volume of router pair. This element value is appropriately corrected in step S210.

The estimation unit 124 calculates a non-zero number of elements S for the contribution matrix Δ of each router, and based on the calculated S, the router order a [1] = k1, a [2] = k2,. a [N] = kN is determined. Here, a [1] to a [N] are arrays in which values can be set, for example, and k1 to kN are values for identifying routers. Estimation unit 124, for example, from greater router S order k1, k2, · · ·, and kN, the S respectively _{S k1,} S k2, · · _·, expressed as _{S kN.} The estimation unit 124 allocates traffic processing amounts in the order of k1, k2,.

The estimation unit 124 allocates the traffic processing amount t _k1 of the router k1 according to the distribution of ^Δfk1 (step S205). Estimation unit 124, for example, allocated at a rate proportional to the delta ^fk1 _{i, j} traffic throughput _{t k1} router k1 in accordance with the distribution of delta ^fk1. The traffic amount obtained by allocating the traffic processing amount is represented by a matrix as T in Expression (1). Here, the amount of traffic obtained by allocating traffic throughput t _k1 is decided to express the matrix T ^k1 of N × N.

  In the repetitive processing in step S206, the estimation unit 124 allocates the traffic processing amounts of the remaining routers k2, k3,. Here, a [2] = k2, a [3] = k3,..., A [N] = kN, and the initial processing executed when step S206 is started is n = 2 (n = 2 ), The continuation condition for continuing step S206 is n ≦ N, and the termination process executed each time step S206 is repeated is n = n + 1 (n is incremented by 1).

The estimation unit 124 calculates the traffic processing amount ta _[n] ^' to be allocated to the router specified by a [n] in the iterative process of Step S206 (Step S207). The calculation formula is as follows, for example.

The second term on the right-hand side of equation (2) indicates the amount to be processed in the router a [n] out of the traffic amount T ⁿ⁻¹ allocated to the router a [n−1], and the amount obtained by subtracting this from t _n Becomes the traffic processing amount t _{a [n]} ^′ allocated for the router a [n].

The estimation unit 124 allocates the traffic processing amount ta [n] ' according to the distribution of [ Delta ] fa [n] by the same method as the allocation processing in step S205 (step S208). Estimation unit 124, for example, allocated at a rate proportional to Δ fa [n] i, j traffic throughput t a [n] 'according to the distribution of Δ fa [n]. The estimation unit 124 adds each value obtained by allocating the traffic processing amount ta [n] ' to each element of Tn-1 . Traffic after updating in this way the T n.

The estimating unit 124 determines whether or not the difference between the traffic processing amount of each router calculated from ^TN obtained by the repetition processing in step S206 and t _α is within a predetermined range (step S209). Therefore, first, the estimating unit 124, for example, the traffic volume of each router from T ^N: calculating a _{{u α 1 ≦ α ≦ N} } as follows.

For example, the estimation unit 124 determines whether or not the difference between the calculated u _α and the traffic processing amount t _α (or d obtained by the square of the difference) is within a predetermined range. The estimation unit 124 can use, for example, the following expression as d obtained by the square of the difference between the calculated u _α and the traffic processing amount t _α .

When the estimation unit 124 determines that the difference between the traffic processing amount of each router calculated from ^TN and t _α is within a predetermined range, the estimation unit 124 outputs the ^TN as an estimated amount (estimated data). To do. If the estimation unit 124 determines that the difference between the traffic processing amount of each router calculated from ^TN and t _α is not within the predetermined range, the estimation unit 124 proceeds to step S210. Estimating unit 124 to correct the allocation of traffic throughput of each node pair, updating the delta ^f (step S210), and continues the process returns to step S205.

[Description of specific examples of operation]
A specific example of the traffic amount estimation process performed by the information processing apparatus 100 will be described with reference to FIGS. The network 200 is described as being configured by routers R1 to R6 as shown in FIG. 3, but the configuration of the network 200 is not limited to this. The information processing apparatus 100 illustrated in FIG. 3 corresponds to the information processing apparatus 100 illustrated in FIG. Assume that a packet is not sent to the outside of routers R1 to R6, and packets are transmitted and received through links connected to each of routers R1 to R6. Also, a candidate traffic amount between router pairs used in the following estimation is expressed as a 6 × 6 matrix T.

  In the information processing apparatus 100, the traffic amount between the router pairs is estimated from the traffic processing amount processed by the routers R1 to R6. In the information processing apparatus 100, the traffic data reception unit 134 periodically receives the traffic processing amount from the MIB information of SNMP from the routers R1 to R6. The data is output to the processing amount calculation unit 125 received by the traffic data reception unit 134, and the processing amount calculation unit 125 acquires the traffic processing amount within the same period.

For example, it is assumed that the traffic processing amount {t _α : 1 ≦ α ≦ 6} of each router observed at a certain time is acquired as information as shown in FIG. 5 by the processing amount calculation unit 125. However, it is assumed that the traffic processing amount by the packet starting from the router α and ending at the router α is not included. Therefore, the diagonal component of T is “0”, that is, {T _{i, i} = 0: 1 ≦ i ≦ 6}. The traffic processing amount acquired by the processing amount calculation unit 125 is output to the estimation unit 124.

  Next, the process in which the contribution matrix calculation unit 123 calculates the contribution matrix will be described. In step S101, the route information receiving unit 132 receives the respective routing tables from the routers R1 to R6, and the link information acquiring unit 122 obtains the link information based on the routing table received by the route information receiving unit 132. get. Here, the link information acquisition unit 122 selects a route based on the shortest route method, and when there are a plurality of shortest routes, selects one route for each node pair. As an example, a case where the link information acquisition unit 122 acquires link information of the router R5 will be described. When the routing table possessed by the router R5 is as shown in FIG. 6, the link information acquisition unit 122 links the two routers R4 and R6 to the router R5 based on the Next Hop Node. It can be determined that That is, the link information acquisition unit 122 can acquire information for identifying the router R4 and the router R6 from the routing table as the link information of the router R5.

If it is possible to link information acquiring unit 122 acquires the link information in the same manner for each router, the condition of step S103 is satisfied, the contribution matrix calculating unit 123 calculates a delta ^alpha in step S104. If the route selection method is based on the shortest route method, the following matrix can be cited as the contribution matrix {Δ ^α : 1 ≦ α ≦ 6} of the routers R1 to R6.

Also in the following description, Δ ^{1 to} Δ ⁶ are the contribution matrices of the routers R1 to R6, respectively. Then, the contribution matrices Δ ¹ to Δ ⁶ calculated by the contribution matrix calculation unit 123 are output to the estimation unit 124.

Next, a specific example of the operation of the estimation unit 124 will be described with reference to FIG. 7 (refer to other figures as appropriate). As described above, the traffic throughput of _t 1 ~t ₆ are acquired router R1~ router R6 by the processing amount calculating section 125 in step S201, acquires the contribution matrix Δ ¹ ~Δ ⁶ by the contribution matrix calculating unit 123 in step S202 If so, the condition of step S203 is satisfied.

Here, what the contribution matrix delta ^alpha is also symmetric matrix, _{^{i.e., {Δ α i, j =}} Δ α j, i: 1 ≦ i ≦ 6, 1 ≦ j ≦ 6} is because, for the sake of simplicity, in the following The estimated traffic of the router pair does not distinguish the direction of the router pair. Therefore, T to be estimated can be shown as follows.

If to be estimated T is (5) shown is as equation estimation unit ₁₂₄ will estimate the total of 15 components of T 1, 2, · · · _{T 5, 6.}

In step S204, the estimation unit 124 obtains S for each router. For example, since the number of non-zero (eg, “1”) elements of Δ ¹ for the router R1 is “10”, the estimation unit 124 sets the value of S ₁ that is each S for the router R1 to “10”. I ask. Similarly, the estimation unit 124 obtains the values of S _{2 to} S ₆ that are S for the routers R2 to R6 as “10”, “12”, “12”, “14”, and “22”. Therefore, the estimation unit 124 determines the router R6, the router R5, the router R3, the router R4, the router R1, and the router R2 as the order in which the traffic processing amount is allocated. Here, since S ₁ = S ₂ and S ₃ = S _4, it is necessary to determine which of the router R1 and the router R2 or the router R3 and the router R4 is advanced in order. However, the method of determination is not particularly limited.

In step S205, the estimation unit 124 allocates traffic processing amount t ₆ = 61 (see FIG. 5). Estimation unit 124, a copy of the ^{Δ α {Δ fα: 1 ≦} α ≦ N} a when prepared for each router, delta ^f6 are as shown below.

As shown in Expression (6), the element values of ^Δf6 are all “1”. Also, since only 15 components of T _{1,2 to} T _5,6 are considered now, the estimation unit 124 sets the number of non-zero elements corresponding to the components, that is, the element value to be allocated to “11”. And decide. Accordingly, when t ₆ / (element value to be allocated) = 61/11 = 5.55 is allocated, the estimation unit 124 becomes T ⁶ shown below.

Next, in step S206, the estimation unit 124 sequentially allocates traffic processing amounts t ₅ , t ₃ , t ₄ , t ₁ , and t ₂ .

In step S207, the estimation unit 124 calculates the traffic processing amount t ₅ ^′ allocated for the router R5. According to Expression (2), the estimation unit 124 has t ₅ ^′ = t ₅ −Σ ⁶ _{i = 1} Σ ⁶ _{j = 1} (Δ ^f5 _{i, j} T ⁶ _{i, j} ) = 34−27.73 = 6.27. And calculate.

In step S208, the estimation unit 124 allocates t ₅ ^′ = 6.27 calculated in step S207 in the same manner as the traffic processing amount t ₆ ^′ for the router R6. The non-zero values of the elements of Δ ^f5 _{i, j} are all “1”, and the number of elements to be allocated is Δ ^f5 _1,5 , Δ ^f5 _2,4 , Δ ^f5 _2,5 , Δ ^f5 _3,5 , Δ ^f5 _4,5 , Δ ^f5 _4,6 , Δ ^f5 ₄ , 7, the estimation unit 124 calculates t ₅ ^′ / (element value to be allocated) = 6.27 / 7 = 0.90. _Are added to T ⁶ _1,5 , T ⁶ _2,4 , T ⁶ _2,5 , T ⁶ _3,5 , T ⁶ _4,5 , T ⁶ _4,6 , T ⁶ ₄ , ₇ . As a result, T ⁵ is as shown below.

  Similarly, the estimation unit 124 distributes the traffic processing amount for the router R3, the router R4, the router R1, and the router R2. As a result of the distribution, the estimation unit 124 can acquire an estimated value T as shown below.

When the estimated value T is obtained, the iterative process of step S206 is completed. In step S209, the estimation unit 124 calculates u _α from Expression (3), and calculates a difference d between u _α and t _α by Expression (4). Here, the estimation unit 124 estimates T as the traffic amount of the router pair when the following expression is satisfied.

The estimation unit 124 calculates u _α obtained by the equation (3), and can obtain d = 1.1 from the u _α by the equation (4). Since d = 1.1 does not satisfy the condition (for example, Expression (10)) as the traffic amount of the router pair, the estimating unit 124 proceeds to step S210.

Estimation unit 124, in step S210, modifies the delta ^f from T of formula (9). Here, the estimation unit 124 obtains a router k that maximizes | c _α | from Expression (11), and is expressed by Expression (12) for all elements in the k-th row or the k-th column of all Δ ^f. by multiplying the value to update the delta ^f. As a result, in the traffic processing amount allocation cycle for the next router, if the traffic processing amount for the next router is large, it is small, and conversely, if the traffic processing amount is small, the large amount is allocated. Will be able to.

  In this way, as a result of repeating steps S205 to S210 153 times, the estimation unit 124 obtained the following result as the estimated value T of the traffic amount of the router pair.

  When the estimated value T is a value as shown in Expression (13), the estimation unit 124 can obtain d = 0.0000096 from Expression (3) and Expression (4). When d = 0.0000965, the estimation unit 124 determines in Step S209 that Expression (10) is satisfied, and outputs T indicated by Expression (13) as an estimated value of the traffic amount of the router pair.

Here, the estimated value T of the traffic amount of the router pair varies from T _1,2 = 0.45 to T _2,5 = 8.46. This is because the traffic processing amount is allocated so as to match the contribution matrix Δ and the traffic processing amount is allocated based on the above assumption (* 2). Big router S, for example, traffic throughput of the routers R6 performing the first allocation is assigned according to element values of delta ^6. Therefore, the width of the element values of T are between nonzero elements of delta ⁶ is small. However, the amount of traffic between routers with small S, for example, between the router R1 and the router R2, remains as small as T _1,2 = 0.45 from the assumption (* 2).

[Description of effects]
As described above, according to the first embodiment, an estimated value of the traffic amount of the router pair can be obtained from the traffic processing amount of each router within the same period so that the traffic amount of the router pair becomes as equal as possible. As a result, it is possible to know the amount of traffic flowing through the link, avoid routing to links with high traffic volume, and conversely avoid congestion by encouraging routing to links with low traffic volume, and perform routing with low delay Become.

  The data necessary for estimating the traffic volume is the traffic processing volume and the contribution matrix of each router. The traffic processing amount and the contribution matrix can be acquired by receiving a predetermined amount of packets from each router. Compared to the technology that increases the lifetime of search packets one by one and sends them and obtains the round trip time of the route, the amount of packets that flow in the network can be reduced to obtain the data required for traffic estimation. it can.

  Furthermore, if the traffic processing amount and the contribution matrix are aligned, it is possible to proceed to estimation work, and therefore the time required to estimate the traffic amount can also be reduced.

  Since the traffic processing amount in the same period of each router is used, the traffic amount of the estimated node pair can be estimated with a consistent value.

  A routing table can be acquired by a routing protocol, and a contribution matrix can be uniquely determined from the routing table and a routing rule by each router (for example, equally distributed to routes of the same metric). Therefore, it is possible to cope with a change in the network structure when the routing table is completed.

[Modification]
In the above, the preferred embodiments of the present invention have been described with reference to the accompanying drawings, but it goes without saying that the present invention is not limited to such examples. It will be apparent to those skilled in the art that various changes and modifications can be made within the scope of the claims, and these are naturally within the technical scope of the present invention. Understood.

  For example, the information processing apparatus according to the present embodiment can be used by connecting to any network constituted by each node. For example, the amount of traffic can be estimated by connecting to a P2P virtual network structure or a wireless ad hoc network. Further, since the information processing apparatus according to the present embodiment can know the degree of congestion of each link from the estimated value, it is also possible to estimate the link transmission time.

DESCRIPTION OF SYMBOLS 100 Information processing apparatus 110 Storage part 120 Control part 121 Path information acquisition control part 122 Link information acquisition part 123 Contribution matrix calculation part 124 Estimation part 125 Processing amount calculation part 130 Communication part 131 Path information transmission request transmission part 132 Path information reception part 133 Traffic data transmission request transmission unit 134 Traffic data reception unit 200 Network

Claims (6)

Traffic that is the amount of packets that are connected to a network configured by connecting a plurality of nodes to each other by links, and that flow through the link between each pair of nodes obtained by arbitrarily selecting two nodes from the plurality of nodes. An information processing apparatus for estimating a quantity,
A link information acquisition unit for acquiring link information indicating a connection relationship between the nodes by the link;
A throughput calculator that calculates the amount of packets that flow through each node within the same period as each traffic throughput corresponding to that node;
Based on the link information and the traffic processing amount, each contribution matrix corresponding to the node includes a matrix having, as elements, the ratio of the packet amount flowing through the node to the packet amount flowing between the node pairs. A contribution matrix calculation unit that calculates as
An estimation unit that estimates the traffic amount by allocating the traffic processing amounts corresponding to the nodes to the node pairs based on the contribution matrices corresponding to the nodes;
An information processing apparatus comprising: The estimation unit includes
A value proportional to each element value of the contribution matrix is used as each traffic processing amount allocated to each node pair;
The information processing apparatus according to claim 1, wherein: The estimation unit includes
In order from the node having the largest sum of the element values of the contribution matrix, the traffic processing amount corresponding to the node is allocated to each node pair.
The information processing apparatus according to claim 2, wherein: The estimation unit includes
In order from the node having the largest number of non-zero elements of the contribution matrix, the traffic processing amount corresponding to the node is allocated to each node pair.
The information processing apparatus according to claim 2, wherein: The estimation unit includes
A traffic processing amount that is a packet amount flowing through each node is calculated from the estimated traffic amount, and a magnitude relationship between the calculated traffic processing amount and the traffic processing amount calculated by the processing amount calculation unit is determined. When it is determined that the traffic processing amount calculated from the estimated traffic amount is larger, the traffic processing amount allocated to the node pair including the node is decreased, and the traffic calculated from the estimated traffic amount If it is determined that the processing amount is smaller, the traffic amount is allocated again by increasing the traffic processing amount allocated to the node pair including the node.
The information processing apparatus according to claim 2, wherein: Traffic that is the amount of packets that are connected to a network configured by connecting a plurality of nodes to each other by links, and that flow through the link between each pair of nodes obtained by arbitrarily selecting two nodes from the plurality of nodes. An information processing method by an information processing device for estimating a quantity,
Obtaining link information indicating a connection relationship between the nodes by the link;
Calculating the amount of packets flowing through each node within the same period as each traffic processing amount corresponding to the node;
Based on the link information and the traffic processing amount, each contribution matrix corresponding to the node includes a matrix having, as elements, the ratio of the packet amount flowing through the node to the packet amount flowing between the node pairs. Step to calculate as
Estimating the traffic volume by allocating each traffic processing amount corresponding to each node to each node pair based on each contribution matrix corresponding to each node;
An information processing method comprising:

Images