JP5300104B2 - Optical path network configuration equipment - Google Patents

Description

  The present invention relates to an optical path network configuration apparatus that configures an optical path network, which is a virtual network formed on a physical network, according to traffic.

  Conventionally, a network virtualization technique has been proposed as a technique for realizing effective use of network resources. In this network virtualization technology, a logical link that connects nodes in higher layers is provided by setting a wavelength path on a physical network based on wavelength division multiplexing (WDM). A simple topology (virtual network). Here, the wavelength of the physical network is a resource that becomes a bottleneck on the network and needs to be used effectively. Therefore, in order to efficiently accommodate the traffic in the wavelength path of the physical network, many studies on virtual network control that appropriately constructs a virtual network according to the traffic have been made (for example, see Non-Patent Documents 1 and 2).

  Non-Patent Documents 1 and 2 disclose a method for designing a virtual network optimally or heuristically for accommodating a given traffic demand matrix. Here, the traffic demand matrix represents the traffic demand between any two nodes in the network in a matrix format. This traffic demand matrix expresses the traffic exchange of the entire network.

  This conventional virtual network control technology assumes that traffic demand fluctuates periodically and gently. However, in a communication network in which a large number of applications and services are accommodated, the fluctuation of traffic on the virtual network is large and the fluctuation is difficult to predict. Therefore, in general, heuristic approaches that have adaptability to assumed environmental changes cannot cope with unexpected environmental changes.

  Therefore, in recent years, as a technique for solving this problem, a technique has been proposed in which attractor (stable point) selection that models behavior when a living organism adapts to unknown environmental changes is applied to control of a virtual network. (Refer nonpatent literature 3 and patent documents 1-3.).

The system operating by this attractor selection has a fluctuation structure η (fluctuation term) and a control structure f (control term) having an attractor as shown in the fluctuation equation (differential equation) of the following formula (1) that defines the fluctuation scheme. has two behaviors, their behavior is controlled by the activity of a value indicating the state of the system (growth rate) v _g.

Here, i, j, s, and d are numbers indicating nodes that are the arrival and departure points of the optical path (wavelength path), and p _ij and p _sd are node pairs. Further, x _pij and x _psd indicate the expression levels of the node pair p _ij and p _sd , respectively. The expression level is a value indicating a level that determines the number of optical paths constructed between nodes.

The activity _vg is an index that is a target for optimizing the link utilization rate. The activity v _g, when the state of the virtual network is deteriorated (for example, when the maximum link utilization is greater than a predetermined reference) decreased, increased when the state of the virtual network has been restored.
The function f is a sigmoid function. W in the sigmoid function f is a control matrix that represents the relationship between activity and suppression between optical paths having a node pair as a departure and arrival point.

The control matrix _{W (p ij,} p _sd) elements are node pairs _{p sd} is a positive constant when in the correlation of increasing the number of the optical path of the node pair _{p ij,} the node pairs _{p sd} is, the node pair _{p ij} A negative constant is set when there is a correlation that reduces the number of optical paths, and “0” is set when there is no correlation. That is, the control matrix W is a matrix of N rows and N columns when 1 ≦ i, j, s, d ≦ N (number of nodes).

Further, θ in the sigmoid function f is a constant that is dynamically set according to the link load.
Further, η is a fluctuation term, for example, a probability distribution such as a normal distribution, and white Gaussian noise or the like is used.

In the formula (1) described above, when the state of the virtual network is deteriorated, decreasing the value of the activity v _g, the influence of fluctuation η becomes relatively large. Then, a new attractor by the attractor control defined by the sigmoid function f is searched by the fluctuation η. Then, if the state of the virtual network has been restored, to increase the value of the activity v _g, the influence of fluctuations η becomes relatively small, the deterministic behavior, the system becomes stable.
Thus, the conventional system that controls the virtual network by attractor selection can construct a virtual network adapted to the environment in response to an unexpected environmental change.

JP 2011-155507 A JP 2011-155508 A JP 2011-155509 A

B. Mukherjee, D. Banerjee, S. Ramamurthy, and A. Mukherjee, “Someprinciples for designing a wide-area WDM optical network,” IEEE / ACM Transactions on Networking, vol. 4, no. 5, pp. 684-696 , 1996. R. Ramaswami and K. N. Sivarajan, “Design of logical topologies for wavelength-routed optical networks,” IEEE Journal on Selected Areas in Communications, vol. 14, pp. 840-851, June 1996. Y. Koizumi, et al., "Adaptive Virtual Network Topology Control Based on Attractor Selection," Journal of Lightwave Technology, Vol., 28, No. 11, June 1, 2010.

As described above, the method of controlling a virtual network by attractor selection is excellent in that a virtual network adapted to the environment can be constructed even when an unexpected environmental change occurs.
However, as shown in the equation (1), the control matrix W (p _ij , p _sd ) representing the presence / absence of the interaction between the activation and suppression of the node pair is expressed as 1 ≦ i, j , S, d ≦ N (number of nodes). That is, in the conventional method, the amount of calculation is an order of the fourth power of the number N of nodes in the matrix operation. When such a calculation amount is required, the network scale capable of reconfiguring the optical path of the virtual network according to the traffic in a realistic calculation time based on the current CPU capacity is up to 100 nodes. .
Under such circumstances, there has been a demand for a technique capable of controlling a virtual network by attractor selection even in a network scale of 1,000 to 10,000 nodes, for example, assuming a nationwide scale in Japan.

  Therefore, the present invention has been made based on such a demand. When an optical path is configured according to a change in the network environment, the amount of calculation is reduced, and even in a large-scale network, an optical path network is selected by attractor selection. It is an object of the present invention to provide an optical path network configuration device capable of configuring a (virtual network).

  The present invention was created to solve the above-mentioned problems. First, the optical path network configuration device according to claim 1 is a virtual network including optical paths between nodes on a physical network. In an optical path network configuration device configured to adapt to traffic using a fluctuation equation, a network load amount calculation means for calculating a network load amount which is a load applied to the network from a link load between nodes measured periodically, Based on the network load amount, an activity calculation means for calculating the activity used in the fluctuation equation, and a node pair to be controlled are identified from the relationship of activity or suppression with respect to the setting of the optical path between the node pairs, and the fluctuation equation Using the control matrix generation means for generating the control matrix to be used, the control matrix and the activity, the number of optical paths in the node pair to be controlled is determined by the fluctuation equation Expression level calculating means for calculating the expression level of the optical path, optical path number determining means for determining the number of optical paths to be assigned to the node pair to be controlled according to the expression level, and light determined by the optical path number determining means Virtual network setting means for setting the configuration of the optical path of the node pair to be controlled based on the number of paths, and the control matrix generation means has a positive or negative relationship of activation or suppression with respect to the setting of the optical path between the node pairs. The absolute value sum of the control values in relation to all the other node pairs is greater than or equal to a control target determination threshold value that is a preset threshold value for an arbitrary node pair on the physical network. Control matrix design means for designing the control matrix composed of the control values with a node pair as a control target, and the network load amount within a predetermined threshold range within a predetermined number of cycles When it is determined by the redesign necessity determination means that the necessity for redesign of the control matrix is determined by whether or not the control matrix needs to be redesigned, the redesign necessity determination means determines that the control matrix needs to be redesigned. When the network load amount does not fall below the preset upper limit threshold value, the control target determination threshold value is decreased, and when the network load amount does not exceed the preset lower limit threshold value, the control target determination threshold value is increased. And a threshold value changing unit.

  According to such a configuration, the optical path network configuration device reconfigures the virtual network by removing the node pair having a small activation or suppression relationship from the control target and reducing the size of the control matrix used for the fluctuation equation. This can reduce the amount of calculation.

  The optical path network configuration device according to claim 2 is the optical path network configuration device according to claim 1, wherein the control matrix design unit is a node pair to be non-controlled by the control target determination threshold. In addition, for a node pair in which the absolute value of the traffic measured periodically is equal to or greater than an absolute value threshold that is a preset threshold and the variance of the traffic is equal to or greater than a variance threshold that is a preset threshold, And when the threshold change unit determines that the redesign necessity determination unit needs to redesign the control matrix, when the network load amount is larger than a preset upper limit threshold value, The absolute value threshold and the dispersion threshold are decreased, and when the network load amount is smaller than a preset lower limit threshold, the absolute value threshold and the dispersion threshold are increased. The features.

  The optical path network configuration device according to claim 3 is the optical path network configuration device according to claim 1, wherein the control matrix design means is a node pair to be non-controlled by the control target determination threshold. In addition, a node pair whose absolute value of periodically measured traffic is equal to or greater than an absolute value threshold, which is a preset threshold, is set as the node pair to be controlled, and the threshold changing unit is configured by the redesign necessity determining unit. When it is determined that redesign of the control matrix is necessary, if the network load amount is larger than a preset upper limit threshold value, the absolute value threshold value is decreased, and the network load amount is set to a preset lower limit value. When it is smaller than the threshold value, the absolute value threshold value is increased.

  Further, the invention according to claim 4 is the optical path network constituent device according to claim 1, wherein the control matrix design means is periodic even if the control matrix design means is a node pair that is not controlled by the control object determination threshold. The node pair whose distribution of traffic measured in the above is a threshold value that is a threshold value set in advance is set as the node pair to be controlled, and the threshold changing unit redesigns the control matrix by the redesign necessity determination unit. When it is determined that it is necessary, when the network load amount is larger than a preset upper limit threshold, the dispersion threshold is decreased, and when the network load amount is smaller than a preset lower limit threshold, The dispersion threshold is increased.

  According to such a configuration, the optical path network constituent device can control the node pair that affects traffic to other node pairs, even if the node pair has a small activity or suppression relationship, by leaving it as a control target. In addition to reducing the amount of calculation by reducing the size of the matrix, adaptability to environmental changes can be improved.

  ADVANTAGE OF THE INVENTION According to this invention, the computational complexity at the time of comprising the optical path using the fluctuation system which controls a virtual network by attractor selection can be suppressed. As a result, the present invention can construct a virtual network adapted to the environment even when an unexpected environmental change occurs in a larger-scale network than before.

1 is a configuration diagram showing an overall configuration of an optical path network system according to an embodiment of the present invention. It is a block diagram which shows the function structure of the optical path network structure apparatus which concerns on 1st Embodiment of this invention. It is explanatory drawing for demonstrating the relationship between the activity which specifies the element of the control matrix of a fluctuation equation, and suppression. It is a flowchart which shows operation | movement of the optical path network structure apparatus which concerns on 1st, 2nd embodiment of this invention. It is a flowchart which shows the operation | movement of the control matrix design process of the optical path network configuration apparatus which concerns on 1st Embodiment of this invention. It is a flowchart which shows the operation | movement of the control matrix redesign process of the optical path network structure apparatus which concerns on 1st Embodiment of this invention. It is a block diagram which shows the function structure of the optical path network structure apparatus which concerns on 2nd Embodiment of this invention. It is a flowchart which shows the operation | movement of the control matrix design process of the optical path network configuration apparatus which concerns on 2nd Embodiment of this invention. It is a flowchart which shows the operation | movement of the control matrix redesign process of the optical path network structure apparatus which concerns on 2nd Embodiment of this invention.

Embodiments of the present invention will be described below with reference to the drawings.
[Configuration of optical path network system]
First, the configuration of an optical path network system including an optical path network configuration apparatus according to an embodiment of the present invention will be described with reference to FIG.
As shown in FIG. 1, an optical path network system 100 includes a virtual topology (VNT: Virtual Network Topology, hereinafter referred to as a virtual network) of an IP network that is an upper network on a WDM (wavelength division multiplexing) network that is a physical network. Network), and a virtual network is dynamically controlled according to traffic. Although an example in which a WDM network is adopted as a physical network will be described here, the present invention is similarly applied to a case where a layer 2 network such as a fiber network, a TDM (Time Division Multiplexing) network, or Ethernet (registered trademark) is a physical network. Is applicable. Here, an example of an IP-over-WDM network will be described as an example of a network.

  The optical path network configuration apparatus 1 configures a virtual network (optical path network) by attractor selection. In other words, the optical path network configuration apparatus 1 collects the traffic of the entire network and sets the fluctuation equation so as to select an attractor (stable point) in which the load (network load amount) applied to the network falls within a predetermined range. To control the virtual network. The configuration and operation of the optical path network forming apparatus 1 will be described later in detail.

(Network configuration)
Next, an IP-over-WDM network controlled by the optical path network configuration apparatus 1 according to the embodiment of the present invention will be described. As shown in FIG. 1, the IP-over-WDM network has two layers: a WDM network 101 and an upper IP network 102.

The WDM network 101 includes a plurality of optical cross connect (OXC) OXC ₁ , OXC ₂ , OXC ₃ , and OXC ₄ connected by optical fibers F ₁ , F ₂ , F ₃ , and F _4. Yes.
Optical cross connect OXC ₁ is configured with an IP router _{r 1} node _{N 1,} the optical cross connect OXC ₂ are configured with an IP router _{r 2} the node _{N 2.} Further, the optical cross connect OXC ₃ constitutes a node _{N 3} with IP routers _{r 3,} the optical cross connect OXC ₄ constitute the node _{N 4} with an IP router _{r 4.} In other words, in this example, the physical network includes four nodes (physical nodes) N _{1 to} N ₄ and four optical fibers (physical links) F _{1 to} F ₄ .

The optical fiber F ₁ connects between the optical cross-connects OXC ₁ and OXC ₂ , and the optical fiber F ₂ connects between the optical cross-connects OXC ₂ and OXC ₃ . The optical fiber F ₃ connects the optical cross-connects OXC ₃ and OXC ₄ , and the optical fiber F ₄ connects the optical cross-connects OXC ₄ and OXC ₁ . That is, in this example, a shape connected by _four optical fibers F _{1 to} F ₄ having _four nodes N _{1 to} N ₄ as vertices represents the topology of the physical network.
The physical link accommodates a plurality of wavelengths by WDM technology. Here, the wavelengths that can be used in the optical fibers F _{1 to} F ₄ are, for example, two wavelengths of λ ₁ and λ ₂ .

The physical nodes are connected by an optical path (wavelength path) by wavelength routing. Specifically, in the WDM network 101, an optical path is set using the wavelength λ ₁ for each of the four sides having the four nodes N _{1 to} N ₄ as vertices. When the optical path using the wavelength λ ₁ is set in this way, in the upper IP network 102, logical links V _{1 to} V ₄ are respectively connected to the four sides having _four nodes N _{1 to} N ₄ as vertices. Is set.

Further, the WDM network 101, by connecting the wavelength lambda ₂ at a node _{N 2,} to set the optical path with the wavelength lambda ₂ between the nodes _N 1, _{N 3.} When the optical path using the wavelength λ ₂ is set in this way, since the node N ₁ and the node N ₃ are logically adjacent to each other, the logical link V ₅ is set in the upper IP network 102. The In other words, in this example, the virtual network is composed of five logical links V _{1 to} V ₅ and four IP routers r _{1 to} r ₄ . A shape connected by logical links V _{1 to} V ₅ represents a virtual network topology (VNT).

  In the IP network 102, an optical path is created between IP routers passing through the OXC of the WDM network 101 by VNT control, and a VNT is determined from these IP routers and the optical path. That is, since a virtual network constructed on a physical network is specified by VNT, VNT is simply referred to as a virtual network below.

Although FIG. 1 shows four examples of optical cross-connects and IP routers, it is needless to say that this number can be set to an arbitrary number according to the network scale. Moreover, although the physical network accommodates one virtual network, the present invention is not limited to this, and the physical network can accommodate a plurality of virtual networks. The physical network provides a virtual network topology composed of optical paths to the virtual network. Here, the function of designing and controlling the virtual network topology is assumed by the optical path network configuration device 1.
Hereinafter, the configuration and operation of the optical path network configuration device according to the embodiment of the present invention will be described in detail.

<First Embodiment>
[Configuration of optical path network configuration equipment]
First, the configuration of the optical path network configuration apparatus 1 according to the first embodiment of the present invention will be described with reference to FIG.
The optical path network configuration device 1 includes network information collection means 11, storage means 12, virtual network design means 13, and virtual network setting means 14.

  The network information collecting unit 11 collects virtual network information such as the topology of the virtual network, the traffic amount on the link, the traffic distribution, the number of optical paths, the bandwidth of the optical path, and the topology of the physical network, the optical path (wavelength). Physical network information such as path information, number of transmitters, number of receivers, transmission capacity, and the like is collected from each node via a communication control unit (not shown). Note that such virtual network information may be acquired from a network control device (not shown) that centrally manages the network, instead of collecting it directly from the node.

The network information collecting unit 11 acquires these pieces of information (network information) at a predetermined cycle, and writes the acquired virtual network information in the storage unit 12. Further, the network information collecting unit 11 notifies the virtual network designing unit 13 of the fact at the timing when the network information is periodically collected.
If the state of the physical network does not change, the physical network information is written in the storage unit 12 in advance. Only when there is a change, the physical network information is input via the input unit (not shown). It should be updated.

The storage means 12 is a target when the network information (virtual network information and physical network information) collected by the network information collection means 11 or the virtual network design means 13 described later designs the optical path of the virtual network by attractor selection. The target optical path information indicating the node pair is stored.
The storage means 12 can be configured by a general storage medium such as a hard disk or a semiconductor memory. Here, the virtual network information, the physical network information, and the target optical path information are stored in one storage unit 12, but each may be stored in an individual storage device.

The virtual network designing unit 13 designs an optical path network (virtual network) by attractor selection based on network information (virtual network information and physical network information) stored in the storage unit 12. The virtual network design unit 13 periodically operates at the timing when the network information collection unit 11 collects network information.
Here, the virtual network design unit 13 includes a network load amount calculation unit 131, an activity level calculation unit 132, a control matrix generation unit 133, an expression level calculation unit 134, and an optical path number determination unit 135.

The network load amount calculation means 131 calculates a network load amount that is a load applied to the network from a link load between nodes of the network (IP network). For this network load amount, for example, the maximum utilization rate (maximum link utilization rate) among the link utilization rates between nodes can be used.
That is, the network load amount calculating unit 131, the traffic volume on a link between any pair of nodes when _{y pij,} the transmission capacity of the link between _{C pij,} utilization of each link is calculated by _y pij _{/ C pij} The maximum value is defined as the maximum link utilization rate u _max .

The traffic amount for each link is included in the virtual network information stored in the storage unit 12. Similarly, the transmission capacity of a link can be calculated by the product of the number of optical paths set for the link included in the virtual network information and the bandwidth of the optical path.
The network load amount calculating unit 131 outputs the calculated maximum link utilization rate u _max to the activity degree calculating unit 132 and the control matrix generating unit 133.

Here, an example using the maximum link utilization rate as the network load amount will be described. However, if the network load amount is an amount that can quantitatively specify the load on the entire network, the maximum link utilization rate is used. It is not limited. For example, an average of individual link utilization rates between nodes (average link utilization rate) may be used. That is, the network load amount calculating unit 131, the traffic volume on a link between any pair of nodes when _{y pij,} the transmission capacity of the link between _{C pij,} utilization of each link is calculated by _y pij _{/ C pij} A value averaged by the number of links may be used as the network load.

Based on the maximum link utilization rate calculated by the network load amount calculation unit 131, the activity level calculation unit 132 calculates the activity of the fluctuation equation used when the expression level calculation unit 134, which will be described later, calculates the expression level. It is.
Specifically, activity calculation means 132 calculates the activity v _g by the following equation (2).

Here, u _max is the maximum link utilization rate. Γ, δ, δ ′, and ζ are predetermined constants. As shown in equation (2), activity v _g is the maximum link utilization u _max is a large value equal to or greater than the constant zeta, is controlled to a smaller value is less than the constant zeta. Further, the inclination at that time is controlled by constants δ and δ ′, and the inclination becomes steeper as the constants δ and δ ′ are larger.

Thus, activity calculation means 132, if the constant ζ than the maximum link utilization, the performance of the virtual network regarded as bad, to reduce the activity of v _g. As a result, the influence of the fluctuation term of the fluctuation equation in the expression level calculation means 134, which will be described later, becomes relatively large and works in the direction in which a new attractor is searched.

Moreover, activity calculation means 132, the maximum link utilization is less than constant zeta, regarded as the performance of the virtual network is good, increasing the activity v _g. As a result, the influence of the fluctuation term of the fluctuation equation becomes relatively small, and the attractor converges by the control term of the fluctuation equation. That is, this activity serves as an index that is a target for optimizing the link utilization rate.
The activity calculation means 132 outputs the calculated activity _vg to the expression level calculation means 134.

  The control matrix generation unit 133 generates a control matrix of a fluctuation equation used when the expression level calculation unit 134 calculates the expression level. Here, the control matrix generation unit 133 includes a control matrix design unit 133a, a redesign necessity determination unit 133b, and a threshold change unit 133c.

The control matrix design unit 133a designs a control matrix of a fluctuation equation. The control matrix design unit 133a suppresses whether the relationship between the currently set node pairs is an active relationship based on the network information (virtual network information and physical network information) stored in the storage unit 12. And a control matrix is generated. At this time, the control matrix design unit 133a generates a control matrix using a node pair having a large influence, excluding a node pair having a small influence on the calculation of the control term of the fluctuation equation from a target for generating the control matrix.
Here, the active relationship means a relationship in which a certain node pair increases the number of optical paths of other node pairs. The suppression relationship is a relationship in which a certain node pair reduces the number of optical paths of other node pairs.

Here, with reference to FIG. 3, the relationship of activity and the relationship of suppression are typically demonstrated. As shown in FIG. 3 (a), node pair _N 1, _N optical paths _{L 2 2} is set, in a state where the optical path _{L 3} is set to the node pairs _N 2, _{N 3,} as motivation to set the optical path It is possible to divert traffic. In this case, the optical path L ₁ may be newly added to the node pair N ₁ and N ₃ . In this case, the node pair N ₁ , N ₂ (p _sd ) and the node pair N ₂ , N ₃ are in a relationship of activating the node pair N ₁ , N ₃ (p _ij ).

Further, as shown in FIG. 3B, nodes N ₁ , N ₃ , and N ₄ that are non-aggregating nodes exist with respect to the aggregation node N ₂ that is a traffic aggregation point, and the node pair N ₁ , N ₂ the optical path _{L 1} is set, the optical path _{L 2} to node pairs _N 2, _{N 3} are set. Here, if the number of optical paths in the node pair N ₁ , N ₂ is increased, it is necessary to increase the number of optical paths in the node pair N ₂ , N ₃ according to the traffic. In this case, the node pair N ₁ , N ₂ (p _sd ) has a relationship of activating the node pair N ₂ , N ₃ (p _ij ) (the node pair N ₂ , N ₄ is the same).

Further, as shown in FIG. 3C, the node pair N ₁ , N ₅ and the node pair N ₂ , N ₆ share the same optical fiber F ₁ between the node N ₃ and the node N ₄ through which the node pair N ₁ , N ₅ passes. I will do it. Here, when the number of optical paths of the node pairs N ₁ and N ₅ is increased, there is a relationship of reducing the number of optical paths of the node pairs N ₂ and N ₆ . The reverse is also true. In this case, the node pair N ₁ , N ₅ (p _sd ) has a relationship of suppressing the node pair N ₂ , N ₆ (p _ij ). In addition, the node pair N ₂ , N ₆ (p _ij ) has a relationship of suppressing the node pair N ₁ , N ₅ (p _sd ).
As described above, some node pairs have an activity or suppression relationship, but some have no correlation.
Returning to FIG. 2, the description of the configuration of the optical path network forming apparatus 1 will be continued.

The control matrix design unit 133a specifies the relationship of activation or suppression with respect to setting of the optical path between the node pairs with a positive or negative value (control value), and with respect to any node pair on the physical network, A node pair having a sum of absolute values of control values (control values) equal to or greater than a preset threshold value (control target determination threshold value) is set as a control target, and a control matrix composed of the control values is designed.
Specifically, first, the control matrix design unit 133a uses, as components, control values that specify the relationship between the activity and the suppression as described in FIG. 3 for every node pair stored in the storage unit 12. Then, a component of the control matrix W shown in the following formula (3) is generated (specified).

Here, i, j, s, and d are numbers indicating nodes that are 1 or more and the number of nodes N or less that are optical path arrival and departure points, and p _ij and p _sd indicate node pairs. Α _A represents a positive integer (eg, “1”), and α _I represents a negative integer (eg, “−1”).
As described above, the control matrix design unit 133a is a positive integer for a component corresponding to a node pair having an active relationship, a negative integer for a component corresponding to a node pair having a suppression relationship, and any relationship of activity and suppression. A component of the control matrix W is generated by setting “0” to the component corresponding to each other node pair.

In the conventional fluctuation equation shown in the above equation (1), when the absolute value sum of the components in the row direction of the control matrix W is smaller than a predetermined threshold value W ₀ , the expression level x of the node pair corresponding to that row Notwithstanding the _psd, the value of the sum of products of deleted the row direction of the component of the control matrix W and the expression levels x _psd is reduced. Therefore, even if the node pair specified by the row is excluded from the control target, the influence on the result of the matrix operation in the equation (1) is small.

Therefore, the control matrix design unit 133a, among the components of the control matrix W for all the node pairs, is a node pair in which the absolute sum of the row components is smaller than the threshold (control target determination threshold) W ₀ , that is, other absolute value sum of the control value in relation to all node pairs, the threshold W ₀ is smaller than the node pairs are excluded from the control target.

The control matrix design unit 133a is the initial startup is based for determining the control target predetermined initial value as the threshold value W _0. The control matrix design unit 133a, in the threshold value changing means 133c to be described later, if the threshold W ₀ is changed, as a reference for determining the control target of the changed threshold value.
The control matrix design unit 133a, the absolute value sum of the control values in the node pair, i.e., the relationship between all the other node pair that the sum of the absolute values of the line component corresponding to the threshold (control object determination threshold) W ₀ or more rows However, a node pair having a threshold value W ₀ or more is set as a control target.

Then, the control matrix design unit 133a generates a control matrix (reduction control matrix) W ′ consisting only of components of the control target node pairs among the components of the control matrix W specified by the equation (3).
Thus, the control matrix design unit 133a can generate a control matrix with reduced components for the control matrix for all node pairs.

  Then, the control matrix design unit 133a writes information specifying the node pair (for example, a pair of node IP addresses) to the storage unit 12 as the target optical path information for the node pair to be controlled. In addition, the control matrix design unit 133 a outputs the generated control matrix (reduction control matrix) W ′ to the expression level calculation unit 134.

The redesign necessity determination unit 133b redesigns the control matrix depending on whether or not the maximum link utilization rate (network load amount) calculated by the network load amount calculation unit 131 converges within a predetermined threshold. It is determined whether or not it is necessary.
The redesign necessity determination unit 133b determines that the control matrix needs to be redesigned when the maximum link utilization rate does not converge within a predetermined number of times when the virtual network design unit 13 periodically operates. To do. Only when it is determined that the control matrix needs to be redesigned, the redesign necessity determination unit 133b instructs the threshold change unit 133c to change the threshold used by the control matrix design unit 133a. . After the threshold is changed, the control matrix design unit 133a generates the control matrix again, so that the control matrix is redesigned.

The threshold changing unit 133c changes the threshold (control target determination threshold) used by the control matrix design unit 133a when the redesign necessity determination unit 133b determines that the control matrix needs to be redesigned. .
This threshold value changing means 133c is a component of a control matrix (deletion control matrix) W ′ when the maximum link utilization rate (network load quantity) calculated by the network load quantity calculating means 131 is not less than a predetermined _upper limit threshold value U _upper. Is small, that is, the number of node pairs to be controlled is small, and the threshold value W ₀ is decreased by a predetermined ratio X ₁ (positive number)%. For example, if the threshold value W ₀ is to be decreased by 5%, the threshold value changing unit 133c decreases the threshold value W ₀ from the original value to a value of 95% by W ₀ × (100−5) / 100.

Further, the threshold value changing unit 133c can obtain an optimal solution of the virtual network even if the number of node pairs is smaller when the maximum link utilization rate (network load amount) does not exceed the predetermined lower limit threshold value H _lower. And the threshold value W ₀ is increased by a predetermined ratio X ₁ (positive number)%. For example, if the threshold value W ₀ is to be increased by 5%, the threshold value changing unit 133c increases the threshold value W ₀ from the original value to a value of 105% by W ₀ × (100 + 5) / 100.

The threshold value changing means 133c outputs a threshold W ₀ changing to the control matrix design unit 133a, instructs the re-design of the control matrix.
Thus, the control matrix (deletion control matrix) W ′ generated by the control matrix design unit 133a works in a direction in which the number of node pairs to be controlled is optimized.

The expression level calculation unit 134 calculates an expression level indicating a level for determining the number of optical paths constructed between nodes by a fluctuation equation.
This expression level calculation means 134 uses the activity V _g calculated by the activity calculation means 132 and the control matrix (deletion control matrix) W ′ generated by the control matrix generation means 133, using the following formula ( The expression level for each node pair is calculated by the fluctuation equation (differential equation) shown in 4).

Here, i, j, s, and d are numbers indicating nodes that are optical path departure and arrival points, and p _ij and p _sd are node pairs. Further, x _pij and x _psd indicate the expression levels of the node pair p _ij and p _sd , respectively. The node pair is a control target node pair stored in the storage unit 12 as target optical path information.

The function f is a sigmoid function. Θ _pi is a constant that is dynamically set according to the link load. For example, theta _pij is the traffic taking the exponential moving average on the node pairs _{p ij.} Further, η is a fluctuation term, for example, white Gaussian noise which is a probability distribution such as a normal distribution.
By this expression (4), the expression level calculation means 134 calculates the expression level x _pij for each node pair p _ij .
The expression level calculation unit 134 outputs the calculated expression level for each node pair to the optical path number determination unit 135.

The optical path number determining means 135 determines the number of optical paths to be assigned to the node pair to be controlled according to the expression level calculated by the expression level calculating means 134.
The optical path number determination means 135 assigns a larger number as the expression level of the node pair is larger. Note that the number of optical paths is limited by the number of transmitters and receivers of the node.
Therefore, optical path number determination unit 135, for example, by the following equation (5), calculates the light path number _{G pij} node pair _{p ij.}

Here, P _R is the receiver number of nodes, is P _T is the transmitter number of nodes. The lower bracket means a truncation operation.
In other words, the optical path number determination unit 135 determines whether a certain node i has a node pair according to the ratio of the expression level x _{pij to} the sum (Σ _s x _psi ) of the expression level x _pij of the node pair p _ij that uses the transmitter of the node. The number of optical paths for p _ij is calculated. Further, the optical path number determining means 135 is configured such that, in a certain node i, the node pair is set in accordance with the ratio of the expression level x _{pij to} the sum (Σ _d x _pid ) of the expression level x _pij of the node pair p _ij that uses the receiver of the node. The number of optical paths for p _ij is calculated.

Then, the optical path number determination means 135 assigns the smaller one of the optical path number calculated based on the number of transmitters and the optical path number calculated based on the number of receivers to the node pair p _ij. Determine as a number.
The optical path number determining unit 135 outputs the optical path number determined for each node pair to the virtual network setting unit 14.

The virtual network setting unit 14 sets a network configuration in the optical path network system 100 (see FIG. 1) based on the number of optical paths for each node pair designed by the virtual network design unit 13.
That is, the virtual network setting unit 14 compares the number of optical paths of the node pair included in the virtual network information stored in the storage unit 12 with the number of optical paths for each node pair designed by the virtual network design unit 13. To do. If the numbers are different, the virtual network setting unit 14 adds an optical path and adds an unnecessary optical path to each node of the node pair so that the number of optical paths designed by the virtual network design unit 13 is obtained. Instruct to delete. If the traffic that has passed through the optical path is blocked by deleting the optical path, the route change is instructed before the optical path is deleted.

  Here, the virtual network setting unit 14 notifies each node of the increase / decrease of the optical path and the information on the path change (virtual network configuration information). When (not shown) exists, virtual network configuration information may be notified to the device.

When there is no network control device (not shown), the virtual network setting unit 14 initializes the optical path network as the initial state of the optical path network (virtual network) with the same settings as the topology of the physical network. We will make it. Then, the virtual network setting unit 14 writes the set network information in the storage unit 12.
By configuring the optical path network configuration device 1 as described above, when the optical path network configuration device 1 controls the optical path network (virtual network) by attractor selection, the components of the control matrix of the fluctuation equation are used. By reducing and reducing the amount of calculation, it becomes possible to cope with a large-scale network.

[Operation of optical path network configuration equipment]
Next, the operation of the optical path network configuration apparatus 1 according to the first embodiment of the present invention will be described with reference to FIGS. Here, it is assumed that the optical path network (virtual network) is set in an initial state, for example, a state configured with a topology of a physical network.

As shown in FIG. 4, first, the optical path network forming apparatus 1 collects network information (traffic amount, etc.) from each node by the network information collecting means 11, and measures the link load (step S1).
Then, the optical path network forming apparatus 1 performs initial setting (control matrix design processing) of the control matrix of the fluctuation equation by the control matrix design unit 133a of the control matrix generation unit 133 (step S2). At this time, the control matrix design unit 133a does not generate a control matrix for all the node pairs (optical paths), but uses a control matrix with a reduced number of targets. This control matrix design process will be described later with reference to FIG.

  Then, the optical path network constituent device 1 waits until the timing of the predetermined cycle (No in step S3), and repeatedly executes the following operations when the operation cycle is reached (Yes in step S3). In other words, the optical path network constituent device 1 collects network information (traffic amount, etc.) from each node by the network information collecting means 11 at the timing of the operation cycle, and measures the link load (step S4).

Then, the optical path network configuration apparatus 1 uses the network load amount calculation unit 131 of the virtual network design unit 13 to collect the traffic information on the link between nodes collected and measured as network information in step S4 and the transmission capacity of the link. Based on this, the maximum link usage rate that is the maximum among the usage rates for each link is calculated (step S5).
Further, the optical path network forming apparatus 1 calculates the activity of the fluctuation equation from the maximum link utilization rate calculated in step S5 by the activity calculating means 132 by the above equation (2) (step S6).

Then, the optical path network forming apparatus 1 determines whether or not the maximum link utilization rate calculated in step S5 is within a predetermined threshold range by the redesign necessity determination unit 133b of the control matrix generation unit 133. (Step S7).
Here, if the maximum link utilization rate is within a predetermined threshold range (Yes in step S7), the optical path network constituent device 1 advances the operation after step S10.

On the other hand, when the maximum link utilization rate is not within the predetermined threshold range (No in step S7), the optical path network configuration apparatus 1 further determines that the maximum link utilization rate is within the threshold range by the redesign necessity determination unit 133b. It is determined whether or not a state that is not within continues for a predetermined number of times with respect to the operation cycle (step S8).
Here, if the state has not been continued for a predetermined number of times (No in step S8), the optical path network forming apparatus 1 may remain in the current control matrix after step S10 because the state may converge. Advance the operation.

  On the other hand, if the state continues for a predetermined number of times (Yes in step S8), the operation proceeds to step S9 to redesign the control matrix (control matrix redesign process) (step S9), and the operation proceeds from step S10 onward. This control matrix redesign process will be described later with reference to FIG.

  Thereafter, the optical path network forming apparatus 1 uses the expression level calculation means 134 to use the control matrix designed in step S2 or the control matrix redesigned in step S9 and the activity calculated in step S6. Then, the expression level indicating the level for determining the number of optical paths constructed between the nodes is calculated for each individual node pair (optical path) by the equation (4) (step S10).

  Then, the optical path network constituent device 1 calculates the number of optical paths to be allocated between the node pairs by the optical path number determination means 135 according to the expression level for each node pair calculated in step S10 (step S11). . For example, the optical path number determining means 135 determines the number of optical paths according to the magnitude of the expression level within the limits of the number of transmitters and receivers of the node by the above equation (5).

Then, the optical path network configuration apparatus 1 uses the virtual network setting unit 14 to change the number of optical paths of the currently set node pair and the number of optical paths calculated by the operations up to step S11. An optical path is set to (step S12). That is, the virtual network setting unit 14 instructs the node pair in which the increase / decrease in the number of optical paths occurs to add or delete an optical path or change the route.
Thereafter, the optical path network forming apparatus 1 returns to step S3 and repeats the operations from step S4 to step S12 every operation cycle.

(Control matrix design process)
Next, the control matrix design process of the optical path network forming apparatus 1 will be described with reference to FIG. This process corresponds to the operation in step S2 in FIG.
First, the control matrix design unit 133a of the optical path network configuration device 1 specifies whether the relationship between the node pairs is active or suppressed for all currently set node pairs. Then, a control matrix is generated (step S20). That is, the control matrix design unit 133a does not have a positive integer for elements corresponding to node pairs having an active relationship, and a negative integer for elements corresponding to node pairs having a suppression relationship, neither active nor suppressed. A control matrix is generated by setting “0” to an element corresponding to a node pair.

Thereafter, the control matrix design unit 133a calculates the absolute value sum of the components in the row direction for each row in the control matrix generated in step S20 (step S21).
The control matrix design unit 133a compares the absolute value sum calculated in step S21, the threshold value _{W 0} that is set (step S22)

Here, when the absolute value sum of the line component is more than the threshold value _{W 0} (Yes in step S22), and the control matrix design unit 133a sets a node pair corresponding to the row to the controlled object (step S23). That is, a node pair whose control value indicating the relationship of activation or suppression in relation to all other node pairs is equal to or greater than the threshold value W ₀ is set as a control target.

On the other hand, when the absolute value sum of the line component is less than the threshold _{W 0} (No in step S22), and the control matrix design unit 133a sets a node pair corresponding to the row in uncontrolled (step S24). That is, a node pair whose control value indicating the relationship of activation or suppression in the relationship with all other node pairs is less than the threshold W ₀ is set as a non-control target.
Since the control target and the non-control target are exclusive, only one of the control target and the non-control target may be set in either step S23 or step S24.

  Further, the control matrix design unit 133a determines whether or not processing has been performed for all the rows of the control matrix generated in step S20 (step S25), and processing has not been completed for all the rows. (No in step S25), the process returns to step S21 to determine the next line.

Then, after the processing is completed for all the rows of the control matrix (Yes in step S25), the control matrix design unit 133a generates a control matrix for the node pair targeted for control in step S23 (step S23). S26). That is, the control matrix design unit 133a generates a new control matrix obtained by extracting the elements corresponding to the node pair to be controlled in step S23 from the control matrix generated in step S20.
Thereby, the control matrix design unit 133a generates a control matrix in which the target node pairs are limited from the control matrices for all the node pairs.

(Control matrix redesign process)
Next, referring to FIG. 6 (refer to FIG. 2 as appropriate for the configuration), the control matrix redesign process of the optical path network forming apparatus 1 will be described. This process corresponds to the operation in step S9 in FIG.

First, the threshold value changing unit 133c of the optical path network forming apparatus 1 determines whether or not the maximum link utilization rate calculated in step S5 in FIG. 4 exceeds a preset upper limit threshold value (step S30).
Here, only when the maximum link utilization exceeds a preset upper limit threshold (Yes in step S30), the threshold value changing means 133c is a threshold W ₀ used in step S22 in the control matrix design process of FIG. 5 Then, it is decreased by a predetermined ratio X ₁ % (step S31).

Further, the threshold value changing unit 133c of the optical path network constituting apparatus 1 determines whether or not the maximum link utilization rate is below a preset lower limit threshold value (step S32).
Here, only when the maximum link utilization rate is below the preset lower limit threshold value (Yes in step S32), the threshold value changing means 133c increases the threshold value W ₀ by a predetermined ratio X ₁ % (step S31). S33).

The control matrix design means 133a of the optical path network configuration device 1 uses the threshold _{W 0} which is increased or decreased in step S31 or step S33, redesigning the control matrix (step S34).
The operation in step S34 is the same as the control matrix design process (step S2) described in FIG.

  As described above with reference to FIG. 4 to FIG. 6, when designing the control matrix of the fluctuation equation, the optical path network forming apparatus 1 selects a node pair having less relation between activity and suppression in relation to other node pairs. By generating a control matrix excluding the control target by threshold judgment, the amount of calculation for calculating the expression level for each node pair by the fluctuation equation can be suppressed, and a larger optical path network can be controlled. become.

  In addition, since the optical path network configuration device 1 sequentially changes the threshold value for excluding the target node pair so that the maximum link utilization rate converges, even if the target node pair is small, the traffic is reduced. A stable optical path network can be configured.

Second Embodiment
[Configuration of optical path network configuration equipment]
Next, the configuration of the optical path network configuration apparatus 1A according to the second embodiment of the present invention will be described with reference to FIG.

The optical path network constituent device 1A constitutes a virtual network (optical path network) by selecting an attractor, and collects the traffic of the entire network as in the optical path network constituent device 1 described in FIG. The virtual network is controlled so as to select an attractor whose network load amount falls within a predetermined range.
As shown in FIG. 7, the optical path network configuration device 1A basically has the same configuration as the optical path network configuration device 1 described in FIG. Since the configuration other than the control matrix generation unit 133A is the same as that of the optical path network configuration device 1, the same reference numerals are given and description thereof is omitted.

  The control matrix generation unit 133A generates a control matrix of a fluctuation equation used when the expression level calculation unit 134 calculates the expression level. Here, the control matrix generation unit 133A includes a control matrix design unit 133Aa, a redesign necessity determination unit 133b, and a threshold value change unit 133Ac. The redesign necessity determination unit 133b is the same as that of the optical path network forming apparatus 1, and thus the description thereof is omitted.

  The control matrix design unit 133Aa designs a control matrix of a fluctuation equation. The control matrix design unit 133Aa basically has the same function as the control matrix design unit 133a, but further has a function of specifying a node pair as a target of the control matrix based on traffic.

Control matrix design means 133a of Figure 2, based on a threshold W ₀ which is set, for a node pair relationship with little other node pair and the active or suppressed were excluded from designing the control matrix of fluctuation equation. However, the control matrix design unit 133Aa is similar to the control matrix design unit 133a, in addition to reducing the node pair for which to design a control matrix using the threshold W _0, the traffic size and variation, the target Has a function to leave a node pair.

  For example, when the absolute value (bps) of the traffic volume accommodated by the optical path of the node pair is small, the corresponding optical path is deleted by the fluctuation equation, and the traffic accommodated in the optical path is detoured to another optical path. However, there is little possibility of causing congestion at the detour destination. Conversely, if the absolute value of the traffic volume accommodated by the optical path is large, the corresponding optical path is deleted by the fluctuation equation, and if the traffic accommodated in the optical path detours to another optical path, the detour destination There is a high possibility of causing congestion.

  In addition, for example, when the fluctuation (dispersion) of the traffic pattern (temporal transition of the number of packets) accommodated by the optical path is small, the optical path is stable, and it is necessary to add or delete the optical path using a fluctuation equation. Less is. Conversely, when there are many fluctuations in the traffic pattern accommodated by the optical path, the optical path is unstable, and it can be said that there is a high need to add or delete the optical path using a fluctuation equation.

Therefore, the control matrix design unit 133Aa includes the network information collected by the network information collection unit 11 when the absolute value T of the traffic of the node pair (optical path) exceeds a predetermined threshold T _th1 and When the variance ΔT exceeds a predetermined threshold T _th2 , the node pair is a target for designing a control matrix.

The threshold value changing unit 133Ac changes the threshold value used by the control matrix design unit 133Aa when the redesign necessity determination unit 133b determines that the redesign of the control matrix is necessary.
This threshold value changing means 133Ac has the function of changing the threshold value W ₀ used by the control matrix design means 133Aa based on the maximum link utilization rate (network load quantity) calculated by the network load quantity calculating means 131. This is the same as the matrix design means 133a.

This threshold value changing means 133Ac has a function of changing the threshold values T _th1 and T _th1 used by the control matrix designing means 133Aa based on the maximum link utilization rate (network load amount).
That is, the threshold value changing unit 133Ac determines that the component of the control matrix (deletion control matrix) W ′ is small, that is, the target node pair is small when the maximum link utilization rate does not fall below the predetermined _upper limit threshold value U _upper. The threshold _values T _th1 and T _th1 are decreased by predetermined ratios X ₂ (positive number)% and X ₃ (positive number)%, respectively.

Further, the threshold value changing unit 133Ac determines that the optimum solution of the virtual network can be obtained even with a smaller number of node pairs when the maximum link utilization rate does not exceed the predetermined lower limit threshold value H _lower. T _th1 and T _th1 are increased by predetermined ratios X ₂ (positive number)% and X ₃ (positive number)%, respectively.

  By configuring the optical path network configuration device 1A as described above, the optical path network configuration device 1A uses the attractor selection to control the components of the control matrix of the fluctuation equation when controlling the optical path network (virtual network). This can reduce the amount of calculation. Furthermore, the optical path network configuration device 1A can reduce the amount of calculation and efficiently configure the optical path network by leaving a node pair having a large influence on the addition and deletion of the optical path as a control target by the fluctuation equation. .

[Operation of optical path network configuration equipment]
Next, the operation of the optical path network forming apparatus 1A according to the second embodiment of the present invention will be described with reference to FIGS.

  The overall operation of the optical path network constituent device 1A is almost the same as that of the optical path network constituent device 1 described in FIG. 4, but the control matrix design process (step S2) and the control matrix redesign process (step S2) in FIG. Only S9) is different. Here, a control matrix design process (step S2A) different from the control matrix design process (step S2) will be described with reference to FIG. 8, and a control matrix redesign process (step S9) different from the control matrix redesign process (step S9) will be described. S9A) will be described with reference to FIG.

(Control matrix design process)
First, the control matrix design process (step S2A in FIG. 4) of the optical path network forming apparatus 1A will be described with reference to FIG.
This control matrix design process (step S2A) is the same as the control matrix design process (step S2) of the optical path network forming apparatus 1 described in FIG. 5 except that only step S22A is added. About the step, the same step number is added and description is abbreviate | omitted.

In step S22, the absolute value sum of the line component of less than the threshold value W ₀ (No in step S22), and the control matrix design unit 133Aa further in node pair corresponding to the row, the absolute value T of the traffic is the predetermined It is determined whether or not the threshold value T _th1 is exceeded and the traffic variance ΔT exceeds a predetermined threshold value T _th2 (step S22A).

Here, when the conditions based on the thresholds T _th1 and T _th2 are satisfied (Yes in step S22A), the control matrix design unit 133a proceeds to step S23, and sets the node pair corresponding to the row to be controlled (step S23). ). On the other hand, when the conditions based on the threshold values T _th1 and T _th2 are not satisfied (No in step S22A), the control matrix design unit 133a sets the node pair corresponding to the row as a non-control target (step S24).

As a result, even if the absolute value sum of the row components is less than the threshold value W _0, a node pair having a large traffic absolute value and a large variance becomes a control target.
After Steps S23 and S24, the operations after Step S25 are the same as those of the optical path network constituting apparatus 1.

(Control matrix redesign process)
Next, the control matrix redesign process (step S9A in FIG. 4) of the optical path network forming apparatus 1A will be described with reference to FIG.
Note that this control matrix redesign process (step S9A) adds steps S31A and S33A to the control matrix redesign process (step S9) of the optical path network forming apparatus 1 described in FIG. Since it is changed to step S34A, the same step number is added to the same step, and the description is omitted.

After step S31, the threshold value changing unit 133Ac decreases the threshold values T _th1 and T _th2 used in step S22A of the control matrix design process of FIG. 8 by predetermined ratios X ₂ % and X ₃ %, respectively (step S31A). ). Note that the order of step S31 and step S31A may be changed.

Further, after step S33, the threshold value changing unit 133Ac increases the threshold values T _th1 and T _th2 by predetermined ratios X ₂ % and X ₃ %, respectively (step S33A). Note that the order of step S33 and step S33A may be changed.

Then, after step S33A or when the maximum link utilization rate is greater than or equal to a preset lower limit threshold value in step S32 (No in step S32), the threshold value changing unit 133Ac is increased or decreased in step S31 or step S33. The control matrix is redesigned using the threshold value W ₀ and the threshold values T _th1 and T _th2 increased or decreased in step S31A or step S33A (step S34A).
The operation in step S34A is the same as the control matrix design process (step S2A) described in FIG.

  In the second embodiment, in the control matrix generation unit 133A, both the absolute value T of traffic and the traffic variance ΔT are used as a reference for reducing the node pairs for which the control matrix is designed. Only one of them may be used as a criterion for determination.

  As described above, the optical path network configuration apparatus 1A can reduce the amount of calculation and configure an optical path network by attractor selection, and can deal with a large-scale network. Furthermore, the optical path network configuration apparatus 1A can efficiently configure an optical path network even if the amount of calculation is reduced by leaving a node pair having a large influence on the addition and deletion of an optical path as a control target.

DESCRIPTION OF SYMBOLS 1 Optical path network configuration apparatus 11 Network information collection means 12 Storage means 13 Virtual network design means 131 Network load amount calculation means 132 Activity calculation means 133 Control matrix generation means 133a Control matrix design means 133b Reset necessity determination means 133c Threshold change Means 134 Expression level calculation means 135 Optical path number determination means 14 Virtual network setting means

Claims (4)

In an optical path network configuration device that configures a virtual network composed of optical paths between nodes on a physical network by adapting to traffic using fluctuation equations,
A network load amount calculating means for calculating a network load amount which is a load applied to the network from a link load between nodes measured periodically;
An activity calculation means for calculating an activity used in the fluctuation equation based on the network load;
A control matrix generating means for specifying a control target node pair from the relationship of activation or suppression with respect to the setting of an optical path between the node pairs, and generating a control matrix used in the fluctuation equation;
Expression level calculation means for calculating an expression level for determining the number of optical paths in the node pair to be controlled by the fluctuation equation using the control matrix and the activity,
According to the expression level, optical path number determining means for determining the number of optical paths to be allocated to the node pair to be controlled;
Virtual network setting means for setting the configuration of the optical path of the node pair to be controlled based on the number of optical paths determined by the optical path number determination means,
The control matrix generation means includes
The relationship of activation or suppression for the setting of the optical path between the node pairs is specified by a positive / negative control value, and for any node pair on the physical network, the absolute value sum of the control values in the relationship with all other node pairs Is a control matrix design means for designing the control matrix composed of the control values with a node pair equal to or higher than a control target determination threshold being a preset threshold as a control target;
Redesign necessity determination means for determining whether the control matrix needs to be redesigned according to whether the network load amount converges within a predetermined threshold range within a predetermined number of cycles;
When it is determined by the redesign necessity determination means that the control matrix needs to be redesigned, when the network load amount does not fall below a preset upper limit threshold, the control target determination threshold is decreased, When the network load amount does not exceed a preset lower limit threshold value, threshold value changing means for increasing the control target determination threshold value;
An optical path network constituting apparatus comprising: The control matrix design means includes:
Even for a node pair that is a non-control target by the control target determination threshold, the absolute value of the periodically measured traffic is equal to or greater than an absolute value threshold that is a preset threshold, and the variance of the traffic is preset. For a node pair that is equal to or greater than the dispersion threshold that is the threshold value, the node pair is the control target,
The threshold value changing means includes
When it is determined that the redesign necessity determination unit needs to redesign the control matrix, the absolute value threshold and the dispersion threshold are decreased when the network load amount is larger than a preset upper limit threshold. 2. The optical path network configuration device according to claim 1, wherein when the network load amount is smaller than a preset lower limit threshold, the absolute value threshold and the dispersion threshold are increased. The control matrix design means includes:
Even for a node pair that is not controlled by the control target determination threshold, a node pair whose absolute value of periodically measured traffic is greater than or equal to an absolute value threshold that is a preset threshold is set as the node pair to be controlled. ,
The threshold value changing means includes
When it is determined by the redesign necessity determination means that redesign of the control matrix is necessary, if the network load amount is larger than a preset upper limit threshold, the absolute value threshold is decreased, and the network 2. The optical path network constituting apparatus according to claim 1, wherein when the load amount is smaller than a preset lower limit threshold, the absolute value threshold is increased. The control matrix design means includes:
Even for a node pair that is a non-control target by the control target determination threshold, a node pair that is equal to or greater than a dispersion threshold that is a predetermined threshold for the distribution of periodically measured traffic is the node pair to be controlled,
The threshold value changing means includes
When it is determined by the redesign necessity determination means that the control matrix needs to be redesigned, if the network load amount is larger than a preset upper limit threshold, the dispersion threshold is decreased, and the network load is reduced. 2. The optical path network forming apparatus according to claim 1, wherein when the amount is smaller than a preset lower threshold, the dispersion threshold is increased.

Images