JP2014053770A - Scheduler, scheduling method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce scheduling time.SOLUTION: The present invention is applied to a scheduler 40 which constitutes a network system of a TDM system. The scheduler 40 of the present invention has: an AC traffic amount totalization section 401 which, from each node constituting the network system, periodically totals the traffic amount requested by each path which uses each node as a transmission source node; a traffic deviation detection section 403 which generates traffic deviation information indicating the deviation state of traffic when traffic deviation for each path in a network is detected on the basis of the traffic amount requested by each path; a traffic deviation information holding section 403 which holds the traffic deviation information; and a node grouper 404 which groups a plurality of nodes constituting the network system on the basis of the traffic deviation information.

Description

  The present invention relates to a technique for scheduling nodes and links in a TDM (time division multiplexing) network system.

  In a TDM system network system, processing is repeated with a period (t) that is an integral multiple of the TS (time slot) length as a cycle. Hereinafter, t is referred to as a TDM frame length.

  Hereinafter, a conventional scheduling method for scheduling nodes and links in a TDM network system will be described (for example, see Non-Patent Documents 1 and 2). Here, as shown in FIG. 31, it is assumed that the network system includes nodes A to E and links a to e and has a configuration of one wavelength and a unidirectional ring. Further, it is assumed that the scheduler is arranged in the node A.

Step S100:
First, referring to FIG. 31, the scheduler converts the amount of requested traffic requested by each path into the number of requested TS, and converts the traffic matrix into a TS matrix. A path refers to a communication path that connects a source node and a destination node. Here, the traffic amount of 50 Mbps is converted to 1 TS. For example, since the traffic volume of the path from node A to node D is 150 Mbps, the number of requested TS is three.

Step S200:
Next, referring to FIG. 32, the scheduler calculates the number of requested TS for each link a to e based on the TS matrix. For example, the number of requested TSs for link a is 18, which is the sum of the number of requested TSs for the path passing through link a (the number of requested TSs surrounded by rounded squares). Next, a TDM frame length that can accommodate all TSs (referred to as superframe length in Non-Patent Document 2) is obtained based on the maximum number of requested TSs for links a to e. Since the number of requested TSs for the links a and b is 18, which is the maximum value, the TDM frame length is 18. In Non-Patent Document 2, there is a restriction that t is an integral multiple of the basic frame length (defined as an integral multiple of the TS length). Therefore, when the basic frame length is 10 TS, all TS are accommodated. The obtained TDM frame length (referred to as superframe length in Non-Patent Document 2) is 20 TS for two frames. Here, the frame length that can accommodate all TSs is variable, but may be fixed.

Step S300:
After that, referring to FIG. 33, the scheduler allocates each path for the number of requested TSs to the empty TS having the frame length obtained in step S200. At this time, for example, there are various allocation policies such as First Fit allocation (immediate allocation when a free space is found) and continuous TS priority allocation (immediate allocation if the requested number of TSs can be secured continuously). In this step, a link schedule table that indicates which path data is to flow in each TS of each link a to e is created.

  The flow of the above scheduling method is shown in the flowchart of FIG. That is, the scheduler performs the processing of the above-described steps S100 to S300 after the traffic volume requested by each path is aggregated, and for each node A to E, the link schedule table of the link having the node as the end point and the link having the node as the start point. To be notified.

X. Zhang and C. Qiao, "Pipelined transmission scheduling in all-optical TDM / WDM rings," in Proc. Int. Conf. Computer Communication and Networks, Sept. 1997, pp. 144-149. K. Gokyu, K. Baba, and M. Murata, "Path accommodation methods for unidirectional rings with optical compression TDM," IEICE Transactions on Communications, vol. E83-B, pp. 2294-2303, Oct. 2000. T. Tatsuta, N. Oota, N. Miki, and K. Kumozaki, "Design philosophy and performance of a GE-PON system for mass deployment," JOURNAL OF OPTICAL NETWORKING, vol. 6, no. 6, pp. 689- 700, June 2007.

  By the way, in the network system, the traffic amount of each path is not fixed but fluctuates at any time.

  For example, referring to FIG. 35, when TS allocation is performed in a state where the traffic volume of the path from the node C to the node A is large and the traffic volume of the path from the node C to the node B is small, then the node C → the node A The amount of traffic on the path of node C may decrease, and the amount of traffic on the path from node C to node B may increase.

  At this time, if the TS is not changed, a large number of TSs remain allocated to the path from the node C to the node A where the traffic volume has decreased, and the state becomes very inefficient. In addition, a small TS remains allocated to the path from node C to node B where the traffic volume has increased, resulting in a shortage of TS.

  Therefore, the scheduler performs the scheduling again and tracks the traffic fluctuation by rewriting the link schedule table.

However, since the total number of paths is represented as O (N ² ) where N is the number of nodes, the average TS allocation calculation time per path is used when performing TS allocation sequentially for each path as in the conventional method. Is set to α, the calculation time of TS allocation for all paths is α × O (N ² ). In addition, since the number of links through which each path passes, that is, the number of links that need to check the status of the blockage at the time of TS allocation is expressed as O (N), α itself also depends on N. Therefore, the calculation time for TS allocation increases with the scale of the network system.

  Therefore, in order to realize a large-scale network system that follows traffic fluctuations, it is important to reduce the scheduling time by reducing the TS allocation calculation time.

  Therefore, an object of the present invention is to provide a technique capable of realizing a large-scale system by reducing scheduling time in a TDM network system.

The scheduler of the present invention
A scheduler constituting a TDM network system,
Aggregating means for periodically totaling the amount of traffic required by each path having the node as a transmission source node from each node constituting the network system;
Based on the amount of traffic requested by each path, when detecting the traffic bias of each path in the network, traffic bias detection means for creating traffic bias information indicating the traffic bias state;
Traffic bias information holding means for holding the traffic bias information;
Grouping means for grouping a plurality of nodes constituting the network system based on the traffic bias information.

The scheduling method of the present invention includes:
A scheduling method by a scheduler constituting a TDM network system,
Periodically totaling the amount of traffic required by each path having the node as a transmission source node from each node constituting the network system;
When detecting traffic deviation of each path in the network based on the traffic amount requested by each path, creating and maintaining traffic deviation information indicating the traffic deviation state; and
And grouping a plurality of nodes constituting the network system based on the traffic bias information.

The program of the present invention
In the scheduler that configures the TDM network system,
A procedure for periodically totaling the amount of traffic required by each path having the node as a transmission source node from each node constituting the network system;
A procedure for creating and maintaining traffic bias information indicating a traffic bias state when detecting a traffic bias of each path in the network based on the traffic volume requested by each path;
And a procedure for grouping a plurality of nodes constituting the network system based on the traffic bias information.

  According to the present invention, grouping of a plurality of nodes constituting a network system is performed. As a result, after TS allocation between groups, TS allocation of paths between nodes in the group is executed step by step, or after TS allocation within groups, step by step between nodes between groups Since it is possible to execute TS allocation of paths, it is possible to reduce the scheduling time and achieve the effect of realizing a large-scale system.

  In addition, according to the present invention, since the grouping is performed in consideration of the traffic bias, an effect of suppressing an increase in the schedule length (time required to transfer the requested traffic) can be obtained.

It is a figure which shows the structure of the network system of this invention. It is a figure which shows the structure of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the outline of the whole operation | movement of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the conventional grouping method. It is a figure explaining the grouping method of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the precondition of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the detail of the whole operation | movement of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the detail of the whole operation | movement of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the detail of the whole operation | movement of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the detail of the whole operation | movement of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the operation | movement sequence of the whole network system of this invention. It is a figure explaining the operation | movement sequence of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the operation | movement sequence of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the operation | movement sequence of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the operation | movement sequence of the scheduler of the 1st Embodiment of this invention. It is a figure explaining the operation | movement sequence of the scheduler of the 1st Embodiment of this invention. It is a figure which shows the structure of the scheduler of the 2nd Embodiment of this invention. It is a figure explaining the detail of the whole operation | movement of the scheduler of the 2nd Embodiment of this invention. It is a figure explaining the operation | movement sequence of the scheduler of the 2nd Embodiment of this invention. It is a figure which shows the structure of the scheduler of the 3rd Embodiment of this invention. It is a figure explaining the detail of the whole operation | movement of the scheduler of the 3rd Embodiment of this invention. It is a figure explaining the operation | movement sequence of the scheduler of the 3rd Embodiment of this invention. It is a figure which shows the structure of the scheduler of the 4th Embodiment of this invention. It is a figure explaining the grouping method of the scheduler of the 4th Embodiment of this invention. It is a figure explaining the grouping method of the scheduler of the 4th Embodiment of this invention. It is a figure explaining the effect of the grouping method of the scheduler of the 4th Embodiment of this invention. It is a figure explaining the specific example of the grouping method of the scheduler of the 4th Embodiment of this invention. It is a figure explaining the operation | movement sequence of the scheduler of the 4th Embodiment of this invention. It is a figure which shows the structure of the scheduler of the 5th Embodiment of this invention. It is a figure which shows the structure of the scheduler of the 6th Embodiment of this invention. It is a figure explaining the conventional scheduling method. It is a figure explaining the conventional scheduling method. It is a figure explaining the conventional scheduling method. It is a flowchart explaining the flow of the conventional scheduling method. It is a figure explaining the subject of the conventional scheduling method.

EMBODIMENT OF THE INVENTION Below, the form for implementing this invention is demonstrated with reference to drawings.
(1) First Embodiment FIG. 1 shows the configuration of a network system to which the scheduler of the present invention is applied.

  Referring to FIG. 1, the network system of the present invention is a ring network system in which N (N is a natural number of 2 or more) nodes 10 are connected by a link 20. In FIG. 1, N = 5, but the number of N is not limited to this.

  Each node 10 is connected to a host computer 30 and transfers data signals between the host computers 30. At this time, the node 10 processes the data signal sent from the previous node on the ring, and sends the data signal to the next node.

  The scheduler 40 manages the entire network system and schedules each link 20 and each node 10. The placement location of the scheduler 40 is not limited to that shown in FIG. 1 and may be placed in another node.

  In the present invention, the scheduler 40 has a main feature, and the node 10 may have a configuration in which a function of dynamically collecting a traffic amount of a path and notifying the scheduler 40 is added to a known configuration. The host computer 30 may have a known configuration. Therefore, only the configuration of the scheduler 40 will be described in detail below.

  FIG. 2 shows the configuration of the scheduler 40 of the first embodiment.

  Referring to FIG. 2, the scheduler 40 according to the first embodiment includes an AC traffic amount totaling unit 401, a traffic bias detection unit 402, a traffic bias information holding unit 403, a node grouper 404, a selector 405, and a group configuration table. 406, an inter-group bandwidth allocation unit 407, a link table (group unit) 408, an intra-group (inter-node) bandwidth allocation unit 409, a table conversion unit 410, a link schedule table 411, a table conversion unit 412, A node group table 413, a table conversion unit 414, a node schedule table 415, a table transmitter 416, and a timer 417 are provided.

  The AC traffic volume totalization unit 401 periodically totals the traffic volume of requested traffic requested from each node 10 by each path having the node as a transmission source node.

  The traffic bias detection unit 402 detects the traffic bias of the path in the network based on the traffic amount requested by each path, and creates traffic bias information indicating the traffic bias state when the traffic bias is detected. To do. It should be noted that the traffic bias information indicates information such as how much / in which communication the node 10 is involved in.

  The traffic bias information holding unit 403 holds traffic bias information.

  The node grouper 404 groups (groups) the N nodes 10 logically into G (G is a natural number of 2 or more satisfying the relationship of N> G) groups based on the traffic bias information. Means. At this time, the node grouper 404 performs one or more groupings to create one or more group configuration candidates.

  The selector 405 selects one of the group configuration candidates.

  The group configuration table 406 is a table representing the nodes 10 belonging to the group for each group configuration selected by the selector 405.

  The inter-group bandwidth allocation unit 407 allocates TS between each group (including both the same group and between different groups), and based on the TS allocation result, the link table (group unit) 408 of each link 20 is stored. create. That is, the inter-group band allocation unit 407 performs scheduling between groups.

  The link table (group unit) 408 is arranged as many as the number of links (N), and is a table showing which group data flows in each TS of the link 20.

The intra-group (inter-node) bandwidth allocation unit 409 is arranged for the square of the number of groups (G ² ), and is located between the nodes 10 in the group (inter-nodes 10 belonging to the same group, belonging to different groups. TS is assigned to a path including both nodes 10). The intra-group (inter-node) bandwidth allocation unit 409 can perform parallel processing. That is, the intra-group (inter-node) bandwidth allocation unit 409 performs intra-group scheduling.

Note that the breakdown of the G ² intra-group (inter-node) bandwidth allocation unit 409 is to allocate TSs to paths between nodes 10 belonging to the same group (source and destination nodes belong to the same group). G (G-1) is assigned to the paths between nodes 10 belonging to different groups (paths belonging to different groups in which the transmission source node and the destination node are different), a total of G ² It becomes.

  The table conversion unit 410 is arranged for the number of links (N), and the link table (group unit) 408 of the link 20 is determined based on the TS allocation result of the intra-group (inter-node) bandwidth allocation unit 409. , Converted into a link schedule table 411. Each table conversion unit 410 can perform parallel processing.

  The link schedule table 411 is arranged for the number of links (N), and is a table representing the schedule of the link 20 (specifically, which path data is to flow in each TS).

  The table conversion unit 412 is arranged for the number of groups (G), and creates a node group table 413 for the group based on the link table (group unit) 408 of each link 20. Each table conversion unit 412 can perform parallel processing.

  The node group table 413 is arranged for the number of groups (G), and is a table representing the contents of node processing (data transmission, route switching, etc.) in each TS of the group.

  The table conversion unit 414 is arranged for the number of groups (G), and converts the node group table 413 of the group into the node schedule table 415 based on the link schedule table 411 of each link 20. Each table conversion unit 414 can perform parallel processing.

  The node schedule table 415 is arranged for the number of nodes (N), and represents the schedule of the node 10 (specifically, the contents of node processing (data transmission, route switching, etc.) in each TS). It is a table.

  The table transmitter 416 transmits the node schedule table 415 of the node to each node 10. Here, in the conventional method, each node 10 has to uniquely recognize the node processing in each TS based on the link schedule table 411. However, in the present invention, since the node schedule table 415 is transmitted to each node 10, it is not necessary to provide a function for each node 10 to recognize node processing from the link schedule table 411.

  The timer 417 manages the time during which each node 10 performs processing.

  It is assumed that the group configuration table 406, the link table (group unit) 408, the link schedule table 411, the node group table 413, and the node schedule table 415 are held in a memory (not shown) inside the scheduler 40.

Hereinafter, the operation of the scheduler 40 of the first embodiment will be described.
(I) General Operation First, the general operation of the scheduler 40 will be described.

  FIG. 3 shows an outline of the overall operation of the scheduler 40 of the first embodiment.

  Referring to FIG. 3, here, for convenience of explanation, the network system has a unidirectional ring configuration with N = 6, and one channel that can be simultaneously connected to the same slot per link (fiber multiplexing or wavelength multiplexing). Not). In addition, the six nodes 10 are nodes A to F, respectively, and the six links 20 are links a to f, respectively. However, the present invention is not limited to this, and can also be applied to a case where an arbitrary topology and wavelength multiplexing are performed including a bidirectional ring (a link schedule table 411 is created for each wavelength of each fiber).

  Further, it is assumed that the scheduler 40 is provided in the node A.

  First, the node grouper 404 logically groups the six nodes A to F into three groups, creates one or more group configuration candidates, and the selector 405 selects one of them. Here, nodes A and B are selected as group G1, nodes C and D as group G2, and nodes E and F as group G3, respectively.

  Next, the inter-group bandwidth allocation unit 407 converts the node-unit TS matrix into a group-unit TS matrix. For example, the number of requested TSs of G3 → G3 (a set of paths in which the source node belongs to G3 and the destination node belongs to G3. The same applies hereinafter) is that one of nodes E and F belonging to G3 is the source node and the other is Based on the number of request TSs for the path as the destination node (the number of request TSs enclosed in a rounded rectangle), the number is 4.

  Next, the inter-group bandwidth allocation unit 407 allocates TSs between the groups based on the TS matrix in units of groups. For example, the TSs S1 to S4 of the links a to e are allocated to G1 → G3. Note that the middle table on the right side of FIG. 3 represents the status of TS vacancy defined on each link in the network, and is hereinafter referred to as a schedule table.

  Next, the inter-group bandwidth allocation unit 407 creates a link table (group unit) 408 for each link. For example, in the link table (group unit) 408 of link a, G1 → G3 is scheduled for the TS of S1.

  On the other hand, the intra-group (inter-node) bandwidth allocation unit 409 allocates TS to each path between the nodes in the group. For example, the intra-group (inter-node) bandwidth allocation unit 409 that performs TS allocation of paths between nodes in G1 → G3 allocates S1 TSs of links a to e to the path of node A → node F. At this time, the TS allocation process for the path between the nodes in G1 → G3 and the TS allocation process for the path between the nodes in G1 → G2 can be performed independently.

  Thereafter, each table conversion unit 410 converts the link table (group unit) 408 of each link into a link schedule table 411. For example, in the link schedule table 411 for link a, the path from node A to node F is scheduled for the TS of S1.

  In the present invention, N nodes 10 are grouped. Therefore, TS allocation of paths between nodes in the group can be executed in stages after TS allocation between groups, so that the calculation time of TS allocation can be reduced compared with the conventional method. it can.

That is, in the conventional method, O (N ² ) paths are allocated on O (N) links. On the other hand, in the present invention, in the group unit calculation, O (G ² ) paths are allocated on the O (G) links, and in the node unit calculation in the group, per group, Assign O {(N / G) ² } paths on O {(N / G)} links. Further, in the present invention, since the table is divided, the number of TSs that need to check the status of the empty block at the time of TS allocation is smaller than that in the conventional method. As described above, in the present invention, the calculation time for TS allocation can be reduced as compared with the conventional method.

  Further, for example, on the table of FIG. 3, the portion of G1 → G3 and the portion of G1 → G2 are exclusively divided. Therefore, the TS allocation process for the path between the nodes in G1 → G3 and the TS allocation process for the path between the nodes in G1 → G2 can be executed independently. Therefore, the TS allocation calculation time by the intra-group (inter-node) band allocation unit 409 is processed in parallel, so that the TS allocation calculation time can be further reduced.

  As described above, in the present invention, the TS allocation calculation time can be reduced by grouping N nodes 10.

  However, depending on the grouping method, the schedule length (the time required to transfer the requested traffic) increases due to the uneven traffic of each path in the network. There may be a problem that request traffic cannot be transferred. Hereinafter, this problem will be described in detail.

  FIG. 4 shows an example of a grouping method in which the same number of nodes are fixedly grouped.

  Referring to FIG. 4, here, for convenience of explanation, the network system has a unidirectional ring configuration with N = 8, and one channel that can be simultaneously connected to the same slot per link (fiber multiplexing or wavelength multiplexing). Not). Eight nodes 10 are respectively designated as nodes A to H (the same applies to FIG. 5 described later). However, the present invention is not limited to this, and can also be applied to a case where an arbitrary topology and wavelength multiplexing are performed including a bidirectional ring (a link schedule table 411 is created for each wavelength of each fiber).

  Here, it is assumed that the node grouper 404 groups the eight nodes A to F into four groups, two at a fixed number from the lowest node number, and the selector 405 selects them. That is, nodes A and B are grouped in group G1, nodes C and D are grouped in group G2, nodes E and F are grouped in group G3, and nodes G and H are grouped in group G4.

  At this time, in G1 → G1, the number of requested TSs for the path from node A → node B is 5, the number of requested TSs for the path from node B → node A is 1, and the traffic volume of the path from node A → node B is In particular, there is a large amount, and there is a large bias in the traffic between the two. For this reason, the empty area increases on the schedule table, and TS allocation becomes inefficient.

  Further, the number of requested TSs from G1 to G2 is 11, and the number of requested TSs from G3 to G4 is 4, and there is a large deviation in the traffic between the two. For this reason, the empty area increases on the schedule table, and TS allocation becomes inefficient.

  Here, if a set of consecutive TSs on the schedule table is referred to as a block for the sake of convenience, when blocks having different heights (here, corresponding to the number of requested TSs) are packed on the schedule table, an empty space is generated. It becomes easy.

  As a result, TS allocation becomes inefficient, the schedule length increases, and all the requested traffic cannot be transferred within the time t corresponding to the TDM frame length.

  Therefore, in the first embodiment, the node grouper 404 groups the N nodes 10 in consideration of traffic bias based on the traffic bias information in order to suppress an increase in the schedule length.

  FIG. 5 shows an example of a grouping method according to the traffic distribution.

  Referring to FIG. 5, the node grouper 404 makes the traffic amount within the group as uniform as possible among the groups based on the traffic bias information (ie, the height of each block is as equal as possible). ) Grouping is performed.

  For example, in the example of FIG. 5, the number of requested TSs for the path from node A → node C is five, and the number of requested TSs for the path from node E → node G is one, which is small.

  Therefore, the node grouper 404, for example, sets the node A to the group G1 and the nodes B and C to the group G2 so that the number of nodes in the group to which the node A belongs is small and the number of nodes in the group to which the node E belongs is large. , Nodes D, E, and F are grouped into group G3, and nodes G and H are grouped into group G3.

  As a result, the number of required TSs for G1 → G2 and G3 → G4 are both 8, and the heights of these blocks are equal.

  As described above, when blocks having the same height (here, corresponding to the number of requested TSs) are packed on the schedule table, it becomes difficult to generate empty space.

As a result, the efficiency of TS allocation is improved, an increase in schedule length is suppressed, and all requested traffic can be transferred within a time t corresponding to the TDM frame length.
(Ii) Detailed Operation Next, the detailed operation of the scheduler 40 will be described.

  FIG. 6 shows preconditions for the detailed operation of the scheduler 40.

  Referring to FIG. 6, here, for convenience of explanation, the network system has a unidirectional ring configuration with N = 8, and one channel that can be simultaneously connected to the same slot per link (fiber multiplexing or wavelength multiplexing). Not). It is assumed that the eight nodes 10 are nodes A to H, respectively. However, the present invention is not limited to this, and can also be applied to a case where an arbitrary topology and wavelength multiplexing are performed including a bidirectional ring (a link schedule table 411 is created for each wavelength of each fiber).

  Further, it is assumed that the scheduler 40 is provided in the node A.

  FIG. 7 shows the operation until the traffic bias information is held among the detailed operations of the entire scheduler 40.

  Referring to FIG. 7, first, in step S <b> 1, the AC traffic volume totaling unit 401 periodically determines the traffic volume of the requested traffic requested by each path having the node as a transmission source node from each of the nodes A to H. Totaling is performed, the traffic volume is converted into the required number of TSs, and the traffic matrix is converted into a TS matrix. The TS matrix represents the number of requested TSs required by each path between the nodes, and is a two-dimensional array having elements of the number of nodes × the number of nodes. Here, the traffic amount of 10 Mbps is converted to 1 TS. For example, since the traffic volume of the path from node A to node B is 100 Mbps, the number of requested TS is 10. The TS matrix is output to the traffic deviation detection unit 402, the inter-group band allocation unit 407, and the intra-group (inter-node) band allocation unit 409, and is held there.

  Next, in step S2, the traffic bias detection unit 402 creates traffic bias information indicating the traffic bias state when the traffic bias of each path in the network is detected based on the TS matrix. Here, the traffic volume of the path having node A as the transmission source node is large, and the traffic volume of the path having nodes D, E, and F as the transmission source node and nodes A, B, and C as the destination nodes is small. Yes.

  Next, in step S3, the traffic bias information holding unit 403 holds the traffic bias information. The traffic bias information is a one-dimensional array indicating the traffic bias state of each path, that is, how much / smaller the communication is related to which node.

  Here, the operation for detecting the traffic bias in step S2 shown in FIG. 7 will be described in detail.

  FIG. 8 and FIG. 9 show details of the operation for detecting the traffic bias in step S2.

Referring to FIG. 8, the traffic deviation detection unit 402 includes four counters (not shown) for each node. First, the TS matrix values are input, and the traffic is detected using these four counters. Analyze the next four items on the distribution.
Dst: the number of destination nodes with which the relevant node communicates Tx: the sum of the number of TSs that the relevant node transmits (that is, the sum of the traffic volume of the path where the relevant node is the source node)
Src: the number of transmission source nodes with which the relevant node communicates Rx: the sum of the number of TSs received by the relevant node (that is, the sum of the traffic volume of the path where the relevant node is the destination node)
At this time, the counter corresponding to each item searches the TS matrix and performs the following calculation operation.
Dst (n): Calculates the number of non-zero elements in the nth row (n = 1,..., N. The same applies hereinafter)
-Tx (n): Add the element in the nth row-src (n): Calculate the number of non-zero elements in the nth column-Rx (n): Add the element in the nth column See also FIG. Then, the traffic deviation detection unit 402 includes not only four counters (not shown) for each node but also a comparator (not shown), and the counter of each counter is used by using the comparator. By performing various comparisons on the values, whether there is a path bias or limitation for each source node, whether there is a path bias or limitation for each destination node, or the vertical ratio of each node (that node is the source The ratio between the number of TSs of the path serving as a node and the number of TSs of the path serving as the destination node is detected.
Detection of a node whose destination is limited First, one of N nodes n is selected.

  Next, dst (n) of the node n selected above is compared with the value of (N−1).

  Here, if the two match, it is detected that there is no limitation on the destination node in the path having the node n as the transmission source node.

  On the other hand, if the two do not match, it is detected that there is a limitation of destination nodes in the path having node n as the transmission source node.

The above is performed for all N nodes n.
-Detection of a node whose transmission source is limited First, one of N nodes n is selected.

  Next, src (n) of the node n selected above is compared with the value of (N−1).

  Here, if the two match, it is detected that there is no limitation on the source node in the path having node n as the destination node.

  On the other hand, if the two do not match, it is detected that there is a limitation on the source node in the path having node n as the destination node.

The above is performed for all N nodes n.
Detection of traffic bias for each transmission source First, an average value of Tx (1),..., Tx (N) is obtained.

  Next, one of the N nodes n is selected, and Tx (n) of the selected node n is compared with the average value.

  Here, if both are of the same level, it is detected that there is no bias in the traffic of the path having the node n as the transmission source node.

  On the other hand, if Tx (n) is greater than or equal to 平均 times the average value or less than or equal to ●, it is detected that there is a bias in traffic on the path having node n as the transmission source node.

The above is performed for all N nodes n.
Detection of traffic bias for each destination First, an average value of Rx (1),..., Rx (N) is obtained.

  Next, one of the N nodes n is selected, and Rx (n) of the selected node n is compared with the average value.

  Here, if both are of the same level, it is detected that there is no bias in the traffic of the path having the node n as the destination node.

  On the other hand, if Rx (n) is greater than or equal to 平均 times the average value or less than or equal to ●, it is detected that there is a bias in traffic on the path having node n as the destination node.

The above is performed for all N nodes n.
Detection of vertical asymmetric traffic First, one of N nodes n is selected.

  Next, Tx (n) and Rx (n) of the node n selected above are compared.

  Here, if both are the same level, it is detected that the node n is not vertically asymmetric.

  On the other hand, if Tx (n) is greater than or equal to Δ times Rx (n) or less than 1 / min, node n is detected as being vertically asymmetric.

  The above is performed for all N nodes n. The values of ◯, ●, Δ, and ▲ can be set to appropriate values according to the design conditions, but may be set to 2 or more, for example.

  At this time, the traffic bias detection unit 402 may detect all of the information or only a part of the information.

  The traffic bias detection unit 402 detects traffic bias information when it detects a node whose destination or transmission source is limited, when it detects traffic bias at the destination or transmission source, or when it detects vertical asymmetric traffic. Is generated and output to the traffic bias information holding unit 403. Here, the traffic bias information indicates that the traffic of the path having the node A as the transmission source node is biased about twice as large as the average value, and the node A is asymmetrical so that the top / bottom ratio is about 1: 4. Show.

  FIG. 10 shows operations performed after the operation of FIG. 7 until the group configuration table 406 is created, out of the entire detailed operations of the scheduler 40.

  Referring to FIG. 10, in step S4, the traffic bias detection unit 402 reads the traffic bias information from the traffic bias information holding unit 403 and passes the traffic bias information to the node grouper 404. The nodes 10 are logically grouped into G groups, and one or more group configuration candidates are created.

  For example, when the node grouper 404 creates a candidate in consideration of traffic bias of the transmission source node, if a path having a certain node as the transmission source node is a multiple of the average value, the node in the group to which the node belongs Try to reduce the number. Further, if the path having a certain node as a transmission source node is less than 1 / of the average value, the number of nodes in the group to which the node belongs is increased.

  In addition, when the node grouper 404 creates candidates in consideration of traffic bias of the destination node, if the path having a certain node as the destination node is a multiple of the average value, the number of nodes in the group to which the node belongs is Try to reduce it. Further, if the path having a certain node as the destination node is less than 1 / of the average value, the number of nodes in the group to which the node belongs is increased.

  Note that the grouping method when there is a limitation on the communication destination (source node or destination node) corresponds to the configuration of the fourth embodiment described later.

  For each group configuration candidate created by the node grouper 404, the selector 405 calculates, for each group, SumTS (G), which is the sum of the number of request TSs of a path having a node belonging to the group as a transmission source node. A variance value V {SumTS (G)} is calculated for SumTS (G). The selector 405 holds a one-dimensional array table that represents the group configuration of each candidate, the sum of the number of requested TSs of each group, and the variance value.

  Then, the selector 405 selects a candidate having the smallest variance value from among group combination candidates. Here, the second candidate having a variance value of 73 is selected.

  Next, in step S5, the selector 405 creates a one-dimensional array group configuration table 406 indicating the nodes and the number of nodes belonging to the group configuration selected above for each group, and stores the memory (not shown). ).

Thereafter, although illustration is omitted, TS allocation between groups is executed, and thereafter, TS allocation of paths between nodes in the group is executed step by step.
(Iii) Operation Sequence Next, the operation sequence of the network system of the present invention will be described.

  FIG. 11 shows an operation sequence of the entire network system.

  Referring to FIG. 11, it is assumed that the host computer Aa connected to the node A transmits an inter-host data signal addressed to the host computer Bb connected to the node B to the node A in steps A1 and A2.

  Then, in step A3, the node A receives and buffers the host-to-host data signal from the host computer Aa, and monitors the traffic amount required by the path from the node A to the node B for transmission of the host-to-host data signal. To do.

  Next, in step A4, the node A transmits to the scheduler 40 inter-host traffic information indicating the traffic volume (including the traffic volume monitored above) requested by each path whose source is the node A. Note that transmission of traffic information between hosts is periodically performed regardless of the presence or absence of data signals between hosts. Therefore, in step A5, the node A is also transmitting inter-host traffic information to the scheduler 40.

  Next, in steps A6, A7, A8, the scheduler 40 performs scheduling (scheduling between groups and scheduling within the group) based on the traffic information between hosts, and for each node A, B, each of the nodes. A node schedule table 415 representing the contents of node processing in the TS is transmitted.

  Thereafter, in steps A9, A10, and A11, the nodes A and B transfer the inter-host data signal addressed to the host computer Bb from the host computer Aa according to the contents of the node schedule table 415.

  12 and 13 show an operation sequence up to the creation of the group configuration table 406 in the operation sequence of the scheduler 40. FIG.

  Referring to FIG. 12, first, in step B1, the AC traffic volume totaling unit 401 periodically tabulates the traffic volume of the requested traffic requested from each node 10 by each path having the node 10 as a transmission source node. To do.

  Next, in step B2, the AC traffic volume totaling unit 401 determines whether the traffic volume has been totaled from all the nodes 10, and if so, in step B3, converts the traffic volume into the requested number of TSs, and traffic. Convert matrix to TS matrix.

  Next, in step B4, the AC traffic volume totaling unit 401 outputs and holds the TS matrix to the traffic deviation detection unit 402, the inter-group band allocation unit 407, and the intra-group (inter-node) band allocation unit 409. .

  The above steps B1 to B4 correspond to step S1 in FIG.

  Next, in step B5, the traffic bias detection unit 402 creates traffic bias information when detecting the traffic bias of each path in the network based on the TS matrix.

  The above step B5 corresponds to step S2 in FIG.

  Next, in step B6, the traffic bias detection unit 402 outputs the traffic bias information to the traffic bias information holding unit 403, and in step B7, the traffic bias information holding unit 403 holds the traffic bias information.

  The above steps B6 to B7 correspond to step S3 in FIG.

  Referring to FIG. 13, in step B8, the traffic bias detection unit 402 reads out the traffic bias information from the traffic bias information holding unit 403 and passes it to the node grouper 404. The node grouper 404 in step B9 receives the traffic bias information. Based on the above, N nodes 10 are logically grouped into G groups to create group configuration candidates. In step B10, the created group configuration candidates are notified to the selector 405.

  Next, in step B11, the selector 405 stores the group configuration candidates notified from the node grouper 404. In step B12, for each candidate group, the selector 405 stores a path whose source node is the node 10 belonging to the group. The sum of the requested TS numbers is calculated, and further, a variance value is calculated for the sum.

  Next, in step B13, the node grouper 404 determines whether all group configuration candidates have been created, and if so, gives a trigger for selecting a group to the selector 405 in step B14.

  Next, in step B <b> 15, the selector 405 selects a candidate having the smallest variance value from all the group configuration candidates notified from the node grouper 404.

  The above steps B8 to B15 correspond to step S4 in FIG.

  Next, in step B16, the selector 405 creates a group configuration table 406 representing the nodes 10 and the like belonging to the group for each group configuration selected above, and stores it in a memory (not shown) in step B17. Remember.

  The above steps B16 to B17 correspond to step S5 in FIG.

  14 to 16 show an operation sequence performed after the operation of FIG. 13 among the operation sequences of the scheduler 40.

  Referring to FIG. 14, next, in step B18, the group configuration table 406 is read from the memory, and the inter-group band allocation unit 407, the intra-group (inter-node) band allocation unit 409, the table conversion unit 410, and the table The data is output to the conversion unit 414.

  Next, in step B19, the inter-group band allocation unit 407 allocates TSs between the groups (inter-group scheduling).

  Next, in step B20, the intra-group (inter-node) bandwidth allocation unit 409 allocates TS to the path between the nodes 10 in the group (intra-group scheduling).

  At this time, the group requested TS number may be calculated by the intra-group (inter-node) band allocation unit 409 and notified to the inter-group band allocation unit 407.

  Next, in step B21, the inter-group band allocation unit 407, based on the TS allocation result between the groups, a link table (group unit) that indicates which group data flows in each TS of each link 20 408 is created and stored in the memory in step B22.

  Next, in step B23, the link table (group unit) 408 of each link is read from the memory and output to the table conversion unit 410 and the table conversion unit 412.

  Next, in step B24, the intra-group (inter-node) bandwidth allocation unit 409 outputs the TS allocation result of the path between the nodes 10 in the group to the table conversion unit 410. In step B25, the table conversion unit 410 A link table (group unit) 408 of each link 20 indicates which path data is to flow in each TS of each link 20 based on the TS allocation result from the intra-group (inter-node) bandwidth allocation unit 409. Convert to schedule table 411.

  Referring to FIG. 15, next, in steps B26 and B27, the table conversion unit 410 stores the link schedule table 411 of each link in the memory.

  Next, in step B28, the link schedule table 411 of each link is read from the memory and output to the table conversion unit 414.

  Next, in step B29, the table conversion unit 412 represents a node processing content (data transmission, route switching, etc.) in each TS of each group based on the link table (group unit) 408 of each link 20. A group table 413 is created, and the node group table 413 of each group is stored in the memory in steps B30 and B31.

  Next, in step B32, the node group table 413 of each group is read from the memory and output to the table conversion unit 414.

  Next, in step B33, the table conversion unit 414 converts the node group table 413 of the group into the node processing contents (data transmission, route) in each TS of each node 10 based on the link schedule table 411 of each link 20. To a node schedule table 415 representing switching etc.).

  Referring to FIG. 16, next, in steps B34 and B35, the table conversion unit 414 stores the node schedule table 415 of each node 10 in the memory.

  Next, in step B36, the node schedule table 415 of each node 10 is read from the memory and output to the table transmitter 416.

  Thereafter, in steps B37 and B38, the table transmitter 416 transmits the node schedule table 415 of the node to each node 10.

  In the first embodiment, since the grouping of N nodes 10 is performed, the TS allocation of the paths between the nodes in the group can be executed step by step after the TS allocation between the groups. Compared with the method, the calculation time of TS allocation can be reduced.

In the first embodiment, N nodes 10 are grouped in consideration of traffic bias, so that an increase in the schedule length (the time required to transfer the requested traffic) can be suppressed. it can.
(2) Second Embodiment As a period for performing grouping, it is conceivable to perform grouping for each control period in which the traffic amount is periodically counted from each node 10.

  In this case, it is possible to appropriately cope with the traffic bias, but on the other hand, the grouping is performed every time, and the calculation time of the grouping is required, so that the schedule time increases every time.

  Therefore, in the second embodiment, a detection timer is provided, and grouping is performed once every a plurality of times in the above control cycle.

  FIG. 17 shows the configuration of the scheduler 40 of the second embodiment. In FIG. 17, the same components as those in FIG.

  Referring to FIG. 17, the scheduler 40 of the second embodiment includes a detection timer 418, a traffic amount history holding unit 419, an average value calculating unit 420, and the first embodiment shown in FIG. , Is different.

  The detection timer 418 performs a countdown operation. Note that the time from the start to the completion of the count is longer than the control cycle in which the AC traffic volume totalization unit 401 periodically totals the traffic volume from each node 10.

  The traffic volume history holding unit 419 holds the traffic volume history requested by each path, which is totaled by the AC traffic volume totaling unit 401.

  The average value calculation unit 420 receives the timeout of the detection timer 418 and detects the average value of the traffic amount requested by each path.

  The traffic bias detection unit 402 instructs the node grouper 404 to perform grouping again based on the average value of the traffic amount requested by each path.

  FIG. 18 shows the entire detailed operation of the scheduler 40. The detailed operation of the scheduler 40 of the second embodiment is different from that of the first embodiment only in that steps S1 and S2 in FIG. 7 are replaced with steps S1-1 and S2-1. Is the same.

  Referring to FIG. 18, first, in step S <b> 1-1, the AC traffic amount totaling unit 401 periodically requests traffic from each of the nodes A to H to be requested by each path having that node as a transmission source node. The amount is totaled, the traffic amount is converted into the required number of TSs, and the traffic matrix is converted into a TS matrix. The TS matrix is output to and held in the traffic volume history holding unit 419, the inter-group band allocation unit 407, and the intra-group (inter-node) band allocation unit 409.

  In addition, the average value calculation unit 420 receives the timeout of the detection timer 418 and creates an average TS matrix based on the most recent predetermined number of TS matrix histories. The average TS matrix represents an average value of the number of requested TSs required by each path between the nodes, and is a two-dimensional array having elements of the number of nodes × the number of nodes. Note that the TS matrix history can be used as t1 to t10, t11 to t20, t21 to t30,... (= Jumping window method), t1 to t10, t2 to t12, t3 to t13, There is a usage method (= moving window method). The average TS matrix is output to the traffic deviation detection unit 402 and held.

  Next, in step S 2-1, when the traffic bias detection unit 402 detects traffic bias based on the average TS matrix, it generates traffic bias information indicating the traffic bias status of each path in the network. Then, the node grouper 404 is instructed to perform grouping again. Here, the traffic volume of the path having node A as the transmission source node is large, and the traffic volume of the path having nodes D, E, and F as the transmission source node and nodes A, B, and C as the destination nodes is small. This is detected, and another grouping is instructed.

  FIG. 19 shows an operation sequence of the scheduler 40. FIG. 19 shows an operation sequence corresponding to steps S1-1 and S2-1 different from the first embodiment.

  Referring to FIG. 19, first, in step C <b> 1, the AC traffic volume totaling unit 401 periodically tabulates the traffic volume of the requested traffic requested by each path that has that node 10 as the transmission source node. To do.

  Next, when the traffic volume totaling unit 401 totals the traffic volume from all the nodes 10, in step C2, the traffic volume is converted into the requested TS number, the traffic matrix is converted into the TS matrix, and the TS matrix is converted into the TS matrix. The traffic volume history holding unit 419, the inter-group band allocation unit 407, and the intra-group (inter-node) band allocation unit 409 are output and held.

  Next, in step C3, the traffic amount history holding unit 419 updates the history of the TS matrix.

  On the other hand, in step C4, the detection timer 418 performs a countdown operation, and when the count is completed in step C5, in step C6, the average value of the number of requested TSs of each path is sent to the average value calculation unit 420. Is calculated, and then reset is performed in step C7.

  Next, in step C8, when the average value calculation unit 420 receives the above instruction from the detection timer 418, in step C9, the history of the TS matrix corresponding to the latest predetermined number is read from the traffic amount history holding unit 419, and the step In C10, the average value of the number of requested TSs of each path is calculated, and an average TS matrix is created. In step C11, the average TS matrix is passed to the traffic bias detection unit 402.

  The above steps C1 to C11 correspond to step S1-1 in FIG.

  Next, in step C12, when the traffic bias detection unit 402 detects traffic bias of each path in the network based on the average TS matrix, in step C13, the node biaser 404 is instructed to perform grouping again. To do.

  The above steps C12 to C13 correspond to step S2-1 in FIG.

  In the second embodiment, the traffic bias detection unit 402 creates traffic bias information based on the average TS matrix, holds the traffic bias information in the traffic bias information holding unit 403, and the traffic bias information. Information is passed to the node grouper 404 to instruct grouping.

  As described above, in the second embodiment, grouping is performed again only when the traffic bias is large on average.

  Therefore, since the number of groupings can be reduced, the scheduling time can be further reduced.

Other effects are the same as those of the first embodiment.
(3) Third Embodiment As a period for performing grouping, it is conceivable to perform grouping for each control period in which the traffic amount is periodically counted from each node 10.

  In this case, it is possible to appropriately cope with the traffic bias, but on the other hand, the grouping is performed every time, and the calculation time of the grouping is required, so that the schedule time increases every time.

  Therefore, in the third embodiment, grouping is performed only when there is a large change in traffic bias.

  FIG. 20 shows the configuration of the scheduler 40 of the third embodiment. In FIG. 20, the same components as those in FIG.

  Referring to FIG. 20, the scheduler 40 of the third embodiment is different from the first embodiment shown in FIG. 2 in that a threshold determination unit 421 is added.

  When the traffic bias detection unit 402 newly creates traffic bias information, the threshold determination unit 421 uses a difference between the new traffic bias information and the current traffic bias information held in the traffic bias information holding unit 403 as a threshold. The traffic bias information held in the traffic bias information holding unit 403 is updated based on the comparison result, and the grouping of the node grouper 404 is instructed again.

  FIG. 21 shows the entire detailed operation of the scheduler 40. In the detailed operation of the scheduler 40 of the third embodiment, steps S1, S2, and S3 in FIG. 7 are replaced with steps S1-2, S2-2, and S3-2 as compared with the first embodiment. Only the point is different and the others are the same.

  Referring to FIG. 21, first, in step S1-2, the AC traffic volume totaling unit 401 periodically requests traffic from each of the nodes A to H to be requested by each path having the node as a transmission source node. The amount is totaled, the traffic amount is converted into the required number of TSs, and the traffic matrix is converted into a TS matrix. The TS matrix represents the number of requested TSs required by each path between the nodes, and is a two-dimensional array having elements of the number of nodes × the number of nodes. Here, the traffic amount of 10 Mbps is converted to 1 TS. The TS matrix is output to the traffic deviation detection unit 402, the inter-group band allocation unit 407, and the intra-group (inter-node) band allocation unit 409, and is held there.

  Next, in step S2-2, when the traffic bias detection unit 402 detects traffic bias of each path in the network based on the TS matrix, the traffic bias detection unit 402 creates traffic bias information indicating the traffic bias state. Here, the traffic volume of the path having node A as the transmission source node is large, and the traffic volume of the path having nodes D, E, and F as the transmission source node and nodes A, B, and C as the destination nodes is small. Yes.

  Next, in step S <b> 3-2, the threshold determination unit 421 determines the difference between the traffic bias information newly created by the traffic bias detection unit 402 and the current traffic bias information held in the traffic bias information holding unit 403 as a threshold value. And based on the comparison result, if it is determined that there has been a significant change in the traffic bias, the traffic bias information held in the traffic bias information holding unit 403 is updated and the node grouper 404 is again updated. Instructing grouping.

In addition, the following method is mentioned as a method in which the threshold determination unit 421 determines that there has been a significant change in the traffic bias.
1. Number of nodes involved in bias fluctuation For example, if the number of requested TSs changes in a predetermined ratio (for example, 25%) or more of all nodes (for example, two of eight nodes), there is a significant change in traffic bias. 1. Number of TSs involved in bias fluctuation For example, if there is a node whose request TS number has changed by a predetermined number or more (for example, 50 TS or more) (for example, the request TS number of node A changes from 71 TS to 20 TS), the traffic bias will change greatly. If both events 3.1 and 2 are determined to have occurred, it is determined that there has been a significant change in traffic bias. FIG. 22 shows an operation sequence of the scheduler 40. FIG. 22 shows an operation sequence corresponding to steps S1-2, S2-2, and S3-2 different from the first embodiment.

  Referring to FIG. 22, first, in step D <b> 1, the AC traffic volume totaling unit 401 periodically tabulates the traffic volume of requested traffic requested from each node 10 by each path having the node 10 as a transmission source node. To do.

  Next, in step D2, the AC traffic volume totalizing unit 401 determines whether the traffic volume has been totaled from all the nodes 10, and if so, in step D3, converts the traffic volume into the requested number of TSs and traffic. Convert matrix to TS matrix.

  Next, in step D4, the AC traffic volume totaling unit 401 outputs and holds the TS matrix to the traffic bias detection unit 402, the inter-group band allocation unit 407, and the intra-group (inter-node) band allocation unit 409. .

  The above steps D1 to D4 correspond to step S1-2 in FIG.

  Next, in step D5, when the traffic bias detection unit 402 detects traffic bias of each path in the network based on the TS matrix, in step D6, it newly generates traffic bias information and determines the threshold value. Output to the unit 421.

  The above steps D5 to D6 correspond to step S2-2 in FIG.

  Next, in step D7, the threshold value determination unit 421 reads the current traffic bias information from the traffic bias information holding unit 403, and in step D8, compares the difference between the current traffic bias information and the new traffic bias information with the threshold value. Then, based on the comparison result, it is determined whether or not there has been a significant change in traffic bias.

  Next, when there is a significant change in the traffic bias, the threshold value determination unit 421 notifies the traffic bias information holding unit 403 in step D9, and in step D11, the traffic bias information holding unit 403 Update traffic bias information. In step D10, the updated traffic bias information is passed to the node grouper 404 to instruct further grouping.

  The above steps D7 to D11 correspond to step S3-2 in FIG.

  As described above, in the third embodiment, grouping is performed again only when there is a large change in the traffic bias.

  Therefore, since the number of groupings can be reduced, the scheduling time can be further reduced.

Other effects are the same as those of the first embodiment.
(4) Fourth Embodiment In the fourth embodiment, when communication destinations (source nodes and destination nodes) are limited, grouping is performed considering this.

  FIG. 23 shows the configuration of the scheduler 40 of the fourth embodiment. In FIG. 23, the same components as those in FIG. 2 are denoted by the same reference numerals, and description thereof is omitted.

  Referring to FIG. 23, the scheduler 40 of the fourth embodiment is different from the first embodiment shown in FIG. 2 in that an upper node / lower node management unit 422 is added.

  24 and 25 show the operation of the scheduler 40 of the fourth embodiment.

  Referring to FIG. 24, for example, the network system has a unidirectional ring configuration with N = 8, and there is one channel that can be connected to the same slot per link (no fiber multiplexing or wavelength multiplexing). It shall be. Further, as shown in the TS matrix, it is assumed that a path exists only at a portion surrounded by a rounded rectangle.

  In this case, the network system is a unidirectional ring in terms of the physical network configuration, but logically includes a set of nodes X and Y and a set of nodes A, B, C, D,. Can be considered separately.

  Therefore, among the above two sets, the set of nodes X and Y, that is, the set having the larger number of communication destinations is set as the upper node, the set of nodes A, B, C, D,. The pair with the smaller number is called the lower node, the node whose communication destination is the number of lower nodes is defined as the upper node, the node whose communication destination is the number of upper nodes is defined as the lower node, and the upper node and lower node The management is done separately.

  As described above, the traffic bias detection unit 402 can calculate the dst (n) and src (n): the number of non-zero elements in the nth column of each of the N nodes n by using the TS matrix as an input. It is.

  Therefore, the traffic bias detection unit 402 sets a node where dst () = src () = L as an upper node and dst () = src when the number of upper nodes is U and the number of lower nodes is L (U <L). A node where () = U can be determined as a lower node.

  Therefore, when there is a limitation in the communication destination, the traffic bias detection unit 402 notifies the upper node / lower node management unit 422 of information on the upper node and lower node, and the upper node / lower node management unit 422 Holds node and subordinate node information.

When the communication destination is limited based on the traffic bias information, the node grouper 404 groups the nodes according to the following criteria based on the traffic bias information and information on the upper and lower nodes.
-Upper node and lower node are not included in the same group-Lower nodes that straddle the upper node are not included in the same group-Upper nodes should be separated as much as possible → When entering the same group, physical network Priority is given to adjacent nodes above, and physically separated nodes are not grouped together. When grouping lower nodes, the traffic volume of each group (the path of the path that uses the node belonging to the group as the source node) Referring to FIG. 25, in this example, the node grouper 404 groups the upper nodes A and B in different groups and groups the lower nodes.

  As a result, when the inter-group bandwidth allocation unit 407 allocates TSs between the groups, it is avoided to allocate blocks for one ring (Effect 1).

  Further, when the inter-group bandwidth allocation unit 407 allocates TSs between the groups, it is possible to allocate a block having a higher node as a transmission source node and a block having a destination node as the same TS (Effect 2).

  Further, when the inter-group bandwidth allocation unit 407 allocates TSs between the groups, the block heights can be made uniform (effect 3).

  Thereby, the empty area on the schedule table can be reduced.

  FIG. 26 shows examples of effects 1, 2, and 3 of the fourth embodiment.

  Referring to FIG. 26, the effect 1 avoids the allocation of blocks for one round of the ring such as node A → node X.

  Also, due to the effect 2, it is possible to assign a block having the upper node X as a transmission source node and a block having a destination node as the same TS.

  In addition, due to the effect 3, the heights of the blocks of the group G1 → G2 and the group G3 → G4 can be made uniform.

  FIG. 27 shows a specific example of the operation of the scheduler 40 of the fourth embodiment.

  Referring to FIG. 27, for example, the network system has a unidirectional ring configuration with N = 8, and one channel that can be simultaneously connected to the same slot per link (no fiber multiplexing or wavelength multiplexing is performed). It shall be. In addition, it is assumed that node numbers are assigned in the order of clockwise or counterclockwise (in FIG. 27, clockwise) on the ring.

  In the case of the above assumption, the number of lower nodes existing between the upper nodes on the physical network can be calculated from the node numbers X and Y of the upper nodes (in FIG. 27, there are two nodes between the upper nodes X → Y, 4 nodes between nodes Y → X).

  Based on these, the node grouper 404 performs grouping according to the situation (in FIG. 27, grouping into 5 groups).

A group is composed of only the upper node (in FIG. 27, the upper node X constitutes the group G1 and the upper node Y constitutes the group G2)
The grouping of the lower nodes does not cross the upper nodes, group adjacent nodes, and then group them so that the sum of the requested TS numbers of each group is as uniform as possible (in FIG. The lower nodes A and B constitute a group lower G3, the lower nodes C and D constitute a group lower G4, and the lower nodes E and E do not make C the same group, so that the number of requested TS is uniform. Group G5)
FIG. 28 shows an operation sequence of the scheduler 40. The detailed operation of the scheduler 40 of the fourth embodiment is different from that of the first embodiment only in that step S3 in FIG. 7 is replaced with step S3-3 (not shown), and the other operations are the same. It is. FIG. 22 mainly shows an operation sequence corresponding to step S3-2 different from that of the first embodiment.

  Referring to FIG. 28, first, operations in steps E1 to E5 similar to steps B1 to B5 shown in FIG. 12 are performed.

  The above steps E1 to E4 correspond to step S1 in FIG. 7, and step E5 corresponds to step S2 in FIG.

  Next, in step E6, when there is a limitation on the communication destination (source node or destination node), the traffic bias detection unit 402 notifies the upper node / lower node management unit 422 of the information on the upper node and the lower node. In step E7, the upper node / lower node management unit 422 holds information on the upper node and the lower node. This information is read to the node grouper 404 in step E8 and used for grouping. This reading may be performed in advance or may be performed at the time of grouping.

  Next, in step E9, the traffic bias detection unit 402 outputs the traffic bias information to the traffic bias information holding unit 403, and in step E10, the traffic bias information holding unit 403 holds the traffic bias information.

  The above steps E6 to E10 correspond to the above-described step S3-3.

  As described above, in the fourth embodiment, when communication destinations are limited, grouping can be performed in consideration of this.

Other effects are the same as those of the first embodiment.
(5) Fifth Embodiment In the first to fourth embodiments described above, the configuration is based on a single ring having one wavelength (no fiber multiplexing or wavelength multiplexing).

  On the other hand, in the fifth embodiment, the present invention can be applied to a configuration in which the number of wavelengths is W (W is a natural number of 2 or more) in a single ring.

  When W waves are multiplexed on each fiber using WDM (Wavelength Division Multiplex), each link has a W schedule schedule table.

  However, on the W-side link schedule table, the same node is not allowed to become a transmission source node or a destination node at the same timing.

  Therefore, the scheduler 40 of the fifth embodiment has a management function for managing the W-side link schedule table.

  FIG. 29 shows the configuration of the scheduler 40 of the fifth embodiment. 29, the same components as those in FIG. 2 are denoted by the same reference numerals, and description thereof is omitted.

  Referring to FIG. 29, the scheduler 40 according to the fifth embodiment has a point having a W plane of a link table (group unit) 408 and a link schedule table 411, as compared with the first embodiment shown in FIG. It differs from the point which added the space | interval adjustment parts 423,424.

  The inter-wavelength adjustment unit 423 performs adjustment so that the same node 10 does not become a transmission source node or a destination node at the same timing on the link schedule table 411 on the W plane, and the adjustment result is within the group (between nodes). ) Feedback to the bandwidth allocation unit 409.

  The inter-wavelength adjustment unit 424 performs adjustment so that the same node 10 does not become a transmission source node or a destination node at the same timing on the link table (group unit) 408 on the W plane, and the adjustment result is obtained between the groups. This is fed back to the bandwidth allocation unit 407.

  When the table conversion unit 414 creates the node schedule table 415, the link schedule table 411 for the W plane of the link starting from the relevant node, the link schedule table 411 for the W plane of the link having the end point, To determine the contents of node processing (data transmission, route switching, etc.) in each TS of the corresponding node.

  As described above, the fifth embodiment can be applied to a network system having a single ring and W wavelengths.

Other effects are the same as those of the first embodiment.
(6) Sixth Embodiment In the first to fifth embodiments described above, it is assumed that the node grouper 404 logically groups a plurality of nodes 10 into a group number G set in advance. .

  At this time, the node grouper 404 performs grouping so that the number of requested TSs in each group is uniform in order to cope with the traffic bias, but the number of requested TSs in each group may not be almost uniform.

  Therefore, in the sixth embodiment, a function for adjusting the number of groups G is provided in order to make the number of required TSs of each group substantially uniform.

  For example, it is assumed that the preset number of groups G is 10 for 100 nodes.

  Here, when the number of groups G = 10, it is assumed that the minimum value of the variance value (see FIG. 10) of the sum of the number of requested TSs of each group is 150.

  On the other hand, if the number of groups G = 8, 9, 11, and 12 can be minimized to the variance values 140, 170, 90, and 130, respectively, the group number G can be minimized 11. Adjust to the operation.

  Thereby, it becomes possible to further suppress the increase in the schedule length.

  FIG. 30 shows a part of the configuration of the scheduler 40 of the sixth embodiment. In FIG. 30, the same components as those in FIG. 2 are denoted by the same reference numerals, and description thereof is omitted.

  Referring to FIG. 30, the scheduler 40 of the sixth embodiment is different from the first embodiment shown in FIG. 2 in that a group number adjustment unit 425 is added. The components other than the components shown in FIG. 30 are the same as those in FIG.

  For example, the group number adjustment unit 425 extracts four adjustment candidates within a range of ± 30% of the preset group number G, and calculates the minimum variance value in the number of groups of each adjustment candidate as the original value. Compared with the minimum value of the variance value in the group number G, the group value is adjusted so as to minimize the variance value.

Therefore, the group number adjustment unit 425 has the following functions.
A function for extracting adjustment candidates for the number of groups G. The node grouper 404 is notified of the five types of groups (the original group number G and the number of groups of the four adjustment candidates) one by one to instruct grouping. Function ・ Function to memorize “minimum variance value and group configuration at that time” for each group number ・ Function to compare minimum variance value for each group number ・ Each group in the number of groups with the smallest variance value Next, the function of creating the node configuration table 406 representing the nodes 10 belonging to the group and storing them in the memory. As described above, the sixth embodiment has the function of adjusting the number of groups. The increase can be further suppressed.

  Other effects are the same as those of the first embodiment.

  The present invention has been described above with reference to the embodiments, but the present invention is not limited to the above embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present invention without departing from the gist of the present invention.

  For example, in the first to sixth embodiments, an example in which scheduling within groups (TS allocation) is executed step by step after scheduling between groups (TS allocation) has been described. The order is not limited to this, and scheduling between groups may be executed step by step after scheduling within the group. Even in this case, the calculation time for TS allocation can be reduced.

  The first to sixth embodiments may be applied in combination.

  The method performed by the scheduler 40 of the present invention may be applied to a program for causing a computer to execute. In addition, the program can be stored in a storage medium and can be provided to the outside via a network.

10 nodes 20 links 30 host computer 40 scheduler 401 AC traffic volume totaling unit 402 traffic bias detecting unit 403 traffic bias information holding unit 404 node grouper 405 selector 406 group configuration table 407 inter-group bandwidth allocation unit 408 link table (group unit)
409 Intra-group (inter-node) bandwidth allocation unit 410 Table conversion unit 411 Link schedule table 412 Table conversion unit 413 Node group table 414 Table conversion unit 415 Node schedule table 416 Table transmitter 417 Timer 418 Detection timer 419 Traffic amount history holding unit 420 Average Value calculation unit 421 Threshold determination unit 422 Upper node / lower node management unit 423,424 Inter-wavelength adjustment unit 425 Group number adjustment unit

Claims (8)

A scheduler constituting a TDM network system,
Aggregating means for periodically totaling the amount of traffic required by each path having the node as a transmission source node from each node constituting the network system;
Based on the amount of traffic requested by each path, when detecting the traffic bias of each path in the network, traffic bias detection means for creating traffic bias information indicating the traffic bias state;
Traffic bias information holding means for holding the traffic bias information;
And a grouping means for grouping a plurality of nodes constituting the network system based on the traffic bias information. Traffic volume history holding means for holding a history of traffic volume requested by each path;
A timer that performs a countdown operation;
An average value calculating means for calculating an average value of the traffic amount requested by each path based on the most recent predetermined number of histories in response to a timeout of the timer;
The traffic bias detection means includes:
2. The scheduler according to claim 1, wherein when the traffic deviation of each path in the network is detected based on an average value of the traffic amount requested by each path, the grouping unit is instructed to perform grouping again. 3. .   When the traffic bias detection means newly creates traffic bias information, the difference between the new traffic bias information and the current traffic bias information held in the traffic bias information holding means is compared with a threshold value, and the comparison result The system according to claim 1, further comprising: a threshold value determination unit that updates the traffic bias information held in the traffic bias information holding unit and instructs the grouping unit to perform grouping again based on the information. Scheduler. It further has an upper node / lower node management means,
The traffic bias detection means includes:
When a node whose communication destination is limited is detected among the nodes constituting the network system, a node whose communication destination is limited to the number of lower nodes is defined as an upper node, and the communication destination is equal to the number of upper nodes. Define a limited node as a lower node, and hold the upper node and lower node information in the upper node / lower node management means,
The grouping means includes
The scheduler according to any one of claims 1 to 3, wherein a plurality of nodes constituting the network system are grouped based on the traffic bias information and information on the upper node and the lower node. The network system has a unidirectional ring configuration with a wavelength number of W (W is a natural number of 2 or more),
A link schedule table representing the schedule of each link constituting the network system is provided on the W plane,
5. The inter-wavelength adjustment unit further adjusts so that the same node does not become a transmission source node and a destination node at the same timing on the W-side link schedule table. The scheduler described in the section.   The system further comprises group number adjusting means for notifying the grouping means of a plurality of groups one by one, instructing grouping, and adjusting the number of groups based on the grouping result in each group number. Item 6. The scheduler according to any one of Items 1 to 5. A scheduling method by a scheduler constituting a TDM network system,
Periodically totaling the amount of traffic required by each path having the node as a transmission source node from each node constituting the network system;
When detecting traffic deviation of each path in the network based on the traffic amount requested by each path, creating and maintaining traffic deviation information indicating the traffic deviation state; and
Grouping a plurality of nodes constituting the network system based on the traffic bias information. In the scheduler that configures the TDM network system,
A procedure for periodically totaling the amount of traffic required by each path having the node as a transmission source node from each node constituting the network system;
A procedure for creating and maintaining traffic bias information indicating a traffic bias state when detecting a traffic bias of each path in the network based on the traffic volume requested by each path;
A program for executing a procedure for grouping a plurality of nodes constituting the network system based on the traffic bias information.