JP2015029183A - Tdm network system and scheduling method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a TDM network system which makes it possible to perform, at high speed and efficiently, time slot allocation calculation in a large scale network.SOLUTION: A TDM network system comprises: a network including a plurality of nodes for performing data transfer to another adjacent node according to a time slot schedule; and a scheduler which is connected to any one node and creates the time slot schedule on the basis of traffic aggregated by the network. The scheduler calculates the number of required time slots from the aggregated traffic; groups the plurality of nodes; calculates traffic fluctuation difference of respective groups; performs scheduling inside each group and scheduling among the respective groups on the basis of information on the difference; and determines a time slot schedule of the entire system by applying results of the scheduling inside the groups to a result of the scheduling among the groups.

Description

  The present invention relates to a device for dynamically allocating resources and a resource allocation method (scheduling method) in a network system using Time Division Multiplexing (TDM) technology.

  In the TDM network, data transfer between nodes is performed by determining the timing (referred to as a time slot (TS)) at which data is transmitted in advance. FIG. 18 is a diagram for explaining a data transfer method by a conventional TDM network system.

  In the TDM network system, it is necessary to allocate TSs as much as possible for efficient data transfer. Therefore, scheduling for determining the communication TS between the nodes is performed in advance.

  A table representing the allocation status of TSs defined on each link in the network (NW) is called a schedule table. Scheduling corresponds to an operation of assigning communication on this table. In the TDM network system, the operation is repeated with a TDM frame length (t) composed of a plurality of TSs as a unit.

  Referring to FIG. 18, when a common time slot (S1) is assigned to the via link of node A → node B → node C, the same time slot cannot be used on the same link of node B → node C. Another time slot (S2, S3) is assigned to the transit link from node B to node C to node D.

  In this way, the TDM network system determines and operates the data transfer schedule in advance.

  There are various heuristic methods for the TDM scheduling method, for example, the method of Non-Patent Document 1. In this method, TSs are sequentially assigned to a given topology and traffic one path at a time.

  In order to increase the scale of an NW system that dynamically allocates bandwidth within an NW according to communication traffic (for example, in the order of seconds or less), the inventors of the present application have a hierarchical system (see Non-Patent Document 2). And Rip-up & Re-allocate method (Non-Patent Document 3).

  In the hierarchization method disclosed in Non-Patent Document 2, the resource allocation problem (synonymous with the above-mentioned scheduling) is divided into small units, and each is solved independently to speed up the allocation calculation.

  In the Rip-up & Re-allocate method disclosed in Non-Patent Document 3, resource allocation (synonymous with the above-mentioned scheduling) focusing on only the difference in traffic fluctuation is performed, and the part where fluctuation does not occur is not subject to calculation. By doing so, the allocation calculation is speeded up.

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 Nakagawa, Hattori, Kimijima, Katayama, Misawa, Hiramatsu, “Proposal of optical TDM scheduling method HOTS using layering,” Society of Science Society Conference, B-6-25, September 2012 Nakagawa, Hattori, Kimijima, Katayama, Misawa, "Proposal of Rip-up & Re-allocate Dynamic Bandwidth Allocation Method Focusing on Traffic Fluctuation Difference in Optical L2 Switch Networks," IEICE General Conference, B-6-3, 2013 March

  In the conventional method, when the scale of the NW increases, it takes time to calculate TS allocation necessary for NW control with the same calculation resource, and there is a possibility that the effect of speeding up is insufficient.

  The present invention has been made to solve the above-described problems of the technology, and provides a TDM network system and a scheduling method capable of performing time slot allocation calculation in a large-scale network at high speed and efficiency. The purpose is to provide.

To achieve the above object, the TDM network system of the present invention comprises:
A network including a plurality of nodes performing data transfer to other adjacent nodes in units of time slots according to a time slot schedule;
A scheduler that is connected to any one of the plurality of nodes, aggregates traffic of the network, and creates the time slot schedule based on the aggregated traffic;
The scheduler
Calculate the required number of time slots from the aggregated traffic, group a plurality of nodes included in the network, calculate the traffic fluctuation difference of each group, and based on the information of the difference, schedule and group within the group The time slot schedule of the entire system is determined by performing scheduling between each other and applying the scheduling result within the group to the scheduling result between the groups.

The scheduling method of the present invention is a scheduling method by a scheduler that provides the time slot schedule to a network including a plurality of nodes that perform data transfer to other adjacent nodes in time slot units according to the time slot schedule. There,
Calculate the required number of time slots from the traffic aggregated in the network,
Grouping a plurality of nodes included in the network, calculating a traffic fluctuation difference of each group,
Based on the calculated difference information, scheduling within the group and scheduling between the groups, respectively,
The scheduling result within the group is applied to the scheduling result between groups to determine the time slot schedule of the entire system.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to shorten the time of TS allocation calculation required for network control, and can provide a larger-scale network system.

It is a block diagram which shows the example of 1 structure of the TDM network system of 1st Embodiment. FIG. 2 is a block diagram illustrating a configuration example of a node illustrated in FIG. 1. FIG. 2 is a block diagram illustrating a configuration example of a scheduler illustrated in FIG. 1. The combination of the arithmetic unit and the memory shown in FIG. 2 is replaced with a functional block. It is a flowchart which shows the operation | movement procedure of the scheduler of 1st Embodiment. It is a flowchart which shows the process in step 203 shown in FIG. It is a figure for demonstrating the mechanism of scheduling acceleration. It is a figure for demonstrating the subject which arises by the bias of traffic. It is a figure for demonstrating the operation | movement by the TDM network system of 2nd Embodiment. It is the schematic diagram which imaged the mode of the dynamic group rearrangement. FIG. 3 is a block diagram illustrating a configuration example of a scheduler in the TDM network system according to the first embodiment. FIG. 6 is a block diagram illustrating a configuration example of a TDM network system according to a second embodiment. FIG. 10 is a block diagram illustrating a configuration example of a TDM network system according to a third embodiment. FIG. 10 is a block diagram illustrating a configuration example of a scheduler in a TDM network system according to a fourth embodiment. It is a table | surface for demonstrating the various signals shown in FIG. It is a block diagram which shows the structural example of the TDM network system of 3rd Embodiment. It is a figure for demonstrating operation | movement of the TDM network system of 3rd Embodiment. It is a figure for demonstrating the data transfer method by the conventional TDM network system.

  An embodiment of the TDM network system and scheduling method of the present invention will be described when the network configuration is a ring topology.

  In the embodiment, for convenience of explanation, the number of channels that can be simultaneously connected to the same slot per one-way ring and link is assumed to be one (no fiber multiplexing or wavelength multiplexing is performed). For example, the number of wavelengths is 1, the TDM frame length is 10 ms, and the TS length is 100 μs. The present invention can also be applied to a case where an arbitrary topology and wavelength multiplexing are performed including a bidirectional ring. Thereby, a link schedule table is created on each wavelength of each link. Further, it is assumed that the scheduling is a process executed inside the scheduler.

(First embodiment)
The configuration of the TDM network system of this embodiment will be described.

  FIG. 1 is a block diagram showing a configuration example of the TDM network system according to the present embodiment.

  As shown in FIG. 1, nodes 20a to 20e are connected in a ring shape, and a host computer 30 is connected to each node. The scheduler 10 is connected to the node 20a.

  FIG. 2 is a block diagram showing a configuration example of the node shown in FIG. FIG. 3 is a block diagram showing a configuration example of the scheduler shown in FIG.

  As illustrated in FIG. 2, each of the nodes 20 a to 20 e includes a controller 21, a switch 22, a Tx (Transmitter) 23, an Rx (Receiver) 24, a buffer 25, and an IO 26. As shown in FIG. 3, the scheduler 10 includes an arithmetic device 11, a memory 13, and an IO (input / output unit) 15.

  Table 1 shows an outline of input signals, processing contents, and output signals for each of the nodes 20a to 20e.

  Table 2 shows an outline of input signals, processing contents, and output signals for the scheduler 10.

  FIG. 4 is obtained by replacing the combination of the arithmetic unit and the memory shown in FIG. 2 with functional blocks. The arithmetic device 11 and the memory 13 shown in FIG. 2 have a configuration related to the hierarchization method, a configuration related to the Rip-up & Re-allocate method, and a configuration newly added by a combination of these two methods. Have

  Among the configurations shown in FIG. 4, the configuration related to the hierarchization method includes a node grouper 121, a traffic information holding unit 122, a node control signal generation unit 123, an actual schedule table holding unit 124, an actual schedule table calculation unit 125, and an intra-group schedule. A table holding unit 126 and an inter-group schedule table holding unit 127.

  The configuration related to the Rip-up & Re-allocate method includes an increment paste change unit 131, a decrease strip removal change unit 132, an increment paste change unit 133, a decrease strip removal change unit 134, and a timer 135. .

  The configuration added in the present embodiment by the combination of these two methods includes an intra-group traffic fluctuation difference calculation unit 111, an intra-group traffic fluctuation difference holding unit 112, an inter-group traffic fluctuation difference calculation unit 113, and an inter-group traffic fluctuation difference. A holding unit 114, an intra-group allocation history holding unit 115, and an inter-group allocation holding unit 116.

  In the present embodiment, the case where each unit illustrated in FIG. 4 is configured by a dedicated electronic circuit will be described. However, a CPU (Central Processing Unit) is provided in the arithmetic device 11, and a program for executing scheduling is stored in the memory 13. 4 may be virtually configured in the scheduler 10 by executing software (program) by a computer (arithmetic apparatus).

  In the present embodiment, a detailed description of the configurations related to each of the “layered scheme” and “Rip-up & Re-allocate scheme” that have already been realized among the configurations shown in FIG. 4 is omitted. The configuration related to the combination of the two methods will be described in detail.

  The intra-group allocation history holding unit 115 holds an intra-group schedule that is a schedule indicating TS allocation within a group based on the previous TS schedule.

  The inter-group allocation holding unit 116 holds an inter-group schedule that is a schedule indicating TS allocation between groups based on the previous TS schedule.

  The traffic information holding unit 122 counts the traffic in the NW, calculates a required TS from the totaled traffic information, and holds the required TS information.

  The node grouper 121 groups the nodes 20a to 20e from the traffic information collected by the traffic information holding unit 122, and notifies the intra-group traffic fluctuation difference calculation unit 111 and the inter-group traffic fluctuation difference calculation unit 113 of the grouping result.

  The intra-group traffic fluctuation difference calculation unit 111 calculates the intra-group TS fluctuation difference from the grouping result notified from the node grouper 121 and the information on the number of required TSs held by the traffic information holding unit 122, and the calculation result is used as the intra-group traffic difference. It is passed to the fluctuation difference holding unit 112.

  The intra-group traffic fluctuation difference holding unit 112 holds the intra-group traffic fluctuation difference information received from the intra-group traffic fluctuation difference calculation unit 111.

  The inter-group traffic fluctuation difference calculation unit 113 calculates an inter-group TS fluctuation difference from the grouping result notified from the node grouper 121 and the information on the required number of TS held by the traffic information holding unit 122, and the calculation result is used as the inter-group traffic fluctuation. It is passed to the difference holding unit 114.

  The inter-group traffic fluctuation difference holding unit 114 holds the inter-group traffic fluctuation difference information received from the inter-group traffic fluctuation difference calculation unit 113.

  The increase pasting change unit 131 and the decrease stripping change unit 132 are based on information held in the intra-group allocation history holding unit 115 and information held in the intra-group traffic fluctuation difference holding unit 112. The intra-group schedule table held by the intra-group schedule table holding unit 126 is updated by increasing or decreasing the TS between nodes where fluctuations occur.

  The increment paste change unit 133 and the decrease strip change unit 134 determine the traffic based on the information held in the inter-group allocation history holding unit 116 and the information held in the inter-group traffic fluctuation difference holding unit 114. The TS is increased or decreased between the groups in which fluctuations occur in the group, and the inter-group schedule table held by the inter-group schedule table holding unit 127 is updated.

  The actual schedule table calculation unit 125 acquires the TS schedule by reflecting the intra-group schedule table in the inter-group schedule table, and stores the new TS schedule in the actual schedule table holding unit 124.

  The node control signal generation unit 123 converts the TS schedule held in the actual schedule table holding unit 124 into node processing corresponding to each node of the nodes 20a to 20e, and sends a scheduling result corresponding to each node via the IO15. Output.

  Next, the operation procedure of the scheduler in the TDM network system of this embodiment will be described.

  FIG. 5 is a flowchart showing the operation procedure of the scheduler of this embodiment, and FIG. 6 is a flowchart showing the processing in step 203 shown in FIG.

  When the scheduler 10 sums up the traffic of the NW (step 201), it calculates the required number of TSs (step 203). Subsequently, the scheduler 10 assigns TSs and creates a TS schedule (step 203).

  In step 203, as shown in FIG. 6, the scheduler 10 calculates the intra-group TS fluctuation difference from the calculated required TS number (step 231a), and calculates the inter-group fluctuation difference from the required TS number (step 231b). ). Subsequently, the scheduler 10 executes Rip-up & Re-allocate calculation processing on the table in the group based on the information on the TS variation difference in the group calculated in Step 231a (Step 232a). This calculation result is referred to as “calculation result CR1”.

  Further, the scheduler 10 executes Rip-up & Re-allocate calculation processing on the table between groups based on the information on the difference between groups calculated in step 231b (step 232b). This calculation result is referred to as “calculation result CR2”. The scheduler 10 acquires the TS schedule by substituting the calculation result CR1 into the calculation result CR2 (step 233).

  After that, as shown in FIG. 5, the scheduler 10 converts to node processing (step 204) and notifies the scheduling result corresponding to each of the nodes 20a to 20e to each node (step 205).

  In this embodiment, it is quickly optimized by combining hierarchization (dividing a large-scale problem into multiple small problems) and Rip-up & Re-allocate (partial change focusing only on communication with fluctuations). It makes it possible to approach the solution.

  An outline of a mechanism that can speed up scheduling according to the present embodiment will be described with reference to the drawings. FIG. 7 is a diagram for explaining a mechanism for speeding up scheduling.

  As a first step, the scale of calculation of the TS allocation problem for the target node is reduced by applying a hierarchization method. For example, the assignment problem of 100 nodes is replaced with “assignment problem of 10 nodes × 10 + 10 group assignment problem”.

  As a second stage, the Rip-up & Re-allocate method is applied to each allocation problem in the first stage. In response to an allocation problem that is becoming smaller in scale, a solution is derived at high speed by performing partial modification of the allocation so that “nodes that do not have traffic fluctuations are not subject to calculation”.

  Next, the mechanism for reducing the calculation time of the assignment problem will be described in detail.

  When a node group is defined to perform allocation by the hierarchization method disclosed in Non-Patent Document 2, and an ordered pair of node groups is called a node group pair (GP), a path in the NW belongs to any GP It will be. Let N be the number of nodes and G be the number of groups.

(1) Since the number of paths and the number of links targeted in the allocation calculation within each GP and the allocation calculation between GPs are smaller than the allocation method based on the existing method disclosed in Non-Patent Document 1, the allocation calculation Scale is reduced.
The number of paths and links targeted by the existing method in the N-node network: O (N ² ) path, the number of links and the number of links targeted by the hierarchical system that defines the G group in the O (N) link N-node network:
→ Allocation calculation within GP ... O [(N / G) ² ] path, O [(N / G)] link → Allocation calculation between GPs ... O (G ² ) path, O (G) link (2) If the Rip-up & Re-allocate method is used for the allocation calculation within the GP and the allocation calculation method between the GPs, the allocation calculation can be omitted for a path whose traffic fluctuation is below a certain value. The scale of calculation is reduced.
If the ratio that the traffic fluctuation amount exceeds the bandwidth for one TS (for example, 100 Mbps) in each path is a (0 ≦ a ≦ 1),
→ Allocation calculation in GP: The number of paths subject to Rip-up & Re-allocate calculation is multiplied by a (≦ 1).
→ Assignment calculation between GPs ... If the ratio of the path where traffic increases and the path where traffic decreases in the NW is the same, there is a high probability that traffic will not increase or decrease when viewed in GP units. In this case, Rip-up & Re-allocate is not subject to calculation.

Therefore, the required calculation time is less than the hierarchization method. As a result, the following effects can be obtained.
-The calculation time in the GP decreases in an area where the traffic fluctuation is relatively moderate.
-If there is no traffic increase / decrease in GP units, calculation time between GPs can be saved.
-The upper limit of the calculation time is equivalent to that of the hierarchization method (the calculation time takes the upper limit or a value near the upper limit when unrealistic heavy traffic fluctuation occurs).

  According to the present embodiment, it is possible to reduce the TS allocation calculation time required for NW control of the same scale in the same calculation resource, and to provide a larger scale NW system.

(Second Embodiment)
In this embodiment, the TDM network system described in the first embodiment can cope with the case where traffic is biased.

  When the hierarchized group is fixed, if the traffic is biased, an event may occur in which the total sum of traffic fluctuations for each group increases. In this case, it may take time to approach the optimal solution using the Rip-up & Re-allocate method. In this case, traffic cannot be transferred as requested, and the amount of transfer traffic is reduced.

  FIG. 8 is a diagram for explaining a problem caused by traffic bias.

  When the traffic is uniformly distributed in the NW, the scheduling method described in the first embodiment is performed as shown in the left diagram of FIG. When the traffic is biased and the traffic distribution changes, as shown in the diagram on the right side of FIG. 8, an attempt is made to eliminate the deterioration in allocation efficiency caused by large traffic fluctuations in units of groups. As a result, the amount of traffic transferred as requested in the NW system having the same equipment amount decreases.

  FIG. 9 is a diagram for explaining the operation of the TDM network system of this embodiment.

  As shown in FIG. 9, in the present embodiment, accumulation of allocation calculation errors is performed by performing dynamic group rearrangement that dynamically rearranges groups so as to equalize the sum of traffic fluctuation amounts in each group. To avoid. This dynamic group rearrangement is realized by a mechanism for observing traffic fluctuation for each group, a mechanism for detecting traffic fluctuation for each group, a mechanism for performing group recombination calculation, and a mechanism for changing settings.

  A specific configuration for realizing dynamic group rearranging in the TDM network system of this embodiment will be described. FIG. 10 is a schematic diagram illustrating the dynamic group rearrange.

  Design items for dynamic group rearrangement include: (A) where to detect bias for each group, (B) how to detect (what to monitor and how to detect bias), and (C) how to rearrange the group. There are three possible items, and various patterns can be considered depending on what is selected for each item.

(A) Where is the bias for each group detected?
1. 1. Scheduler 2. The controller of the child node closest to the parent node among the nodes belonging to the group Parent node controller (B) What to monitor?
a. Sum of traffic fluctuation amount of group → Threshold comparison b. Sum of group loss rates → threshold comparison c. Sum of group delays → threshold comparison (C) How to rearrange groups?
i. Rearrange so that the total amount of variation is uniform in each group.
ii. Rearrange so that the total delay is uniform in each group.

  Adjust the frequency of rearranging the group as an option when rearranging the group (once done, do not rearrange until a certain time has elapsed). This is because, if group rearrangement processing is frequently performed, the group in the network changes and the allocation state that does not need to be changed is also changed.

  Of the above three items (A) to (C), an example in which “a” is selected in the item (B) in each of “1 to 3” in the item (A) will be described. A detailed description of the same configuration as that described in the first embodiment is omitted.

  The present embodiment is a case where the scheduler monitors the traffic for each group and detects a large traffic fluctuation, and corresponds to a combination pattern of “1, a”.

  FIG. 11 is a block diagram showing a configuration example of the scheduler in the TDM network system of the present embodiment.

  As shown in FIG. 11, the scheduler 10 according to the present embodiment further includes a group traffic monitoring unit 152 and a group recombination notification unit 151. In FIG. 11, for the sake of explanation, a configuration is added and shown separately from the arithmetic device 11 of the scheduler 20 described in the first embodiment, but the functions of the group traffic monitoring unit 152 and the group rearrangement notification unit 151 are illustrated. You may make the arithmetic unit 11 perform. In the present embodiment, a process of rearranging the node groups is added to the process executed by the arithmetic device 11.

  The group traffic monitoring unit 152 periodically acquires traffic information between the nodes via the IO 15, and groups the threshold excess information that is information as to whether or not the inter-group traffic fluctuation amount exceeds a preset threshold. The recombination notification unit 151 is notified.

  The group reassignment notification unit 151 sends a group reassignment instruction notification for instructing group reassignment when the threshold excess information notified from the group traffic monitoring unit 152 is information indicating that the inter-group traffic fluctuation amount exceeds the threshold. It transmits to the arithmetic unit 11.

  The arithmetic unit 11 regroups the nodes 20a to 20e so that the total traffic fluctuation amount is uniform.

  In this embodiment, when the traffic between the groups fluctuates, the scheduler detects the fluctuation and regroups so that the traffic fluctuation amount becomes uniform, so that the traffic amount transferred by the same equipment decreases. It becomes possible to suppress.

  In this embodiment, the traffic of each group is monitored by the controller of the child node, and the scheduler detects a large traffic fluctuation, which corresponds to a combination pattern of “2, a”.

  FIG. 12 is a block diagram illustrating a configuration example of the TDM network system according to the present embodiment. FIG. 12A shows a configuration example of a scheduler and a node connected to the scheduler, and FIG. 12B shows a configuration example of a node closest to the node.

  The node 20a connected to the scheduler 10 is referred to as a parent node, and the controller of the parent node is referred to as a parent controller.

  Among the nodes belonging to the same group, nodes other than the parent node are referred to as child nodes, and the controller of the child node is referred to as a child controller. Here, attention is paid to the child node that is the most downstream (one hop before the parent node) with respect to the direction of the control signal toward the parent node. In this embodiment, the node 20e is the above-mentioned most downstream child node. The case will be explained. The same applies to other examples described later.

  As shown in FIG. 12A, the scheduler 10 has the same configuration as that of the first embodiment. As shown in FIG. 12B, the node 20e of this embodiment further includes a group traffic calculation / notification unit 161.

  The group traffic calculation / notification unit 161 periodically acquires traffic information between the nodes, calculates an inter-group traffic fluctuation amount from the acquired traffic information, and notifies the child controller 21b of the result. The child controller 21b transfers the inter-group traffic fluctuation amount information received from the group traffic calculation / notification unit 161 to the node 20a.

  The parent controller 21a of the node 20a transfers the inter-group traffic fluctuation amount received from the node 20e to the scheduler 10. The group traffic monitoring unit 152 of the scheduler 10 compares the inter-group traffic fluctuation amount with a preset threshold value, and notifies the group recombination notification unit 151 of threshold excess information indicating the comparison result.

  In this embodiment, since the process of calculating the inter-group traffic fluctuation amount is executed by the child node, the processing load of the scheduler is reduced compared to the first embodiment.

  In this embodiment, the parent node controller monitors the traffic for each group and detects a large traffic fluctuation, which corresponds to a combination pattern of “3, a”.

  FIG. 13 is a block diagram showing a configuration example of the TDM network system of this embodiment. FIG. 13A shows a configuration example of the scheduler and the parent node, and FIG. 13B shows a configuration example of the child node. Also in the present embodiment, description will be given on the case where the most downstream child node is the node 20e.

  As illustrated in FIG. 13A, the scheduler 10 includes a threshold excess notification monitoring unit 153 instead of the group traffic monitoring unit 152 in the configuration illustrated in FIG. Further, the node 20a further includes a group traffic calculation unit 162, a threshold comparison unit 163, and a threshold excess notification unit 164 in the configuration shown in the first embodiment.

  The group traffic calculation unit 162 acquires information on the inter-group traffic fluctuation amount from the scheduler 10 via the parent controller 21a and notifies the threshold comparison unit 163 of the information. The threshold comparison unit 163 compares a predetermined threshold with the inter-group traffic fluctuation amount and notifies the threshold excess notification unit 164 of the result.

  The threshold excess notification unit 164 notifies the parent controller 21a of threshold excess information, which is information indicating whether or not the inter-group traffic fluctuation amount has exceeded the threshold value, based on the comparison result between the inter-group traffic fluctuation amount and the threshold value. The parent controller 21 a transfers the threshold excess information to the scheduler 10.

  When the threshold excess notification information received from the node 20a is information that the inter-group traffic fluctuation amount exceeds the threshold, the threshold excess notification monitoring unit 153 of the scheduler 10 notifies the group rearrangement notification unit 151 to that effect.

  In the present embodiment, the processing load of the scheduler is reduced compared to the first embodiment because the parent node is caused to execute the calculation of the inter-group traffic fluctuation amount and the comparison processing between the traffic fluctuation amount and the threshold value.

  The present embodiment is a case where excessively frequent group changes are suppressed, and corresponds to another pattern of combinations of “1, a”.

  FIG. 14 is a block diagram showing a configuration example of a scheduler in the TDM network system of the present embodiment, and FIG. 15 is a table for explaining various signals shown in FIG.

  As illustrated in FIG. 14, the scheduler 10 according to the present embodiment further includes a reassignment timer 155 and a reassignment permission notification unit 154 in the configuration described in the first embodiment.

  The group traffic monitoring unit 152 obtains traffic information between the nodes via the IO 15, and transmits group excess notification information indicating whether the inter-group traffic fluctuation amount exceeds a preset threshold. 151 is notified. The group reassignment notification unit 151 sends a group reassignment instruction notification for instructing group reassignment when the threshold excess information notified from the group traffic monitoring unit 152 is information indicating that the inter-group traffic fluctuation amount exceeds the threshold. It transmits to the arithmetic unit 11.

  In addition, the group recombination notification unit 151 transmits a group reassignment instruction transmission notification for notifying that the group reconfiguration instruction notification is instructed when the group reconfiguration instruction notification is transmitted to the arithmetic device 11. Send to. The reassignment permission timer 155 activates the timer after receiving the group reassignment instruction transmission notification, and if it does not receive the next group reassignment instruction transmission notification within a predetermined time, the reassignment permission notification unit 154 is notified. When receiving the notification of timeout from the reassignment permission timer, the reassignment permission notification unit 154 permits the group reassignment notification unit 151 to notify the arithmetic device 11 of a group reassignment instruction.

  In the present embodiment, unnecessary allocation state changes due to excessively frequent rearrangement of groups can be suppressed by performing management using a timer.

  According to this embodiment, it is possible to suppress the allocation efficiency deterioration caused by the traffic fluctuation of the group unit and increase the accommodated traffic amount in the same NW facility amount. Therefore, it is possible to dynamically control a large-scale NW at a higher speed than when dynamic scheduling (bandwidth allocation) is performed according to traffic fluctuations according to the conventional technology.

(Third embodiment)
Next, a case where the scheduling method by the TDM network system described in the first and second embodiments is applied to a multistage ring will be described.

  FIG. 16 is a block diagram showing a configuration example of the TDM network system of the present embodiment. FIG. 16 is a diagram showing an example of a multistage ring network, and shows a case where the number of stages is two.

  Five nodes 20 including nodes A to E are connected to the upper ring in a ring shape. In the lower ring A, three nodes 20 of a node A, a node F, and a node G are connected in a ring shape. In the lower ring B, four nodes 20 of a node B, a node H, a node I, and a node J are connected in a ring shape. In the lower ring C, three nodes 20 of a node C, a node K, and a node M are connected in a ring shape.

  Note that the configuration of the node 20 and the scheduler 10 is the same as the configuration described in the first embodiment, and thus detailed description thereof is omitted in this embodiment.

  The operation of the TDM network system of this embodiment will be described.

  FIG. 17 is a diagram for explaining the operation of the TDM network system of this embodiment. FIG. 17 shows that node D and node E are grouped into group Z, lower ring A including node A is grouped into group RA, lower ring B including node B is grouped into group RB, and node C is The case where the lower ring C to be included is grouped into the group RC is shown.

  First, scheduling between upper and lower rings is performed by scheduling between groups. The upper diagram in FIG. 17 shows TS allocation for data transfer from the group RA → RB → RC → Z. Subsequently, scheduling in each ring is performed. In the upper diagram of FIG. 17, TS allocation for data transfer in the lower ring B is shown.

  After that, dynamic change by Rip-up & Re-allocate is performed at each layer. In the lower diagram of FIG. 17, since there is no change from group RB to group Z, it is excluded from calculation, and the lower ring B is dynamically changed by Rip-up & Re-allocate. This indicates that 1TS decrease of “→ node E” and 1 TS increase of “node I → node D” are performed.

  In this way, even in a multistage ring, after defining a hierarchy consisting of scheduling between each ring and scheduling within each ring, dynamic change by Rip-up & Re-allocate can be performed in each hierarchy. It becomes possible.

  In the present embodiment, the case where scheduling is performed in the order of “between groups” → “within group” has been described, but the order may be “within group” → “between groups”.

  In the above embodiment, the case where the hierarchization method and the Rip-up & Re-allocate method are used has been described. However, these methods are examples of the TDM scheduling method constituting the present invention. It is not limited to one method.

  Furthermore, the characteristics of the scheduling method based on the combination of the hierarchization method and the Rip-up & Re-allocate method will be described.

The first characteristic is that the calculation of each of the group layer and the node layer can be performed independently. As a specific example,
-The period may be different. For example, for the node layer T, the group layer may be kT (k is a natural number).
The calculation method may be changed (for example, the heuristic method disclosed in Non-Patent Document 1 or the Rip-up & Re-allocate method).
-An arithmetic unit dedicated to each calculation may be provided in the scheduler (a plurality of arithmetic units may be used).
Variations such as are conceivable.

The second characteristic is that each node layer (within a group) can be calculated independently. As a specific example,
-The period may be different. For example, the period may be independently mT (m is a natural number that is a divisor of k).
The calculation method may be changed (for example, the heuristic method disclosed in Non-Patent Document 1 or the Rip-up & Re-allocate method).
-An arithmetic unit dedicated to each calculation may be provided in the scheduler (a plurality of arithmetic units may be used).
Variations such as are conceivable.

10 scheduler 11 computing device 13 memory 20a-20e node

Claims (3)

A network including a plurality of nodes performing data transfer to other adjacent nodes in units of time slots according to a time slot schedule;
A scheduler that is connected to any one of the plurality of nodes, aggregates traffic of the network, and creates the time slot schedule based on the aggregated traffic;
The scheduler
Calculate the required number of time slots from the aggregated traffic, group the plurality of nodes, calculate the traffic fluctuation difference of each group, and perform scheduling within the group and scheduling between groups based on the difference information A TDM network system that performs each and applies a scheduling result within a group to a scheduling result between groups to determine the time slot schedule of the entire system. The TDM network system according to claim 1, wherein
The scheduler is a TDM network system that groups the plurality of nodes so as to uniformize the total traffic fluctuation amount. A scheduling method by a scheduler that provides the time slot schedule to a network including a plurality of nodes that perform data transfer to other adjacent nodes in time slot units according to the time slot schedule,
Calculate the required number of time slots from the traffic aggregated in the network,
Grouping the plurality of nodes, calculating a traffic fluctuation difference of each group,
Based on the calculated difference information, scheduling within the group and scheduling between the groups, respectively,
A scheduling method for determining the time slot schedule of the entire system by applying a scheduling result in the group to a scheduling result between groups.