Processing apparatus, processing method, and program

The apparatus and method optimize Clos network design by considering non-uniform terminal utilization rates, enhancing network capacity and reducing blockages through strategic switch and link configurations.

JP2026182006APending Publication Date: 2026-11-19NIPPON TELEGRAPH & TELEPHONE CORP
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2025077885
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2025-05-08
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

Existing network design methods for Clos networks assume uniform terminal utilization rates, failing to account for varying utilization rates among terminals, leading to inefficiencies and potential blockages.

Method used

A processing apparatus and method that acquires parameter data including terminal utilization rates and allowable blockage rates to design a network that maximizes capacity while maintaining a predetermined blockage rate, even with non-uniform terminal utilization, by optimizing the number of switches and links.

Benefits of technology

Enables efficient network design with controlled blockage rates, ensuring network capacity is maximized and blockages are minimized even when terminal utilization varies.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2026182006000001_ABST
    Figure 2026182006000001_ABST
Patent Text Reader

Abstract

In situations where the utilization rate differs from terminal to terminal, the network is designed under conditions that the congestion rate remains below a predetermined value. [Solution] The processing device 1 includes a calculation unit 22 that outputs network design data 12 such that, in a network satisfying the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, and the utilization rate pu for each terminal u, the sum of the probabilities of each probability of blockage occurring does not exceed the allowable blockage rate ε when the number of terminals using each input / output switch increases by 1, which is greater than the maximum value of the number of terminals using each input / output switch when the network is not blocked, and the product of the number of terminals n connected to the input and output of each input / output switch and the number of input / output switches k is maximized.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

[0001] This disclosure relates to an apparatus, a processing method, and a program. [Background technology]

[0002] I have design skills for indoor networks such as data centers or telecommunications buildings. Data centers house many racks. Designated rack pairs are connected. Telecommunications buildings house transmission equipment to other buildings. These transmission devices are connected.

[0003] Currently, packet switching technologies such as IP (Internet Protocol) or Ethernet are used. In the future, circuit switching technologies such as optical communication are being considered to improve transmission capacity and power efficiency.

[0004] In circuit switching, if no path exists between the pairs requesting a connection, a circuit is not established between them. This is called a block.

[0005] There is a circuit-switched network structure called TF-Clos (Twisted and Folded Clos Network) (see Non-Patent Document 1). Non-Patent Document 1 discloses a method for designing a network so that network capacity is maximized without blockages occurring.

[0006] Patent Document 1 provides a theoretical guarantee of the occupancy rate in a Clos network. Patent Document 1 utilizes the fact that the utilization rate of all terminals is uniform and defines the probability of a connection request occurring at each inbound / outbound switch using a binomial distribution. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Japanese Patent Publication No. 2024-126364 [Non-patent literature]

[0008] [Non-Patent Document 1] T. Mano, T. Inoue, K. Mizutani, and O. Akashi, “Redesigning the nonblocking Clos network to increase its capacity,” IEEE Transactions on Network and Service Management, pp. 1-1, 2022, [Overview of the Initiative] [Problems that the invention aims to solve]

[0009] Patent Document 1 assumes that the utilization rate of all terminals is uniform, but in an actual Clos network, it is not necessarily uniform and can take different values ​​for each terminal.

[0010] This disclosure is made in view of the above circumstances, and the purpose of this disclosure is to provide a technology that enables the design of a network under conditions where the congestion rate is below a predetermined value, even when the utilization rate differs for each terminal. [Means for solving the problem]

[0011] A processing apparatus according to one aspect of the present disclosure includes an acquisition unit that acquires parameter data including the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, the utilization rate pu for each terminal u, and the allowable blockage rate ε; and a calculation unit that outputs network design data including the number of terminals n connected to the input and output of each input / output switch, the number of input / output switches k, the number of intermediate switches m, and the number of links from one input / output switch to one intermediate switch and the number of links from one intermediate switch to one input switch v, such that in a network satisfying the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, and the utilization rate pu for each terminal u, the sum of the probabilities of blockage occurring when the number of terminals using each input / output switch increases by 1 from the maximum value of the number of terminals using each input / output switch in a state where the network is unblocked, the product of the number of terminals n connected to the input and output of each input / output switch and the number of input / output switches k is maximized, and the probability of blockage occurring is the probability of blockage occurring in the probability density of the sum of the utilization rates pu for each terminal u using each input / output switch.

[0012] A processing method according to an aspect of the present disclosure causes a computer to acquire parameter data including respective parameters of the number N of input terminals and output terminals of each input / output switch, the total number a of input / output switches and intermediate switches, the usage rate pu of each terminal u, and the allowable blocking rate ε, and in a network that satisfies the number N of input terminals and output terminals of each input / output switch, the total number a of input / output switches and intermediate switches, and the usage rate pu of each terminal u, when the number of terminals using each input / output switch increases by 1 from the maximum value of the number of terminals using each input / output switch in a state where the network is non-blocking, the total probability of each probability that blocking occurs does not exceed the allowable blocking rate ε, and network design data including the number n of terminals connected to each of the input and output of each input / output switch, the number k of input / output switches, the number m of intermediate switches, and the number v of links from one input / output switch to one intermediate switch and from one intermediate switch to one input switch is output such that the product of the number n of terminals connected to each of the input and output of each input / output switch and the number k of input / output switches is maximized, wherein the probability that blocking occurs is a probability that blocking occurs in the probability density of the sum of the usage rates pu of each terminal u that uses each input / output switch.

[0013] A program according to an aspect of the present disclosure causes a computer to function as an acquisition unit that acquires parameter data including respective parameters of the number N of input terminals and output terminals of each input / output switch, the total number a of input / output switches and intermediate switches, the usage rate pu of each terminal u, and the allowable blocking rate ε, and in a network that satisfies the number N of input terminals and output terminals of each input / output switch, the total number a of input / output switches and intermediate switches, and the usage rate pu of each terminal u, when the number of terminals using each input / output switch increases by 1 from the maximum value of the number of terminals using each input / output switch in a state where the network is non-blocking, the total probability of each probability that blocking occurs does not exceed the allowable blocking rate ε, to function as a calculation unit that outputs network design data including the number n of terminals connected to each of the input and output of each input-output switch, the number k of input-output switches, the number m of intermediate switches, and the number v of links from one input-output switch to one intermediate switch and also from one intermediate switch to one input switch, such that the product of the number n of terminals connected to each of the input and output of each input-output switch and the number k of input-output switches is maximized, wherein the probability that blocking occurs is the probability that blocking occurs in the probability density of the sum of the utilization rates pu of each terminal u that uses each input-output switch. [Advantageous Effects of Invention]

[0014] According to the present disclosure, it is possible to provide a technology capable of designing a network under the condition that the blocking rate is equal to or less than a predetermined value in a situation where the utilization rate differs for each terminal. [Brief Description of Drawings]

[0015] [Figure 1] Figure 1 is a diagram illustrating functional blocks of a processing apparatus of the present disclosure. [Figure 2] Figure 2 is a diagram illustrating a TF-Clos network structure. [Figure 3] Figure 3 is a flowchart illustrating processing of the processing apparatus. [Figure 4] Figure 4 is a diagram illustrating an example of a network having a TF-Clos network structure according to a second modification. [Figure 5] Figure 5 is a diagram illustrating a Clos network structure. [Figure 6] Figure 6 is a diagram illustrating a hardware configuration of a computer used for the processing apparatus. [Mode for Carrying Out the Invention]

[0016] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. In the description of the drawings, the same reference numerals are assigned to the same parts, and descriptions thereof are omitted.

[0017] (Processing Apparatus) The processing apparatus 1 shown in Figure 1 designs a network under the condition that the blockage rate is below a predetermined value. In this disclosure, the processing apparatus 1 is described in the case of a TF-Clos network, but it may also be applied to a Clos network, as will be described with reference to the modified examples.

[0018] As shown in Figure 2, the TF-Clos network comprises multiple input-output switches in the input-output layer and multiple intermediate switches in the intermediate layer. In the example shown in Figure 2, there are k input-output switches and m intermediate switches. The switches are named S1, ..., S1, ..., S1, ..., S1, ..., S1, ..., S1, ..., S1, ..., S2 k+m An identifier is given as shown.

[0019] In a TF-Clos network, each switch has an input switch face and an output switch face. Inputs from input terminals or intermediate switches are connected to the input switch face of the input / output switch. Outputs to output terminals or intermediate switches are connected to the output switch face. In the example shown in Figure 2, each switch is an N×N switch with 12 ports on each face.

[0020] The input / output switches and intermediate switches are each connected by v links. v links are provided from one input / output switch to one intermediate switch. v links are also provided from one intermediate switch to one input / output switch.

[0021] Multiple terminals are connected to each input / output switch. One input / output switch is connected to n input terminals t and n output terminals r.

[0022] In the network shown in Figure 2, terminals are connected to the input / output switches. When two terminals attempt to communicate, they can communicate if they can be connected using only the available links between the switches. An available link is a link through which no signal is passing.

[0023] Two terminals t1 1 and r1 k attempt to communicate. A path is prepared between t1 1 and r1 k using only idle links. As shown by the solid line in Figure 2, there is an idle link between S1 and S k+1 , there is an idle link, and between S k+1 and S k , there is also an idle link. Furthermore, the internal configurations of switches S k+1 and S k are modified to connect these links. As a result, a path consisting only of idle links is established up to terminal t1 1 and r1 k . Here, when terminal t1 1 transmits a signal, the signal reaches terminal r1 k .

[0024] When communication between terminals starts, the links on the path become in use and are no longer idle. An in-use link will not be used by other communications. As the number of idle links between switches decreases, a state where communication cannot be performed because there is no longer any path consisting of idle links even when attempting to communicate between terminals is referred to as "blocking".

[0025] Patent Document 1 discloses a method for designing a TF-Clos network shown in Figure 2 such that the network capacity is maximized under the condition that the blocking probability is equal to or less than a predetermined value ε. Assume that there are a N×N switches each having N ports on the input plane and N ports on the output plane. The utilization rate p of each terminal is uniform.

[0026] For this problem, Patent Document 1 utilizes the fact that the utilization rates of the terminals are uniform, and defines the probability that a connection request occurs at each ingress-egress switch using a binomial distribution. Patent Document 1 formulates this as a mathematical programming problem according to Equation (1), and designs the TF-Clos network to maximize the capacity, i.e., design a network with maximum network capacity. Specifically, the design of the network determines the number of ingress-egress switches, the number of intermediate switches, and the number of inter-switch links.

[0027] [Formula]

[0028] Here, N is the set of integers of a certain sex. N, a, p, and ε are given as problems. n, k, m, v, and n sub This is the decision variable for the optimization problem.

[0029] Equation (1a) is the objective function, which shows that the network capacity is maximized. Equation (1b) is a condition regarding the number of ports on the input / output switches. Equation (1c) is a condition regarding the number of ports on the intermediate switches. Equation (1f) is a condition regarding the number of switches.

[0030] From equation (1e), the number of terminals used is n. sub The following conditions guarantee that no blockage will occur. Here, SNB stands for strictly nonblocking, meaning that no blockage will occur. Equation (1d) indicates that the terminal used is n sub It is guaranteed that the sum of the occlusion rates in the larger cases does not exceed ε. Equation (1d) utilizes the binomial distribution. Equation (1g) is the domain of each variable.

[0031] The admission control described in Patent Document 1 will also be explained. The admission control in Patent Document 1 is given as the solution to equations (1a) to (1g) n sub The number of connected terminals on either the input or output switch of the connection request is n. sub The following conditions apply: The connection is made using one of the m intermediate switches. The number of connected terminals on either the input or output switch is n. sub If the value exceeds this limit, the connection request will be rejected.

[0032] Patent Document 1 assumes that the utilization rate of each terminal is uniform, but in an actual TF-Clos network, this is not necessarily the case. Each terminal may have a different utilization rate. Patent Document 1 cannot handle cases where the utilization rates of each terminal differ.

[0033] Therefore, the processing device 1 of this disclosure designs a network that maximizes network capacity while allowing a predetermined blockage rate in a TF-Clos network where the utilization rate differs for each terminal.

[0034] As shown in Figure 1, the processing device 1 according to this disclosure includes parameter data 11, network design data 12, and operating condition data 13, as well as the functions of an acquisition unit 21 and a calculation unit 22. Each piece of data is stored in a storage device such as a memory 902 or storage 903. Each function is implemented in a CPU 901.

[0035] Parameter data 11 consists of parameters given in advance for designing the network. Parameter data 11 includes the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, the utilization rate pu for each terminal u, and the allowable blockage rate ε.

[0036] Here, the congestion rate refers to the congestion rate experienced by each terminal, not the congestion rate observed for the entire network. The congestion rate is the percentage of times when communication is impossible because no available links are available to take advantage of an attempt to communicate between terminals. Parameter data 11 includes the allowable congestion rate ε for the network being designed.

[0037] Terminal u utilization rate pu is the percentage of time that terminal u is connected to the input / output switch in an attempt to communicate with other terminals. Terminal u utilization rate pu is the percentage of the links between terminal u and the input / output switch that are in use. Terminal u utilization rate pu is a known value estimated based on past service performance, etc.

[0038] The network design data 12 is data that identifies the network calculated by the calculation unit 22. The network design data 12 includes four parameters: (i) n, the number of terminals connected to the input and output of each input / output switch; (ii) k, the number of input / output switches; (iii) m, the number of intermediate switches; and (iv) v, the number of links from one input / output switch to one intermediate switch and from one intermediate switch to one input switch.

[0039] The operational condition data 13 includes the maximum number of terminals connected to the input and output of each inlet / outlet switch in the network specified by the network design data 12, under conditions where the allowable occlusion rate ε is not exceeded. Specifically, the operational condition data 13 is the maximum number of terminals connected to the inlet / outlet switch when the total probability of not exceeding the allowable occlusion rate ε is calculated in equation (5d) described below. This maximum number of terminals is the n when equation (5a) is maximized. sub It can also be expressed as follows.

[0040] The acquisition unit 21 acquires parameter data 11. Each parameter is specified, for example, by a user.

[0041] The calculation unit 22 designs a network that maximizes network capacity under the conditions specified by the parameter data 11, and outputs network design data 12 and operational condition data 13. Network capacity is calculated by multiplying the number of terminals n connected to the input and output of each inlet / outlet switch by the number of inlet / outlet switches k.

[0042] The calculation unit 22 designs the network such that, in a network satisfying the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, and the utilization rate pu for each terminal u, the sum of the probabilities of each probability of blockage occurring does not exceed the allowable blockage rate ε when the number of terminals using each input / output switch increases by 1, compared to the maximum number of terminals using each input / output switch when the network is not blocked, and the product of the number of terminals n connected to the input and output of each input / output switch and the number of input / output switches k is maximized.

[0043] The calculation unit 22 outputs network design data 12 as a network design, which includes the number of terminals n connected to the input and output of each input / output switch, the number of input / output switches k, the number of intermediate switches m, and the number of links v from one input / output switch to one intermediate switch and from one intermediate switch to one input switch.

[0044] The calculation unit 22 further outputs operational condition data 13, which includes the maximum number of terminals connected to the in / out switch when calculating the total probability that does not exceed the allowable blockage rate ε.

[0045] Here, we will explain the calculation method for this disclosure.

[0046] The network has k input / output switches, each with n terminals, so the network as a whole has nk terminals. Here, we count Tx and Rx in pairs. Let pu be the utilization rate of terminal u∈[1,nk]. The probability that terminal u is "not" in use is 1-pu. To express these uniformly, we use the following function h u We will implement this.

[0047]

number

[0048] function h u is, x u Specifying =1 will show the utilization rate of terminal u, x u Specifying =0 represents the "probability of not using". This h u Using this, we modify equation (1d) so that it can handle the utilization rate for each terminal.

[0049] The probability density H is such that exactly w ≤ n terminals are used at the i-th input / output switch. i n (w) is defined as in equation (3). w is the number of requests from terminals to a single input / output switch, and is the number of terminals to which the path is connected.

[0050]

number

[0051] Using equation (3), replace equation (1d) with equation (4).

[0052]

number

[0053] From these, equations (5a) to (5g) show that in a TF-Clos network where the utilization rate differs for each terminal, it is possible to unpack the network while allowing a predetermined blockage rate and maximizing the network capacity.

[0054]

number

[0055] The method for solving this formula is arbitrary, but if there are around 100 switches, you can exhaustively search for the decision variables (n, n sub Even by examining all possible combinations of values ​​for k, m, and v, a solution can be obtained in practical time.

[0056] (Processing method) The processing method of this disclosure will be explained with reference to Figure 3.

[0057] In step S1, the processing unit 1 acquires parameter data 11. The parameter data 11 includes each parameter necessary for the processing unit 1 to design the network.

[0058] In step S2, the processing unit 1 calculates a network that maximizes network capacity by referring to parameter data 11. In step S3, it outputs network design data 12 that identifies the network calculated in step S2, and operational condition data 13 for maintaining an allowable blockage rate ε in the network.

[0059] (First variation) In this disclosure, the process for designing a network when the network is a TF-Clos network has been described, but the processing method described herein can also be applied to general Clos networks.

[0060] Refer to Figure 5 to explain a typical Clos network. Like a TF-Clos network, a Clos network has input / output switches in the input / output layer and intermediate switches in the intermediate layer, but the configuration of the input / output switches is different. In a Clos network, one side of the input / output switch is connected to the terminal, and the other side is connected to the intermediate switch.

[0061] The calculation unit 22, when the network is a Clos network, replaces equation (5b) with equation (5b') to calculate the network design data 12 and the operational condition data 13.

[0062]

number

[0063] This allows for the construction of a larger network than an unblocked network, even in a Clos network, by defining the permissible blockage rate ε.

[0064] (Second variation) The TF-Clos network according to this disclosure preferably includes a control unit D, as shown in Figure 4. The control unit D controls the sum of the utilization rates of each terminal using the access switches so that the gap between each access switch is less than or equal to a predetermined value. The gap is, for example, the standard deviation. The control unit D calculates, for example, the sum of the utilization rates of each terminal using each access switch. The control unit D controls the access switches used by each terminal so that the standard deviation of the sum calculated for each access switch is lower than a predetermined value.

[0065] The control unit D controls each terminal, arranged in a predetermined order of utilization rate, to use the input / output switches, which are also arranged in a predetermined order. The predetermined order of utilization rate is, for example, ascending or descending. The predetermined order of input / output switches is, for example, by identifier. By sequentially assigning each terminal, arranged in ascending order of utilization rate, to the input / output switches, which are arranged in ascending order of identifier, the difference in the sum of utilization rates of each terminal using each input / output switch is reduced.

[0066] The control unit D is, for example, a patch panel. The patch panel is installed between the terminal and the input / output switch. The control unit D controls the connection between the terminal and the input / output switch according to instructions from the processing unit 1 or the like.

[0067] This wiring eliminates the need to change the connection between each terminal and the input / output switch, as long as the utilization rate (pi) of each terminal does not change significantly. Since rewiring is not required every time a connection request occurs at each terminal, the cost of maintaining the network can be reduced.

[0068] As the difference in the sum of the utilization rates of each terminal using each input / output switch decreases, it becomes easier to satisfy the constraint condition, equation (5d). Making it easier to satisfy the constraint may lead to finding a better optimal solution, and in some cases, the network capacity can be increased. On the other hand, if there is an input / output switch i that makes the left side of equation (5d) large, it becomes more difficult to satisfy the constraint condition, equation (5d).

[0069] To make it easier for each input / output switch i to satisfy the constraint equation (5d), it is sufficient to ensure that the probability density of the sum of the utilization rates of any one of the input / output switches i does not become extremely large. Therefore, the probability density H of the total utilization rate of the terminal for any input / output switch is i n Change the correspondence between terminals and input / output switches so that (n) is evenly distributed.

[0070] Although the explanation describes the case where the control unit D is installed in the TF-Clos network, the control unit D may also be installed in the Clos network shown in Figure 5.

[0071] The processing device 1 of this disclosure enables optimal network design even when the utilization rate differs for each terminal, in a network design that theoretically guarantees the occlusion rate. Patent Document 1 described the constraints using a binomial distribution under the assumption that the utilization rate of each terminal is equal. In contrast, the processing device 1 of this disclosure defines the utilization rate for each terminal and describes the constraints accordingly.

[0072] Furthermore, the network designed for the processing unit 1 of this disclosure ensures that the total utilization rate of the terminals accommodated (wired) by each inlet / outlet switch is equal. This disclosure makes it easier to satisfy constraints and obtain a better optimal solution.

[0073] The processing unit 1 described above in this disclosure is, for example, a general-purpose computer system comprising a CPU (Central Processing Unit, processor) 901, memory 902, storage 903 (HDD: Hard Disk Drive, SSD: Solid State Drive), communication device 904, input device 905, and output device 906. In this computer system, each function of the processing unit 1 is realized by the CPU 901 executing a program loaded onto the memory 902.

[0074] The processing unit 1 may be implemented on a single computer, or on multiple computers. Furthermore, the processing unit 1 may be a virtual machine implemented on a computer.

[0075] The program of the processing unit 1 can be stored on a computer-readable recording medium such as an HDD, SSD, USB (Universal Serial Bus) memory, CD (Compact Disc), or DVD (Digital Versatile Disc), or it can be distributed over a network. A computer-readable recording medium is, for example, a non-transitory recording medium.

[0076] The functionality of each element of this disclosure may be implemented using circuits or processing circuits that include general-purpose processors, special-purpose processors, integrated circuits, ASICs (Application Specific Integrated Circuits), FPGAs (Field Programmable Gate Arrays), general circuits, and / or combinations thereof that are programmed using one or more programs stored in one or more memories, or otherwise configured to perform the disclosed functionality. A processor is considered a circuit or processing circuit because it includes transistors and other circuits. A processor may be a programmed processor that executes programs stored in memory. In this disclosure, a circuit, unit, or means is hardware that performs the enumerated functionality, or hardware programmed to perform the enumerated functionality. Hardware may be any hardware of this disclosure that is programmed or configured to perform the enumerated functionality.

[0077] Memory stores computer programs, including computer instructions. These computer instructions provide logic and routines that enable hardware (e.g., circuits or processing circuits) to perform the methods of this disclosure. This computer program can be implemented in common forms, such as computer-readable storage media, computer program products, memory devices, recording media such as CD-ROMs and DVDs, and / or memory of FPGAs or ASICs.

[0078] This disclosure is not limited to the embodiments described above, and numerous modifications are possible within the scope of its essence. [Explanation of symbols]

[0079] 1 Processing Unit 11 Parameter data 12 Network Design Data 13. Operating Conditions Data 21 Acquisition Department 22 Calculation Section 901 CPU 902 memory 903 Storage 904 Communication equipment 905 Input device 906 Output device

Claims

1. An acquisition unit acquires parameter data including the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, the utilization rate pu for each terminal u, and the allowable occlusion rate ε. In a network where the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, and the utilization rate pu for each terminal u are satisfied, When the number of terminals using each input / output switch increases by one, beyond the maximum number of terminals using each input / output switch in a state where the network is not blocked, the sum of the probabilities of each probability of blockage occurring does not exceed the allowable blockage rate ε. The product of the number of terminals n connected to the input and output of each input / output switch and the number of input / output switches k is maximized. A calculation unit outputs network design data including the number of terminals n connected to the inputs and outputs of each input / output switch, the number of input / output switches k, the number of intermediate switches m, and the number of links v from one input / output switch to one intermediate switch and from one intermediate switch to one input switch. Equipped with, The probability of an obstruction occurring is, The probability density of the sum of the utilization rates pu of each terminal u that uses each input / output switch is the probability of blockage occurring. Processing device.

2. The calculation unit outputs the network design data such that, when the network is a TF-Clos network, it satisfies equations (5b) to (5g) and maximizes equation (5a). The apparatus according to claim 1. [Math 1]

3. The calculation unit, when the network is a Clos network, replaces equation (5b) with equation (5b') to calculate the network design data. The apparatus according to claim 2. [Math 2]

4. The calculation unit further outputs operational condition data including the maximum number of terminals connected to the input / output switch when the total probability that does not exceed the allowable blockage rate ε is calculated. The apparatus according to claim 1.

5. The network includes a control unit that controls the total utilization rate of each terminal using the entry / exit switches so that the distance between each entry / exit switch is less than or equal to a predetermined value. The apparatus according to claim 1.

6. The control unit controls each terminal, which is arranged in a predetermined order of utilization rate, to use the input / output switches, which are also arranged in a predetermined order. The apparatus according to claim 5.

7. Computers Parameter data is obtained that includes the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, the utilization rate pu for each terminal u, and the allowable blockage rate ε. In a network where the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, and the utilization rate pu for each terminal u are satisfied, When the number of terminals using each input / output switch increases by one, beyond the maximum number of terminals using each input / output switch in a state where the network is not blocked, the sum of the probabilities of each probability of blockage occurring does not exceed the allowable blockage rate ε. The product of the number of terminals n connected to the input and output of each input / output switch and the number of input / output switches k is maximized. The system outputs network design data including the number of terminals n connected to the inputs and outputs of each input / output switch, the number of input / output switches k, the number of intermediate switches m, and the number of links v from one input / output switch to one intermediate switch and from one intermediate switch to one input switch. The probability of an obstruction occurring is, The probability density of the sum of the utilization rates pu of each terminal u that uses each input / output switch is the probability of blockage occurring. Processing method.

8. Computers, An acquisition unit acquires parameter data including the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, the utilization rate pu for each terminal u, and the allowable occlusion rate ε. In a network where the number of input and output terminals N for each input / output switch, the total number a of input / output switches and intermediate switches, and the utilization rate pu for each terminal u are satisfied, When the number of terminals using each input / output switch increases by one, beyond the maximum number of terminals using each input / output switch in a state where the network is not blocked, the sum of the probabilities of each probability of blockage occurring does not exceed the allowable blockage rate ε. The product of the number of terminals n connected to the input and output of each input / output switch and the number of input / output switches k is maximized. It functions as a calculation unit that outputs network design data including the number of terminals n connected to the inputs and outputs of each input / output switch, the number of input / output switches k, the number of intermediate switches m, and the number of links v from one input / output switch to one intermediate switch and from one intermediate switch to one input switch. The probability of an obstruction occurring is, The probability density of the sum of the utilization rates pu of each terminal u that uses each input / output switch is the probability of blockage occurring. program.

Citation Information

Patent Citations

  • Processing device, processing method, and program

    JP2024126364A