Path control device, path control method and program
The path control device optimizes frequency allocation in optical communication networks by deriving route and feature vectors using convolution processing, addressing efficiency and frequency blocking challenges, thereby enhancing network capacity and load resistance.
Patent Information
- Application Number
- JP2023563463
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-11-26
- Publication Date
- 2024-12-04
- Estimated Expiration
- 2041-11-26
AI Technical Summary
Existing dynamic route frequency allocation algorithms in optical communication networks face challenges in efficiently accommodating large communication capacity while minimizing frequency blocking, with approximate solutions being difficult to derive within a realistic time frame.
A path control device and method that derive route vectors and feature vectors based on network topology, using convolution processing to allocate frequencies efficiently, optimizing route and frequency combinations through feasibility and information distance vectors.
Improves the efficiency of accommodating communications by quickly deriving approximate solutions closer to exact solutions, enhancing network resource utilization and resistance to communication loads.
Smart Images

Figure 0007594223000010 
Figure 0007594223000011 
Figure 0007594223000012
Abstract
Description
[Technical field]
[0001] The present invention relates to a path control device, a path control method, and a program. [Background technology]
[0002] A user terminal may request a communication connection to an optical communication network from a request management device. When a user terminal requests a communication connection to an optical communication network, the request management device instructs a path control device to execute dynamic routing and spectrum allocation (RSA).
[0003] Edges in the physical topology of an optical communication network are realized using optical fibers. Here, there are limitations to the frequencies of optical signals that an optical fiber can accommodate. In other words, the same frequency (wavelength) cannot be assigned to the same optical fiber. Therefore, the path control device assigns (sets and deletes) the optical signal route and frequency band of each user terminal in response to a connection request, so that the assignment of the frequency band of the optical signal does not collide on the route (path) of the optical signal. Hereinafter, the act of preventing the assignment of the same frequency (wavelength) to the same optical fiber is referred to as "frequency blocking."
[0004] Problems related to dynamic path frequency allocation algorithms include the path assignment problem and the frequency assignment problem. The k-shortest-paths method is sometimes applied to the path assignment problem. For this reason, among dynamic path frequency allocation algorithms, algorithms for the frequency assignment problem are mainly devised.
[0005] There are two types of frequency allocation problems: a fixed grid method that allocates frequencies with center frequencies at fixed intervals, and a flexible grid method that allocates frequencies without limiting the bandwidth or center frequency intervals. The fixed grid method is called RWA (Routing and Wavelength Assignment). The flexible grid method is called RSA (Routing and Spectrum Allocation).
[0006] Known algorithms for fixed-grid frequency allocation include the "First-Fit (FF) method," the "Least Used (LU) method," the "Random method," and the "Most Used (MU) method" (see Non-Patent Document 1). Of these, the "First-Fit (FF) method" is known as a simple and excellent algorithm. Also, as an algorithm for fixed-grid frequency allocation, the "ksp-FF" method, which is a combination of the path allocation method "k-shortest-paths (ksp) method" and the frequency allocation method, is known as an excellent algorithm.
[0007] Also, as algorithms for frequency allocation in the flexible grid system, the "First-Fit (FF) method", the "Exact-Fit (EF) method", and the "First-Last-Fit (FLF) method" are known (Non-Patent Documents 2 and 3). [Prior art documents] [Non-patent literature]
[0008] [Non-Patent Document 1] Hui Zang, et al., “A Review of Routing and Wavelength Assignment Approaches for Wavelength-Routed Optical WDM Networks”, 2000, Optical Networks Magazine. [Non-Patent Document 2] Chatterjee, et al., “Routing and Spectrum Allocation in Elastic Optical Networks: A Tutorial”, IEEE Communication surveys & tutorials, vol., 17, no. 3 (2015) [Non-Patent Document 3] Yuan, et al., “A Spectrum Assignment Algorithm in Elastic Optical Network With Minimum Sum of Weighted Resource Reductions in all Associated Paths“, Journal of Lightwave Technology, vol. 37, no. 21 (2019) Summary of the Invention [Problem to be solved by the invention]
[0009] In flexible grid frequency allocation (RSA), the objective function of the algorithm is to accommodate a large communication capacity while keeping the probability of frequency blocking low. Therefore, the challenge in flexible grid frequency allocation is to achieve the above objective more efficiently.
[0010] It is difficult to derive a strictly optimal solution (exact solution) for frequency allocation within a realistic time frame. For this reason, an approximate solution for frequency allocation is derived. For this reason, an algorithm capable of quickly deriving an approximate solution closer to the exact solution is required. Thus, there is a problem in that it is not possible to improve the accommodation efficiency of communication requested for connection.
[0011] In view of the above circumstances, an object of the present invention is to provide a path control device, a path control method, and a program that can improve the efficiency of accommodating communications requested for connection. [Means for solving the problem]
[0012] One aspect of the present invention is a path control device that includes a route derivation unit that derives multiple route vectors representing combinations of multiple edges connecting from a start node to an end node based on physical topology information of a network and a start node and an end node of the network, a feature derivation unit that derives a route bandwidth matrix for each route vector including the route vectors replicated according to a required bandwidth, and derives feature vectors for the multiple frequencies and feature vectors for the multiple edges based on a route frequency matrix and the route bandwidth matrix that represent the allocation status of frequencies to the multiple edges in the network, and an allocation control unit that assigns frequencies to the combinations of the multiple edges of the route vectors based on the feature vectors for the multiple frequencies, the feature vectors for the multiple edges, and the route frequency matrix.
[0013] One aspect of the present invention is a path control method executed by a path control device, comprising the steps of: deriving multiple route vectors representing combinations of multiple edges connecting from a start node to an end node based on physical topology information of a network and an start node and an end node of the network; deriving a route bandwidth matrix for each route vector, the route vectors being replicated according to a required bandwidth, and deriving feature vectors for the multiple frequencies and feature vectors for the multiple edges based on a route frequency matrix and the route bandwidth matrix representing the allocation status of frequencies to the multiple edges in the network; and assigning frequencies to the combinations of the multiple edges of the route vectors based on the feature vectors for the multiple frequencies, the feature vectors for the multiple edges, and the route frequency matrix.
[0014] One aspect of the present invention is a program for causing a computer to function as the above-mentioned path control device. Effect of the Invention
[0015] According to the present invention, it is possible to improve the efficiency of accommodating communications requested for connection. [Brief description of the drawings]
[0016] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of a path control system in a first embodiment. [Diagram 2] 4A to 4C are diagrams illustrating examples of physical topology information, route vectors, and route bandwidth matrices in the first embodiment. [Diagram 3] 4A to 4C are diagrams illustrating examples of a path band matrix, a path frequency matrix, a feasibility determination vector, and an information distance vector in the first embodiment. [Figure 4] FIG. 2 is a diagram illustrating an example of frequency allocation in a fixed grid system in the first embodiment. [Diagram 5] FIG. 4 is a sequence diagram showing an example of the operation of the path control system in the first embodiment. [Figure 6] FIG. 4 is a sequence diagram showing an example of the operation of a path control device in the first embodiment. [Figure 7] 4 is a sequence diagram showing an example of the operation of an allocation unit in the first embodiment. FIG. [Figure 8] FIG. 11 is a diagram illustrating an example of an effect of the path control system in the first embodiment. [Figure 9] FIG. 11 is a diagram illustrating an example of frequency allocation in a flexible grid system in the second embodiment. [Figure 10] FIG. 11 is a diagram illustrating an example of an effect of the path control system in the second embodiment. [Figure 11] FIG. 2 is a diagram illustrating an example of a hardware configuration of a path control device in each embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0017] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail with reference to the drawings. (First embodiment) 1 is a diagram showing an example of the configuration of a path control system 1 in the first embodiment. The path control system 1 is a system that controls the route and frequency of a communication line. The path control system 1 selects a route of the communication line in response to a communication connection request from a user terminal. In addition, the path control system 1 dynamically allocates a frequency to the selected route in response to the communication connection request from the user terminal.
[0018] The path control system 1 includes a communication line 2, a request management unit 4, a storage device 3, and a path control device 5. The path control device 5 includes a condition control unit 51 and an allocation unit 52. The allocation unit 52 includes an allocation control unit 53, a route derivation unit 54, and a feature derivation unit 55. The feature derivation unit 55 includes a matrix derivation unit 551, a determination unit 552, and a distance derivation unit 553.
[0019] FIG. 2 shows physical topology information 6, route vector 7, and route bandwidth matrix 8 “a i,j 1 is a diagram showing an example of "Physical topology information 6." Physical topology information 6 represents the physical topology of a communication line connecting a first user terminal (not shown) and a second user terminal (not shown). Physical topology information 6 comprises a plurality of nodes and "m" edges. The "m" edges connect predetermined nodes together. As an example, physical topology information 6 comprises node 60 to node 65, and edge 70 to edge 85.
[0020] The route derivation unit 54 derives a plurality of candidates of a route connecting a start node (transmitting node) and an end node (receiving node) based on the physical topology information 6. The start node is, for example, a node to which a first user terminal (not shown) is connected, and is, for example, the node 60. The end node is, for example, a node to which a second user terminal (not shown) is connected, and is, for example, the node 65. The first user terminal (not shown) communicates with the second user terminal (not shown) using an optical signal transmitted through a route to which a frequency is assigned by the path control device 5. In FIG. 2, a route including an edge 70, an edge 72, an edge 78, an edge 81, and an edge 83 is an example of a candidate route to be selected. The route derivation unit 54 outputs the physical topology information 6 including route candidate information, which is information representing route candidates, to the allocation control unit 53. The physical topology information 6 is expressed, for example, in the form of a directed graph.
[0021] The matrix derivation unit 551 derives a path vector 7 based on the path candidate information in the physical topology information 6. The path vector 7 is a vector of the total number of edges "m" in the physical topology information 6 (physical topology graph). Vector components (elements) shown in white in FIG. 2 represent components that do not constitute a path. The value of a component that does not constitute a path is 0. Vector components shown in black in FIG. 2 represent components that constitute a path. The value of a component that constitutes a path is 1.
[0022] The band "Δ" is a natural number that indicates the width of a frequency slot. The matrix derivation unit 551 converts (converts the format of) the physical topology information 6 including the route candidate information into a route band matrix 8 by duplicating the route vector 7 in the row direction by the number of bands "Δ". The route band matrix 8 is a matrix with Δ rows and m columns. In other words, the matrix derivation unit 551 converts the route band matrix 8 "a i,j " is defined for each route vector 7 (route candidate).
[0023] Therefore, the bandwidth "Δ" is the row size of the path bandwidth matrix 8. The bandwidth "Δ" is determined based on modulation method information. The modulation method information is the size of the frequency grid (required bandwidth) required to cover the signal band. For example, if the width of the frequency grid is 12.5 GHz and the signal band is 37.5 GHz (= 3 × 12.5 GHz), at least three frequency grids are required. Therefore, the natural number "Δ≧3" is given as the modulation method information.
[0024] In the fixed grid method, the modulation method (frequency grid width) changes according to the length of the path. Therefore, the determination unit 552 executes a convolution process between the matrix representing the usage status of the combination of frequency and edge and the path band matrix 8 (path candidate) using a window function having the frequency grid width of each path candidate. The reason for introducing the path band matrix 8 is to obtain the feature amount (feature vector) of each path in this way.
[0025] Fig. 3 is a diagram showing examples of a path band matrix 8, a path frequency matrix 9, a feasibility determination vector 10, and an information distance vector 11 in the first embodiment. The process shown in Fig. 3 can also be executed in the second embodiment. The path frequency matrix 9 represents the usage status (allocation status) of combinations of frequencies and edges in a network.
[0026] The band "Δ" (width of the frequency slot) is 2, for example. The size of the path frequency matrix 9 (bitmap) is j rows and k columns. The row size "j" is equal to the total number of edges "m" in the physical topology information 6 (physical topology graph). The column size "k" represents the total number of frequency slots. The total number of frequency slots "k" is the sum of the number "l" of frequency slots (allocation slots) that can be allocated to a path and the number of boundaries "2". The frequency slots that can be allocated to a path include frequencies that have not yet been allocated.
[0027] For the purpose of an extension to a flexible grid system (RSA extension) described later in the second embodiment, in the path frequency matrix 9, the value of each frequency slot (external boundary) outside the frequency slots that can be assigned to paths is fixed to the value "1" indicating that the frequency has been assigned, for the sake of convenience. That is, the determination unit 552 fixes the value of the component of the frequency slot other than the frequency slots that can be assigned to paths in the path frequency matrix 9 to the value "1" indicating that the frequency has been assigned.
[0028] Boundaries 90 and 91 are located outside the frequency slots that can be assigned to paths (assigned slots). The frequency slot number of boundary 90 is 0. The frequency slot number of boundary 91 is 17. In Fig. 3, all component values of frequency slot numbers "0" and "17" in path frequency matrix 9 are fixed to the value "1" indicating that the frequency has been assigned, as a boundary condition setting.
[0029] The determination unit 552 derives a feasibility determination vector 10 (frequency blocking determination vector) by performing a convolution integral (Hadamard integral) of the path band matrix 8 and the path frequency matrix 9. The feasibility determination vector 10 is a vector (integer sequence) that indicates whether or not frequency allocation is possible (degree of suitability in the frequency direction). If frequency blocking occurs, frequency allocation is not possible. The feasibility determination vector 10 "ξ n " is expressed as equation (1).
[0030]
number
[0031] Here, the validity judgment vector 10 "ξ n The frequency of a frequency slot in which the component (decision value) of " is 0 can be assigned to a route.
[0032] The distance derivation unit 553 derives an information distance vector 11 (integers) by performing a convolution integral (integral of the number of permutations) between the path band matrix 8 and the path frequency matrix 9. The information distance vector 11 is a vector that represents an index (information distance) of the path utilization efficiency. The path utilization efficiency is represented as, for example, a Hamming distance. The information distance vector 11 "ζ n " is expressed as equation (2).
[0033]
number
[0034] Here, the information distance vector 11 "ζ n It is optimal to assign to a path the frequency slot for which the component value of " is the largest." In other words, it is optimal to assign to a path the frequency slot for which the utilization efficiency of the path is the highest.
[0035] The allocation control unit 53 derives a path frequency allocation decision vector. The path frequency allocation decision vector is calculated by subtracting the possibility decision vector 10 "ξ n " in the denominator, and the information distance vector 11 "ζ n " in the numerator. The path frequency assignment decision vector is expressed as in equation (3).
[0036]
number
[0037] The allocation control unit 53 prioritizes a route with high route utilization efficiency and allocates frequencies to that route. For example, the allocation control unit 53 determines the start slot number with the longest Hamming distance "max" by identifying the slot number (index) that is the maximum among the components of the route frequency allocation decision vector. That is, the allocation control unit 53 determines the start slot number "n a " is derived as shown in equation (4).
[0038]
number
[0039] The allocation control unit 53 selects a frequency slot number "n" having the longest information distance "max" from among the allocation slots having the component value (decision value) of the possibility decision vector 10 of 0. a The allocation control unit 53 selects the frequency slot number "n a " based on the starting slot number column {n a +1,…,n a +Δ}. The allocation control unit 53 executes allocation and deletion of frequencies to paths as shown in equation (5).
[0040]
number
[0041] FIG. 4 is a diagram showing an example of frequency allocation in a fixed grid system (RWA) in the first embodiment. In the judgment vector 10 shown in FIG. 4, the value of a blank component is 1 or more. n The locations of "ζ = 0" are candidates for frequency allocation. In FIG. 4, the frequencies with frequency slot numbers "5," "10," and "15" are candidates for available frequencies. In the information distance vector 11 illustrated in FIG. 4, n 4, the allocation control unit 53 determines the maximum value of the information distance vector 11 in the path frequency matrix 9 as “ζ n The frequency grid number "n = 16" corresponds to the a =5". The allocation control unit 53 allocates the frequency band of the selected frequency grid number (frequency grid index) to each of the edges with edge numbers "0", "2", "8", "11", and "13".
[0042] Next, an example of the operation of the path control system 1 will be described. 5 is a sequence diagram showing an example of the operation of the path control system 1 in the first embodiment. In response to a request from the condition control unit 51, the storage device 3 transmits the physical topology information 6 and the path frequency matrix to the condition control unit 51. The condition control unit 51 acquires the physical topology information 6 and the path frequency matrix from the storage device 3 (step S101).
[0043] The request management device 4 transmits allocation request information to the condition control unit 51. The condition control unit 51 acquires the allocation request information from the request management device 4 (step S102). The condition control unit 51 transmits the allocation request information, physical topology information, and path frequency matrix to the allocation unit 52. The allocation unit 52 acquires the allocation request information, physical topology information, and path frequency matrix from the condition control unit 51 (step S103).
[0044] The allocation unit 52 derives a path vector based on the allocation request information and the physical topology information. The allocation unit 52 derives a path bandwidth matrix based on the path vector. The allocation unit 52 derives a feasibility determination vector and an information distance vector based on the path bandwidth matrix and the path frequency matrix. The allocation unit 52 updates the path frequency matrix by allocating frequency bands to paths based on the feasibility determination vector and the information distance vector. The allocation unit 52 transmits the updated path frequency matrix to the condition control unit 51. The condition control unit 51 acquires the updated path frequency matrix from the allocation unit 52 (step S104).
[0045] The condition control unit 51 extracts an allocation result indicating paths and frequency bands from the updated path frequency matrix. The condition control unit 51 transmits the allocation result to the request management device 4. The request management device 4 acquires the allocation result from the condition control unit 51 (step S105). The condition control unit 51 transmits the updated path frequency matrix to the storage device 3. The storage device 3 stores the updated path frequency matrix (step S106). Note that the condition control unit 51 may omit execution of step S106 and store the updated path frequency matrix.
[0046] Alternatively, step S101 may be executed only when the service is started, and the processes from step S102 to step S105 may be executed repeatedly.
[0047] 6 is a sequence diagram showing an example of the operation of the path control device 5 in the first embodiment. The condition control unit 51 transmits the allocation request information, the physical topology information, and the path frequency matrix to the allocation control unit 53. The allocation control unit 53 acquires the allocation request information, the physical topology information, and the path frequency matrix from the condition control unit 51 (step S103 illustrated in FIG. 5).
[0048] The allocation control unit 53 transmits the allocation request information and the physical topology information to the route derivation unit 54. The route derivation unit 54 derives route candidate information based on the allocation request information and the physical topology information (step S201). The route derivation unit 54 transmits the route candidate information to the allocation control unit 53. The allocation control unit 53 generates modulation method information for each route candidate based on the route candidate information. That is, the allocation control unit 53 determines a bandwidth "Δ" (necessary bandwidth information) required for the communication requested for allocation (step S202).
[0049] The allocation control unit 53 transmits the route frequency matrix, route candidate information, and modulation method information to the feature derivation unit 55. The feature derivation unit 55 derives a route band matrix based on the route candidate information and modulation method information (required band information). The feature derivation unit 55 derives a vector (possibility determination vector) representing whether or not a frequency can be allocated to each route candidate (first feature) based on a result of a convolution integral (compression in the frequency direction) using the route band matrix and the route frequency matrix. The feature derivation unit 55 derives a vector (information distance vector) representing an information distance (second feature) of each route candidate based on a result of a convolution integral (compression in the route direction) using the route band matrix and the route frequency matrix (step S203).
[0050] The feature derivation unit 55 transmits to the allocation control unit 53 a possibility / inhibition determination vector indicating whether or not a frequency can be allocated to each route candidate (whether or not there is no frequency blocking) and an information distance vector indicating the information distance of each route candidate (step S204). The allocation control unit 53 updates the route frequency matrix based on whether or not a frequency can be allocated to each route candidate (first feature) and the information distance of each route candidate (second feature). When there are one or more route candidates, the allocation control unit 53 selects a route from among the one or more route candidates based on whether or not a frequency can be allocated to each route candidate (first feature) and the information distance of each route candidate (second feature). The allocation control unit 53 allocates a frequency band (band) to the selected route. When there is no route candidate, the allocation control unit 53 generates information indicating that allocation is not possible (step S205).
[0051] The allocation control unit 53 transmits the updated path frequency matrix to the condition control unit 51. The condition control unit 51 acquires the updated path frequency matrix from the allocation control unit 53. When information indicating that allocation is not possible is generated, the allocation control unit 53 transmits the information indicating that allocation is not possible to the condition control unit 51. The condition control unit 51 acquires the information indicating that allocation is not possible from the allocation control unit 53 (step S104 illustrated in FIG. 5).
[0052] 7 is a sequence diagram showing an example of the operation of the allocation unit 52 in the first embodiment. The allocation control unit 53 transmits the path frequency matrix, the path candidate information, and the modulation method information to the feature derivation unit 55. The matrix derivation unit 551 derives a path band matrix based on the path candidate information and the modulation method information (necessary band information) (step S203 illustrated in FIG. 6).
[0053] The matrix derivation unit 551 transmits the path band matrix and the path frequency matrix to the determination unit 552 (step S301). The determination unit 552 derives a vector (possibility determination vector) indicating whether each path candidate can be allocated (first feature) based on a result of a convolution integral using the path band matrix and the path frequency matrix (step S302). The determination unit 552 transmits the vector (possibility determination vector) indicating whether each path candidate can be allocated to the allocation control unit 53. The allocation control unit 53 acquires the possibility determination vector from the determination unit 552 (step S204-1 illustrated in FIG. 6).
[0054] The matrix derivation unit 551 transmits the path band matrix and the path frequency matrix to the distance derivation unit 553 (step S303). The distance derivation unit 553 derives a vector (information distance vector) representing the information distance (second feature) of each path candidate based on a result of the convolution integral using the path band matrix and the path frequency matrix (step S304). The distance derivation unit 553 transmits the vector (information distance vector) representing the information distance of each path candidate to the allocation control unit 53. The allocation control unit 53 acquires the information distance vector from the distance derivation unit 553 (step S204-2 illustrated in FIG. 6).
[0055] The allocation control unit 53 allocates a frequency band to a route whose component value (judgment value) of the feasibility judgment vector is feasible and whose component value (information distance) of the information distance vector is the longest (step S205 illustrated in FIG. 6).
[0056] As described above, the route derivation unit 54 derives multiple route vectors 7 representing combinations of multiple edges connecting a start node to an end node based on the physical topology information 6 of the network and the start node and end node of the network. The feature derivation unit 55 derives a route bandwidth matrix 8 including route vectors replicated according to a required bandwidth for each route vector 7. The feature derivation unit 55 derives a feasibility determination vector 10 (feature vector related to multiple frequencies) and an information distance vector 11 (feature vector related to multiple edges) based on the route frequency matrix 9 and the route bandwidth matrix 8 representing the allocation status of frequencies to multiple edges in the network. The allocation control unit 53 allocates frequencies to the combinations of multiple edges of the route vector 7 based on the feasibility determination vector 10, the information distance vector 11, and the route frequency matrix 9.
[0057] This makes it possible to improve the efficiency of accommodating communications that have been requested for connection.
[0058] Since convolution processing has been introduced into the frequency allocation calculation, an algebraic algorithm is used instead of an algorithm in the form of a "for statement" and an "if statement", making it easier to introduce complex processing (for example, information distance derivation processing). As a result, the path control device 5 only needs to execute calculations of a fixed size (for example, a matrix), and the range of application to high-speed calculations using a GPU (Graphics Processing Unit) or the like in the path control device 5 increases. The path control device 5 can derive an approximate solution that is closer to the exact solution at high speed using an algorithm that allows it to derive an approximate solution.
[0059] The feasibility determination vector (frequency blocking determination vector) represents the degree of suitability in the frequency direction. The information distance vector represents the degree of suitability in the route direction. Since the size of the feasibility determination vector and the size of the information distance vector are the same, an optimal index (for example, the vector of formula (3)) that combines the frequency direction and the route direction is created. This makes it possible to quickly derive an approximate solution that is closer to the exact solution.
[0060] In the conventional method, an index based on either frequency or path information is derived, and optimality is determined based on such an index. For example, in the "First-Fit (FF) method," optimality is determined based only on information in the frequency direction.
[0061] In the conventional "Least Used (LU) method", the Random method, and the Most Used (MU) method, optimality is determined based only on information on the route direction. In the conventional methods, even if a route frequency matrix (bitmap) is derived, only information on how many free areas exist for only the route direction in the route frequency matrix can be handled. In contrast, in the path control device 5, the relationship between the route (path) that is a candidate for allocation and the route frequency matrix is defined in a generalized form as information distance. This makes it possible to express the degree of suitability for the route direction for each route that is a candidate for allocation in more detail.
[0062] Next, an example of the effect of the path control system 1 in the first embodiment will be described using "JPN25" representing the physical topology of a Japanese network (the number of nodes is 25).
[0063] FIG. 8 is a diagram showing an example of the effect of the path control system 1 in the first embodiment. When frequency blocking is constant, the larger the value on the horizontal axis (the communication load of the entire network) is, the more efficient the algorithm is. In other words, the further to the right the graph is in FIG. 8, the larger the total communication capacity that can be accommodated in the network is. The method of the path control system 1 in the first embodiment is a more effective method than, for example, the FF method. The path control system 1 in the first embodiment is capable of efficiently utilizing network resources.
[0064] 8, a user terminal (not shown) is placed at each node of the physical topology “JPN25.” The condition control unit 51 acquires the physical topology “JPN25” from the storage device 3.
[0065] In the fixed grid method, the required bandwidth is determined regardless of the modulation method. The condition control unit 51 acquires allocation request information from the request management device 4. The allocation request information is determined in advance as a parameter of the communication load of the entire network. In order to make the communication load of the entire network equal to the parameter (allocation request information), each user terminal generates a connection request with equal probability so that the service usage time is exponentially distributed.
[0066] For each parameter (allocation request information), the frequency blocking rate is calculated multiple times for multiple user terminals. The frequency blocking rate is the ratio of the number of failed connection requests to the number of successful connection requests. In other words, the frequency blocking rate is the value obtained by dividing the number of failed connection requests by the number of successful connection requests.
[0067] The path control device 5 creates an optimal index that combines the frequency direction and the route direction. The path control device 5 also derives detailed characteristics of the route direction as an information distance. This makes the communication network more resistant to communication loads.
[0068] Second embodiment The second embodiment differs from the first embodiment in that frequency allocation is performed based on a flexible grid system. The second embodiment will be described focusing on the differences from the first embodiment.
[0069] FIG. 9 is a diagram showing an example of frequency allocation in the flexible grid method in the second embodiment. In the "Exact-Fit (EF) method", matching is considered not only in the path (edge slot) direction but also in the frequency (frequency slot) direction. In the second embodiment, the "Exact-Fit (EF) method" and the flexible grid method are combined. This combination (extension) is called "EFHDM" (Exact-Fit Hamming Distance Maximization).
[0070] The corresponding arrow 92 is the vector 10 "ξ n The correspondence between the convolution matrix product in the derivation of " and the information distance vector 11 "ζ n In addition, to derive the degree of matching (fitness) in the frequency direction (horizontal axis), the frequency matching vector "θ S,n " and "θ L,n " is defined.
[0071] Frequency adaptation vector "θ S,n " is the boundary 93 with the smaller frequency slot number and the path band matrix 8 "a i,j The frequency match vector "θ S,n " is expressed as equation (6).
[0072]
number
[0073] Frequency adaptation vector "θ L,n " is the boundary 93 on the side with a larger frequency slot number and the path band matrix 8 "a i,j The frequency match vector "θ L,n " is expressed as equation (7).
[0074]
number
[0075] The vector (frequency compatibility decision vector) that is the largest when the degree of compatibility in the frequency direction is the highest and the frequency can be assigned (when frequency blocking does not occur) is the frequency compatibility vector "θ S,n " and the frequency matching vector "θ L,n " and the decision vector 10 "ξ n " and is expressed as in equation (8).
[0076]
number
[0077] When equation (8) is equal to equation (9), the degree of fit in the frequency direction is the highest.
[0078]
number
[0079] When formula (8) is equal to formula (9), that is, when the degree of match in the frequency direction is the highest, the determination unit 552 determines the result as "true" in the determination process. When the result as "true" in the determination process, the determination unit 552 assigns a frequency to the path using formula (5). When the result as "false" in the determination process, the determination unit 552 assigns a frequency to the path using, for example, the FF method.
[0080] As described above, the feature derivation unit 55 fixes the values of frequency slots other than the frequency slots that can be assigned to the path in the path frequency matrix to the value "1" which indicates that the frequency has already been assigned. The feature derivation unit 55 identifies the maximum slot number (index) among the components of the path frequency assignment judgment vector using a vector (frequency compatibility judgment vector) that has the highest degree of compatibility in the frequency direction and is maximum when frequency assignment is possible. In this way, the feature derivation unit 55 determines the start slot number with the longest Hamming distance "max". The feature derivation unit 55 assigns frequencies to the path by giving priority to the frequency slot with the start slot number.
[0081] This makes it possible to improve the efficiency of accommodating communications requested for connection in frequency allocation using the flexible grid method.
[0082] Next, an example of the effect of the path control system 1 in the second embodiment will be described. FIG. 10 is a diagram showing an example of the effect of the path control system 1 in the second embodiment. In the flexible grid method, the required bandwidth changes depending on the modulation method. When frequency blocking is constant, the larger the value on the horizontal axis (the communication load of the entire network) is, the more efficient the algorithm is. In other words, the further to the right the graph is in FIG. 10, the larger the total communication capacity that can be accommodated in the network is. The method of the path control system 1 in the second embodiment is a more effective method than, for example, the FF method. The path control system 1 in the second embodiment is capable of efficiently utilizing network resources.
[0083] (Hardware configuration example) FIG. 11 is a diagram showing an example of a hardware configuration of the path control device 5 in each embodiment. A part or all of the functional units of the path control device 5 are realized as software by a processor 101 such as a CPU (Central Processing Unit) executing a program stored in a storage device 102 having a non-volatile recording medium (non-transient recording medium) and a memory 103. The program may be recorded in a computer-readable non-transient recording medium. The computer-readable non-transient recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM (Read Only Memory), a CD-ROM (Compact Disc Read Only Memory), or a storage device such as a hard disk built into a computer system. The communication unit 104 executes a predetermined communication process. The communication unit 104 may acquire data and a program.
[0084] A part or all of the functional units of the path control device 5 may be realized by using hardware including an electronic circuit or circuitry using, for example, an LSI (Large Scale Integrated circuit), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), an FPGA (Field Programmable Gate Array), etc. The same applies to the request management device 4.
[0085] Although an embodiment of the present invention has been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment, and designs that do not deviate from the gist of the present invention are also included. [Industrial Applicability]
[0086] The present invention is applicable to communication systems. [Explanation of symbols]
[0087] 1...path control system, 2...communication line, 3...storage device, 4...request management device, 5...path control device, 6...physical topology information, 7...route vector, 8...route bandwidth matrix, 9...route frequency matrix, 10...possibility determination vector, 11...information distance vector, 51...condition control unit, 52...assignment unit, 53...assignment control unit, 54...route derivation unit, 55...feature derivation unit, 60-65...nodes, 70-85...edges, 90...boundary, 91...boundary, 92...corresponding arrow, 93...boundary, 94...boundary, 101...processor, 102...storage device, 103...memory, 104...communication unit, 551...matrix derivation unit, 552...determination unit, 553...distance derivation unit
Claims
1. a route derivation unit that derives a plurality of route vectors representing a combination of a plurality of edges connecting the start node to the end node based on physical topology information of the network and a start node and an end node of the network; a feature derivation unit that derives a route bandwidth matrix for each of the route vectors, the route bandwidth matrix including the route vectors replicated according to a required bandwidth, and derives feature vectors for the plurality of frequencies and feature vectors for the plurality of edges based on a route frequency matrix and the route bandwidth matrix that represent a frequency allocation status of the plurality of edges in the network; an allocation control unit that allocates frequencies to combinations of the plurality of edges of the path vector based on a feature vector related to the plurality of frequencies, a feature vector related to the plurality of edges, and the path frequency matrix; A path control device comprising:
2. the feature derivation unit derives a feature vector for the plurality of frequencies and a feature vector for the plurality of edges based on a result of a convolution integral of the path frequency matrix and the path band matrix; 2. The path control device according to claim 1.
3. the feature vector for the plurality of frequencies is a vector indicating whether a frequency can be assigned; The feature vector for the plurality of edges is a vector representing a utilization efficiency of a route in the network.
3. The path control device according to claim 2.
4. The efficiency of route utilization in the network is expressed as a Hamming distance.
4. The path control device according to claim 3.
5. the allocation control unit allocates the frequency by giving priority to a path that has a high utilization efficiency.
5. The path control device according to claim 3 or 4.
6. The feature derivation unit fixes values of frequency slots other than frequency slots that can be assigned to paths in the path frequency matrix to values indicating that the frequencies have been assigned. The path control device according to any one of claims 1 to 5.
7. A path control method executed by a path control device, comprising: deriving a plurality of path vectors representing a combination of a plurality of edges connecting the start node to the end node based on physical topology information of a network and a start node and an end node of the network; deriving a route band matrix for each of the route vectors, the route vectors being duplicated according to a required band, and deriving feature vectors for the plurality of frequencies and feature vectors for the plurality of edges based on a route frequency matrix and the route band matrix that represent an allocation status of frequencies to the plurality of edges in the network; assigning frequencies to combinations of the edges of the path vector based on the feature vectors for the frequencies, the feature vectors for the edges, and the path frequency matrix; A path control method including:
8. 7. A program for causing a computer to function as the path control device according to claim 1.
Citation Information
Patent Citations
Link band fluctuation upper limit calculation apparatus and link band fluctuation upper limit calculation method
JP2012004639A
