A method for recommending the optimal path for service activation

By building a topological path selection model, using multi-dimensional attribute features and business demand grouping, and combining with machine learning algorithms, the problem of single-dimensional determination of path recommendation rules in the existing technology is solved, and the high comprehensive performance and fast accuracy of dedicated line business opening paths is achieved. , improve operation and maintenance efficiency.

CN116232978BActive Publication Date: 2024-11-01INSPUR TIANYUAN COMM INFORMATION SYST CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310207317.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-03-01
Publication Date
2024-11-01
Estimated Expiration
2043-03-01

AI Technical Summary

Technical Problem

The existing dedicated line service opening path recommendation rules mainly determine the optimal route from a single direction or dimension, resulting in the low comprehensive performance of the recommended route path and the inability to quickly and accurately determine the optimal route.

Method used

By obtaining site information and topology structure diagrams, we build a topology path selection model, including the optimal path generation module and constraint module, use multi-dimensional attribute characteristics and business requirements grouping, adjust model parameters to generate the optimal path, and combine machine learning algorithms to achieve path recommendations .

Benefits of technology

It realizes the fast and accurate determination of the optimal opening path in the dedicated network topology, and improves the comprehensive performance and operation and maintenance efficiency of dedicated business opening.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116232978B_ABST
    Figure CN116232978B_ABST
Patent Text Reader

Abstract

The present invention provides a method for recommending an optimal path for opening a service, which belongs to the field of data transmission technology, and includes: obtaining the service requirements for opening a dedicated line service and a topological structure diagram of an existing optical cable; based on the obtained site information and the topological structure diagram, obtaining all topological nodes located between sites; constructing a topological path selection model, inputting the attribute characteristics of the newly obtained topological nodes located between sites into the constructed model, and obtaining the optimal path for opening this dedicated line service through the model. The present invention can quickly issue new services based on the existing dedicated line network topology and improve the opening efficiency of dedicated line services.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data transmission, and in particular to a method for recommending an optimal path for service activation. Background Art

[0002] At present, government and enterprise dedicated lines mainly include Internet dedicated lines, data dedicated lines, and virtual private network (VPN) dedicated lines. Internet dedicated lines are usually for small and medium-sized customers, mainly providing Internet access and access, using passive optical network (PON), packet transport network (PTN), multi-service access platform (MSAP) or protocol conversion equipment, which is basically consistent with the architecture of ordinary home broadband networks. Data dedicated lines are usually for large enterprise customers, mainly providing data transmission channels within the enterprise, mainly using PTN, multi-service transmission platform / synchronous digital hierarchy (MSTP / SDH) and other equipment. VPN dedicated lines are usually for large enterprise customers to achieve VPN interconnection between networks in different office areas, mainly using multi-protocol label switching-virtual private network (MPLS-VPN) equipment.

[0003] With the development of globalization, informatization and cloud services, the demand for dedicated lines will increase. The dedicated line service market will become the focus of competition in the 2B market in addition to operators' 2C services (mobile services, broadband services).

[0004] When opening a government and enterprise dedicated line service, the most important step is to design and plan a suitable optical path based on the selected starting and ending points.

[0005] The topology of a dedicated line network includes nodes composed of network element devices and pipelines connecting the nodes. Dedicated line opening means allocating a pipeline resource that can realize network data transmission between a given starting point and end point (for example, A-Z).

[0006] Usually, when recommending pipeline routes, the main recommendation rules can be classified into: optimal idle connection path, short distance, few jumps, non-co-routing, low latency, etc. These existing recommendation rules all determine the optimal route from a single direction or dimension, which will result in relatively low comprehensive performance of the recommended routing path. Summary of the invention

[0007] In order to solve the above technical problems, the present invention provides a method for recommending the optimal path for service activation. When solving the problem of OSS field dedicated line service activation, considering that the service activation needs to be carried out between the starting point and the end point after the two points are determined, it helps business personnel to quickly and accurately determine the optimal dedicated line activation path based on the existing dedicated line network topology.

[0008] The technical solution of the present invention is:

[0009] A method for recommending an optimal path for service activation, comprising:

[0010] Obtaining site information in the private line service activation requirements and a topological structure diagram of existing optical cables, wherein the site information includes start site information and end site information;

[0011] Based on the obtained site information and topology diagram, all topology nodes located between the sites are obtained;

[0012] A topology path selection model is constructed, and the attribute characteristics of the newly acquired topology nodes between the sites are input into the constructed model, and the optimal path for the opening of this dedicated line service is obtained through the model.

[0013] Furthermore,

[0014] The several sites in the site information are divided into several site intervals according to their location proximity, and then the topology structure diagram is traversed according to the intervals to obtain a set of child nodes. The child nodes are combined and merged to obtain a topology node set, that is, all topology nodes located between the sites.

[0015] The topological path selection model includes an optimal path generation module and one or more constraint modules.

[0016] The training steps of the path selection model are as follows:

[0017] 1) Input the multi-dimensional attribute features of the m topological nodes (which can be called transfer points) X = {x_1, x_2, ..., x_m} located between the starting site and the ending site into the optimal path generation module;

[0018] 2) Grouping all service activation requirements to obtain several service activation requirement groups;

[0019] 3) Inputting the activation attribute characteristics and several service activation demand groups into the constraint module to obtain multiple groups of topological path sets corresponding to the demand groups;

[0020] 4) The parameter adjustment coefficient is added to the model parameters in the optimal path generation model until the optimal path generated by the optimal path generation model after parameter adjustment is kept within a relatively constant range with the path distance in the multiple groups of topological path sets obtained by the constraint model, and the model training is completed.

[0021] Going further,

[0022] The multi-dimensional attribute features include the allocated bandwidth of the node on the topological link, the topological relationship of the node, and the QOT value of the topological link where the node is located. The QOT value can reflect the physical performance of the link.

[0023] The grouping of all the service activation requirements to obtain a plurality of service activation requirement groups includes: clustering all the service activation requirements to obtain a plurality of service activation requirement groups.

[0024] The path distances between one or more topological paths and the optimal path generated by the optimal path generation model are calculated respectively, and then the set distance threshold (the average of the path distances obtained each time) is used as the parameter adjustment coefficient of the optimal path generation model. The parameter adjustment coefficient can be converted into a suitable value range by conversion.

[0025] Going further,

[0026] The newly acquired attribute features of the m topological nodes X = {x_1, x_2, ..., x_m} located between the sites are input into the constructed optimal path generation module, and the optimal path for the opening of this dedicated line service is obtained through the model.

[0027] The beneficial effects of the present invention are

[0028] By allocating bandwidth of nodes on topological links, topological relationships of nodes, and QOT values ​​of topological links where nodes are located, and fitting recommended values ​​of paths for different business requirements in the path selection model, it is possible to quickly and accurately determine the optimal dedicated line opening path based on the existing dedicated line network topology. Compared with the prior art solutions that achieve optimal route determination from a single direction or dimension, the routing path recommended by the solution of the present invention can determine a dedicated line service opening path with high comprehensive performance as a whole, thereby improving the operation and maintenance efficiency of dedicated line service opening. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a working schematic diagram of the present invention. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0031] The technical solution of the present invention can ensure that the recommended routing path achieves the minimum overall delay and the fewest hops by designing a method for recommending the optimal optical path for opening a dedicated line service.

[0032] The optimal path recommendation method of the technical solution of the present invention realizes the optimal path recommendation by combining the machine learning algorithm. The detailed implementation scheme is as follows:

[0033] S1. Obtain the business requirements for opening a dedicated line service and the topological structure diagram of the existing optical cable. The business requirements include the site information for opening the dedicated line, such as A-G-Y.

[0034] S2. Based on the acquired site information and topology diagram, all topology nodes located between the sites are acquired;

[0035] Specifically, the multiple sites in the site information can be divided into multiple site intervals according to their location proximity, and then the topology structure diagram can be traversed according to the intervals, so that multiple groups of child node sets can be obtained. The child nodes can be combined and merged to obtain a topological node set, that is, all topological nodes located between the sites.

[0036] S3, constructing a topological path selection model, which includes an optimal path generation module and multiple constraint modules;

[0037] Specifically, the role of multiple constraint modules is to constrain the training of the optimal path generation module. The training logic of the path selection model is as follows:

[0038] S3-1. Obtain m topological nodes (which may be called transfer points) located between sites X = {x1, x2, ..., x m}’s attribute features are input into the optimal path generation module;

[0039] Among them, the attribute characteristics include the allocated bandwidth of the node on the topological link, the topological relationship of the node and the QOT value of the topological link where the node is located. The QOT value can reflect the physical performance of the link. The existing QOT performance prediction methods include Gaussian noise (GN) model, Network Kriging (NK) algorithm and machine learning methods (decision tree, support vector machine, random forest, neural network, etc.).

[0040] S3-2, cluster all business requirements to obtain multiple business requirement groups;

[0041] By clustering the acquired business demand sample data, the business demands can be grouped. The demand classification after clustering may include: optimal idle connection path, short distance, few hops, non-co-routing, low latency, etc.

[0042] S3-3, inputting the attribute features and the multiple demand groups after clustering into multiple constraint modules to obtain multiple sets of topological path sets corresponding to the demand groups;

[0043] The constraint module is actually not based on a prediction model built on a neural network, through which the topological paths corresponding to different topological recommendation rules can be obtained.

[0044] S3-4. Calculate the path distances between multiple groups of topological paths and the optimal path generated by the optimal path generation model respectively, and then use the set distance threshold (the average of the path distances obtained each time) as the parameter adjustment coefficient of the optimal path generation model (the parameter adjustment coefficient can be converted into a value range).

[0045] S3-5. Add the parameter adjustment coefficient to the model parameters in the optimal path generation model until the optimal path generated by the optimal path generation model after parameter adjustment is kept within a relatively constant range with the path distance in the multiple groups of topological path sets obtained by the constraint model, and the model training is completed.

[0046] S4. Input the attribute characteristics of the newly acquired m topological nodes (which can be called transfer points) X={x_1, x_2,..., x_m} located between the sites into the constructed optimal path generation module, and obtain the optimal path for the activation of this dedicated line service through the model.

[0047] The above description is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, and is not used to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A method for recommending an optimal path for service activation, characterized in that: include Obtaining site information in the private line service activation requirements and a topological structure diagram of existing optical cables, wherein the site information includes start site information and end site information; Based on the obtained site information and topology diagram, all topology nodes located between the sites are obtained; Construct a topology path selection model, input the multi-dimensional attribute characteristics of topology nodes into the constructed model, and obtain the optimal path for opening dedicated line services; The several sites in the site information are divided into several site intervals according to the location adjacency relationship, and then the topological structure diagram is traversed according to the interval to obtain the child node set, and the child nodes are combined to obtain the topological node set, that is, all the topological nodes located between the sites are obtained; The topological path selection model includes an optimal path generation module and one or more constraint modules; The training steps of the path selection model are as follows: 1) Input the multi-dimensional attribute features of m (>=0) topological nodes or transfer points X={x_1,x_2,…,x_m} between the starting site and the ending site into the optimal path generation module; 2) Group all service activation requirements to obtain several service activation requirement groups; 3) Input the activation attribute characteristics and several service activation requirement groups into the constraint module to obtain a set of topological paths corresponding to the requirement groups; 4) The parameter adjustment coefficient is added to the model parameters in the optimal path generation module until the distance between the optimal path generated by the optimal path generation module after parameter adjustment and the path in the topological path set obtained by the constraint module remains in a relatively constant range, and the model training is completed; The multi-dimensional attribute features include the allocated bandwidth of the node on the topological link, the topological relationship of the node, and the QOT value of the topological link where the node is located; The grouping of all the service activation requirements to obtain a plurality of service activation requirement groups includes: clustering all the service activation requirements to obtain a plurality of service activation requirement groups.

2. The method according to claim 1, characterized in that: The path distances between one or more topological paths and the optimal path generated by the optimal path generation module are calculated respectively, and then a set distance threshold is used as a parameter adjustment coefficient of the optimal path generation module, and the distance threshold is the average of the path distances obtained each time.

3. The method according to claim 2, characterized in that The parameter adjustment coefficient can be converted into a value range.