CN117201322A - Path planning method, system and medium of hierarchical SDN network - Google Patents

Abstract

The application discloses a path planning method, a system and a medium of a hierarchical SDN network, belonging to the field of path planning. Aiming at the problems of large calculation amount and low efficiency of the existing path planning, the application provides a path planning method of a hierarchical SDN network, which comprises the steps of carrying out hierarchical operation on network topology: dividing the network topology into N layers, wherein the N-layer network topology comprises a plurality of first layers, each first layer comprises a plurality of second layers, and the N-th layer comprises a plurality of devices; calculating all available paths between n+1th layers in the n-th layer; and calculating all available paths between the nth layer and the n-1 th layer; calculating all available paths between the nth layer and each of the n+1th layers; n is more than or equal to 2 and N is less than N; calculating all available paths between the plurality of first layers and between the plurality of devices within the nth layer; an optimal path between any two devices is calculated. The method can save time and resource cost, and the calculation of 10w+ available paths between any two devices is controlled within 200ms, so that the efficiency is high.

Description

Path planning method, system and medium of hierarchical SDN network

Technical Field

The application belongs to the technical field of path planning, and particularly relates to a path planning method, system and medium of a hierarchical SDN network.

Background

The path planning of the traditional network depends on a routing protocol, and the Dijkstra algorithm is adopted, so that only the shortest path of the router local can be calculated, and the global optimal path can not be calculated. The application of SDN (software defined networking) technology separates a network control plane and a forwarding plane, so that the control plane can acquire global network topology information, and a foundation is provided for calculating a global optimal path. SDN technology is applied to a DCN (Data Center Network ) network environment at the earliest time, an Openflow switch is used as a forwarding plane, and an ODL controller is used as a control plane. The ODL controller dynamically collects network link information and node information, generates a global network topology structure, calculates the shortest path between any two points by using an SPF (Shortest Path First ) algorithm, and sends the path to the switch in the form of an Openflow flow table; the ODL controller can simultaneously realize functions of congestion control, load balancing and the like in traffic engineering (Traffic Engineering), but the preconditions still depend on global topology and SPF algorithm. Application of SDN technology in SPN (Service Provider Network, operator network) networks uses SR-TE (Segment Routing Traffic Engineering) technology, with a router with SR capability as a forwarding plane and an ODL controller as a control plane. The ODL controller collects network link state information and node information by using BGP-LS protocol, generates global network topology structure, calculates shortest path between any two points by using centralized or distributed PCE (PathComputation Element, path calculation unit), and issues path to router in SR Policy form. The inside of the PCE uses an SPF algorithm such as Dijkstra or A; SR-TE provides a variety of traffic engineering techniques.

In the SDN technology architecture, the SPF algorithm adopted by the controller is an algorithm for calculating the shortest path in the network, which has some drawbacks: the complexity is high: the SPF algorithm is a greedy algorithm that requires traversing the entire network topology to calculate the shortest path from each node to the other nodes. For a large network, the calculation complexity is high, and a large amount of resources and time are consumed; topK pathway: the SPF algorithm can only calculate one path, and cannot cope with a scene requiring to provide a topK path; multi-constraint control: the SPF algorithm can only calculate the shortest path according to a single constraint condition, and cannot support a complex multi-constraint scenario, for example, the SPF algorithm can calculate a path with the minimum value of the metric of the link, but cannot calculate a path with the minimum value of the metric within a given delay range. Multiple constraint scenes require additional mechanisms and algorithms to assist, but for complex constraint scenes, the SPF algorithm cannot calculate the optimal path.

Corresponding improvements are also made to the above problems, for example, chinese patent application number CN202111031564.8, publication date 2021, 12 months, 17 days, and the patent discloses an in-band network telemetry detection path planning method of an SDN network, which includes: dividing the network topology into a plurality of topology subgraphs; and planning an in-band network telemetry detection path for the topology subgraph. Although the method can effectively reduce huge expenditure caused by rewriting the configuration program due to topology change, the whole method still cannot calculate the optimal path.

Disclosure of Invention

1. Problems to be solved

Aiming at the problems of large calculation amount and low efficiency of the existing path planning, the application provides a path planning method, a system and a medium of a hierarchical SDN network. The application carries out layering operation on the network topology by adopting the dynamic programming idea, and calculates all available paths in each layer or between adjacent layers, thereby eliminating repeated calculation amount, greatly reducing calculation amount and improving calculation efficiency; the whole flow is simple, and the time cost and the resource cost are saved, so that the requirements of a scene constrained by a service are met.

2. Technical proposal

In order to solve the problems, the application adopts the following technical scheme.

A path planning method of a hierarchical SDN network, comprising:

layering the network topology: dividing the network topology into N layers, wherein the N-layer network topology comprises a plurality of first layers, each first layer comprises a plurality of second layers, and the N-th layer comprises a plurality of devices;

calculating all available paths between n+1th layers in the n-th layer; and calculating all available paths between the nth layer and the n-1 th layer; and calculating all available paths between the nth layer and each of the n+1th layers; n is more than or equal to 2 and N is less than N;

calculating all available paths between the plurality of first layers and between the plurality of devices within the nth layer;

and splicing available paths, and calculating an optimal path between any two devices.

Further, calculating all available paths by adopting a BFS algorithm; the BFS algorithm comprises the steps of starting from a starting node, expanding outwards step by step, accessing all adjacent nodes of the starting node, then sequentially accessing all adjacent nodes of the adjacent nodes, and so on until all reachable nodes or target nodes are traversed.

Furthermore, during the process of traversing adjacent nodes, heuristic factors are added; the heuristic factors comprise a non-return path heuristic factor, a hop limit heuristic factor and a time delay limit heuristic factor.

Further, the network topology is divided into three layers according to the areas, the cities and the devices, wherein the three layers of network topology comprise a plurality of area topologies, each area topology comprises a plurality of city topologies, and each city topology comprises a plurality of devices; i.e. the first layer is a regional topology, the second layer is a city topology, and the third layer is a device topology.

Further, when the network topology is divided into three layers, it calculates all available paths specifically including:

calculating all available paths from one region to another region;

calculating all available paths from the city in the area to the exit city of the area;

calculating all available paths from other entrance cities to other exit cities in the area;

calculating all available paths from the entrance city to the city in the other region;

all available paths from the ingress device to the egress device for each of all cities within the region and the other region are calculated.

Further, when the optimal path selection is performed, the optimal path selection may be performed after all the available paths are calculated; the optimal path can also be selected for all available paths of the city and the area before the computing device, and then the final selection can be performed on the basis of the optimal paths of the city and the area after the optimal paths pass through all available paths after the computing device.

A system using a path planning method of a hierarchical SDN network as defined above, comprising:

a network topology layering module: dividing the network topology into N layers, wherein the N-layer network topology comprises a plurality of first layers, each first layer comprising a plurality of second layers, and the N-th layer comprising a plurality of devices;

and a path calculation module: for calculating all available paths between each n+1 layer within the n-th layer; and calculating all available paths between the nth layer and the n-1 th layer; and calculating all available paths between the nth layer and each of the n+1th layers; calculating all available paths between the plurality of first layers and between the plurality of devices within the nth layer;

an optimal path generation module: and the method is used for splicing available paths and calculating the optimal path between any two devices.

A computer readable storage medium storing a computer program which when executed by a processor implements a path planning method of a hierarchical SDN network as described above.

3. Advantageous effects

Compared with the prior art, the application has the beneficial effects that:

(1) The application carries out layering operation on the network topology by adopting the dynamic programming idea, and calculates all available paths in each layer or between adjacent layers, thereby eliminating repeated calculation amount, greatly reducing calculation amount and improving calculation efficiency; and by calculating all available paths and then splicing all available paths, the optimal path between any two devices is obtained, the whole flow is simple, the path planning is not needed by complex and lengthy steps, the optimal path determining efficiency is quickened, and the working efficiency of the whole path planning is greatly improved; the whole path planning method is simple and convenient to operate, can meet the requirements of giving an optimal path, simultaneously reduces the calculated amount, saves the time cost and the resource cost, enables the optimal path to meet the requirements of a scene constrained by a service, and has wide application prospect when the calculation of 10w+ available paths between any two devices is controlled within 200 ms;

(2) The calculation of all available paths in the method adopts a BFS algorithm, and the BFS algorithm can exhaust all available paths between any two devices, thereby being applicable to the scene requiring calculation of the topK path; the BFS algorithm can flexibly adapt to the constraint conditions of the service in the process of calculating and splicing paths, and finally solves all paths meeting the multiple constraint conditions; meanwhile, a heuristic factor is added in the calculation process of the BFS algorithm, and the heuristic factor is used for helping the BFS algorithm to better select the next hop node, so that the calculation performance of the algorithm is improved, and the calculation of the algorithm can meet the requirements of accuracy and efficiency;

(3) According to the application, the network topology is divided into three layers according to the region, the city and the equipment, wherein the three layers are in tree-like relation, namely, the region is used as a father node, the city is used as a first-stage child node, and the equipment is used as a second-stage child node; the method can realize orderly and logical layering of the network topology more accurately, and then avoid the problems of low calculation efficiency and poor precision caused by repeated subsequent calculation of the same path; when the optimal path is selected, all available paths can be calculated through selecting all available paths, and the flow operation is simple in this way; the method can also be used for splicing all available paths after calculating part of the available paths to obtain a pre-optimal path strength, and splicing all the remaining available paths on the basis of the pre-optimal path after calculating the last available paths to obtain a finished optimal path.

Drawings

FIG. 1 is a schematic flow chart of the present application.

Detailed Description

The application is further described below in connection with specific embodiments and the accompanying drawings.

Example 1

As shown in fig. 1, a path planning method of a hierarchical SDN network includes:

s1: layering the network topology: dividing the network topology into N layers, wherein the N-layer network topology comprises a plurality of first layers, each first layer comprises a plurality of second layers, and the N-th layer comprises a plurality of devices; layering the network topology to form a plurality of sub-topology modules, and respectively carrying out subsequent calculation operation on each sub-topology module, so that the whole calculation of the network topology is divided into a plurality of sub-topology calculations, and the calculation operation amount is greatly reduced; meanwhile, the problems of long calculation amount and low efficiency caused by repeated operation are avoided;

when dividing the network topology, the division is specifically performed as follows: the first layer and the second layer are in tree-like relation until the N layer is formed, namely the first layer is used as a father node, the second layer is used as a first-level child node, the third layer is used as a second-level child node, and the N layer is used as an N-1 level child node … …; such that the first layer and the second layer are in surrounding relation; the division mode can realize orderly and logical layering of the network topology, and the problems of low calculation efficiency and poor precision caused by subsequent repeated calculation of the same path due to overlapping layering areas are avoided to the greatest extent;

s2: calculating all available paths between n+1th layers in the n-th layer; and calculating all available paths between the nth layer and the n-1 th layer; and calculating all available paths between the nth layer and each of the n+1th layers; n is more than or equal to 2 and N is less than N; calculating all available paths between the plurality of first layers and between the plurality of devices within the nth layer;

specifically, calculating all available paths between each n+1 layer within the n-th layer means: because the same layer contains a plurality of next layers, the next layers belong to the levels, and the n+1 layers refer to the layers between the levels which are the same as the upper layers; such as calculating all available paths between two second layers within a first layer; calculating all available paths between two third layers within the second layer, and so on; calculating all available paths between the nth layer and n-1, calculating all available paths between the nth layer and each n+1 means: the first layer, the second layer and the third layer are in a tree-shaped relation graph from the nth layer, and belong to a top-bottom relation; such as calculating all available paths between the second layer and the first, second and third layers; calculating all available paths between the third layer and the second layer, and so on; because the first layer and the last layer N layer do not involve calculation with the upper and lower layers, only all available paths in the first layer and the last layer are required to be calculated independently, and the problems of large calculation amount and low efficiency caused by repeated calculation are avoided;

in one embodiment, in calculating all available paths for different objects, a BFS algorithm is used to calculate all available paths; the BFS algorithm comprises the steps of starting from a starting node, expanding outwards step by step, accessing all adjacent nodes of the starting node, then sequentially accessing all adjacent nodes of the adjacent nodes, and so on until all reachable nodes or target nodes are traversed. The BFS algorithm can exhaust all available paths between any two points, and is suitable for a scene requiring calculation of a topK path; the BFS algorithm can flexibly adapt to the constraint conditions of the service in the process of calculating and splicing paths, and finally solves all paths meeting the multiple constraint conditions;

splicing available paths, and calculating an optimal path between any two devices; in this step, when all available paths between the n+1th layers in the n-th layer are calculated; and calculating all available paths between the nth layer and the n-1 th layer; and calculating all available paths between the nth layer and each of the n+1th layers; n is more than or equal to 2 and N is less than N; after all available paths among a plurality of first layers and among a plurality of devices in an N layer are calculated, all available paths are spliced to obtain an optimal path, wherein the optimal path refers to a path required according to the requirement of service constraint, can be a shortest path, and can be an optimal path meeting a plurality of constraint conditions such as bandwidth, time delay range, link utilization, hop count, link/node must be wound;

the application carries out layering operation on the network topology by adopting the dynamic programming idea, and calculates all available paths in each layer or between adjacent layers, thereby eliminating repeated calculation amount, greatly reducing calculation amount and improving calculation efficiency; and by calculating all available paths and then splicing all available paths, the optimal path between any two devices is obtained, the whole flow is simple, the path planning is not needed by complex and lengthy steps, the optimal path determining efficiency is quickened, and the working efficiency of the whole path planning is greatly improved; the whole path planning method is simple and convenient to operate, can meet the requirements of giving an optimal path, simultaneously reduces the calculated amount, saves the time cost and the resource cost, enables the optimal path to meet the requirements of a scene constrained by a service, and has wide application prospect when the calculation of 10w+ available paths between any two devices is controlled within 200 ms.

In a specific embodiment, the network topology is divided into three layers according to areas, cities and devices, wherein the three layers of network topology comprise a plurality of area topologies, each area topology comprises a plurality of city topologies, and each city topology comprises a plurality of devices; namely, the first layer is regional topology, the second layer is city topology, and the third layer is equipment topology; specifically, the division of the regions is generally performed according to the following table 1:

table 1: region dividing table

Region code	Description of the application
		CN	China (excluding Kong Australian platform)
US	United states district
		AP	asia-Tai region (excluding China and the United states)
EU	European district

The city is divided according to three codes of airport, which are shown in the following table 2:

table 2: city dividing table

The division of the devices is based on the HOSTNAME of the router, and the naming rules are: region code, city code, data center name, equipment number, for example: CN-PEK-CTG-1. To further facilitate understanding of the present application, it is assumed in this embodiment that all paths between two devices across the region need to be calculated, where HOSTNAME of one device is CN-SHA-ZRT-1 and HOSTNAME of the other device is AP-SIN-ZRT-2; according to the idea of dynamic planning, the network topology is divided into three sub-topologies, and the calculation of all available paths comprises the following steps:

the first step: calculating all available paths from one region to another region; i.e. calculating all available paths from the CN area to the AP area;

and a second step of: based on the calculation result in the first step, in topology with granularity of city, respectively calculating:

calculating all available paths from the city in the area to the exit city of the area; i.e. calculate all available paths of SHA to CN area exit cities;

calculating all available paths from other entrance cities to other exit cities in the area; i.e. calculating all available paths from other regional ingress cities to other regional egress cities;

calculating all available paths from the entrance city to the city in the other region; i.e. calculating all available paths from the AP area entry city to the SIN;

and a third step of: according to the calculation results of the first step and the second step, all available paths from CN-SHA-ZRT-1 to AP-SIN-ZRT-2 taking the city as granularity are spliced;

fourth step: based on the splicing result of the third step, calculating all available paths from the entrance device to the exit device of each city in all cities in the region and the other region in a topology with granularity of the devices; i.e. calculating all available paths from the ingress device to the egress device inside each city separately; such as:

CN-SHA-ZRT-1 to CN-SHA-ZRT-2

AP-HKG-ZRT-2 to AP-HKG-ZRT-4

AP-SIN-ZRT-8 to AP-SIN-ZRT-2

Fifth step: splicing the splicing result of the third step and the calculation result of the fourth step to obtain all available paths from CN-SHA-ZRT-1 to AP-SIN-ZRT-2 by taking equipment as granularity;

sixth step: and splicing the optimal path according to the requirement of service constraint.

Specifically, in the process of traversing adjacent nodes by the BFS algorithm, adding heuristic factors in the steps; the heuristic factors comprise a non-return path heuristic factor, a hop limit heuristic factor and a time delay limit heuristic factor; the heuristic types and descriptions are shown in Table 3 below:

table 3: heuristic factor description

The addition of the heuristic factors is used for helping the BFS algorithm to better select the next hop node, so that the calculation performance of the algorithm is improved, and the calculation of the algorithm can meet the requirements of accuracy and efficiency.

In one embodiment, when optimal path selection is performed, optimal path selection may be performed after all available paths are calculated; all paths among three levels of the area, the city and the equipment or between the upper layer and the lower layer are completely calculated, and then the optimal paths are spliced; the method has simple flow logic, is easy and convenient to operate, and is spliced after all calculation is completed.

The optimal path can also be selected for all available paths of the city and the area before the computing device, and then the final selection is performed on the basis of the optimal paths of the city and the area after the optimal paths pass through all available paths behind the computing device (such as the above-mentioned example of calculating all paths between two devices crossing the area); the method is that all available paths with cities as granularity are spliced after city topology calculation is carried out; then calculating all available paths from the ingress device to the egress device inside each city; and combining all the calculated available paths among all the devices on the basis of all the spliced available paths with the cities as granularity, and finally splicing the optimal paths according to the requirements of service constraints. The method effectively reduces the calculated amount, saves time and improves efficiency.

Example 2

A system using the path planning method of a hierarchical SDN network as set forth in the above embodiment, comprising:

a network topology module: dividing the network topology into N layers, wherein the N-layer network topology comprises a plurality of first layers, each first layer comprising a plurality of second layers, and the N-th layer comprising a plurality of devices;

and a path calculation module: for calculating all available paths between each n+1 layer within the n-th layer; and calculating all available paths between the nth layer and the n-1 th layer; and calculating all available paths between the nth layer and each of the n+1th layers; calculating all available paths between the plurality of first layers and between the plurality of devices within the nth layer;

an optimal path generation module: and the method is used for splicing available paths and calculating the optimal path between any two devices.

It is worth noting that the system is deployed on a k8s management platform in the form of micro-service and serves as a Web service. The service comprises the three modules: the system comprises a network topology module, a path calculation module and an optimal path generation module. The specific operation process is as follows:

1. the client sends a path calculation request to the service;

2. service reception request:

2.1, resolving the calculation request into a starting node, a terminating node and a constraint set;

2.2, calling a network topology module to generate a global topology matrix diagram;

2.3, transmitting the start-stop node, the constraint set and the global topology matrix diagram to a calculation module;

2.4, the calculation module respectively calls different calculation subtasks and splicing tasks according to the position relation of the start and stop nodes (crossing regions, crossing cities in regions and crossing cities in the same city), and calculates all available paths meeting constraint conditions;

2.5, generating an optimal path generating module according to all available paths according to service requirements;

3. the service returns the path computation results to the customer.

The system meets the requirements of a business multi-constraint scene, and provides an optimal path meeting the multi-constraint conditions of bandwidth, time delay range, link utilization rate, hop count, link/node must-be-wound and the like; the method provides possibility and imagination space for expansion of business boundaries of companies; providing a technical foundation for better realization of traffic engineering; the whole system has simple structure and wide application scene.

Example 3

A computer readable storage medium storing a computer program which when executed by a processor implements a path planning method for a hierarchical SDN network as described in the above embodiments. It should be noted that, for a person skilled in the art, it is fully possible to implement the same program in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. by logic programming the method steps, except for implementing the system and the respective modules provided in the present application in a pure computer readable program code. Therefore, the system provided by the present application and each module thereof may be regarded as a hardware component, and the modules included therein for implementing various programs may be regarded as structures in the hardware component, and the modules for implementing various functions may be regarded as structures in both the hardware component and the software program for implementing the method.

The examples of the present application are merely for describing the preferred embodiments of the present application, and are not intended to limit the spirit and scope of the present application, and those skilled in the art should make various changes and modifications to the technical solution of the present application without departing from the spirit of the present application.

Claims (8)

1. A path planning method of a hierarchical SDN network is characterized in that: comprising the following steps:

layering the network topology: dividing the network topology into N layers, wherein the N-layer network topology comprises a plurality of first layers, each first layer comprises a plurality of second layers, and the N-th layer comprises a plurality of devices;

calculating all available paths between n+1th layers in the n-th layer; and calculating all available paths between the nth layer and the n-1 th layer; and calculating all available paths between the nth layer and each of the n+1th layers; n is more than or equal to 2 and N is less than N;

calculating all available paths between the plurality of first layers and between the plurality of devices within the nth layer;

and splicing available paths, and calculating an optimal path between any two devices.

2. The path planning method of a hierarchical SDN network of claim 1, wherein: calculating all available paths by adopting a BFS algorithm; the BFS algorithm comprises the steps of starting from a starting node, expanding outwards step by step, accessing all adjacent nodes of the starting node, then sequentially accessing all adjacent nodes of the adjacent nodes, and so on until all reachable nodes or target nodes are traversed.

3. The path planning method of a hierarchical SDN network of claim 2, characterized by: adding a heuristic factor in the process of traversing adjacent nodes; the heuristic factors comprise a non-return path heuristic factor, a hop limit heuristic factor and a time delay limit heuristic factor.

4. A path planning method of a hierarchical SDN network according to claim 1 or 2, characterized in that: dividing the network topology into three layers according to regions, cities and devices, wherein the three-layer network topology comprises a plurality of region topologies, each region topology comprises a plurality of city topologies, and each city topology comprises a plurality of devices; i.e. the first layer is a regional topology, the second layer is a city topology, and the third layer is a device topology.

5. The path planning method of a hierarchical SDN network of claim 4, wherein: when the network topology is divided into three layers, it computes all available paths specifically including:

calculating all available paths from one region to another region;

calculating all available paths from the city in the area to the exit city of the area;

calculating all available paths from other entrance cities to other exit cities in the area;

calculating all available paths from the entrance city to the city in the other region;

all available paths from the ingress device to the egress device for each of all cities within the region and the other region are calculated.

6. The path planning method of a hierarchical SDN network of claim 5, wherein: when the optimal path selection is performed, the optimal path selection can be performed after all available paths are calculated; the optimal path can also be selected for all available paths of the city and the area before the computing device, and then the final selection can be performed on the basis of the optimal paths of the city and the area after the optimal paths pass through all available paths after the computing device.

7. A system using a path planning method of a hierarchical SDN network as claimed in any of claims 1-6, characterized in that: comprising the following steps:

a network topology layering module: dividing the network topology into N layers, wherein the N-layer network topology comprises a plurality of first layers, each first layer comprising a plurality of second layers, and the N-th layer comprising a plurality of devices;

and a path calculation module: for calculating all available paths between each n+1 layer within the n-th layer; and calculating all available paths between the nth layer and the n-1 th layer; and calculating all available paths between the nth layer and each of the n+1th layers; calculating all available paths between the plurality of first layers and between the plurality of devices within the nth layer;

an optimal path generation module: and the method is used for splicing available paths and calculating the optimal path between any two devices.

8. A computer-readable storage medium storing a computer program, characterized by: the computer program, when executed by a processor, implements a path planning method of a hierarchical SDN network as claimed in any one of claims 1-6.