WO2019134483A1 - 多维约束下路径计算方法、装置、处理器及存储介质 - Google Patents

多维约束下路径计算方法、装置、处理器及存储介质 Download PDF

Info

Publication number
WO2019134483A1
WO2019134483A1 PCT/CN2018/120350 CN2018120350W WO2019134483A1 WO 2019134483 A1 WO2019134483 A1 WO 2019134483A1 CN 2018120350 W CN2018120350 W CN 2018120350W WO 2019134483 A1 WO2019134483 A1 WO 2019134483A1
Authority
WO
WIPO (PCT)
Prior art keywords
path
service
constraint
path calculation
dimensional
Prior art date
Application number
PCT/CN2018/120350
Other languages
English (en)
French (fr)
Inventor
陆钱春
李锋
严峰
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Priority to EP18898081.7A priority Critical patent/EP3720064A4/en
Publication of WO2019134483A1 publication Critical patent/WO2019134483A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/12—Shortest path evaluation
    • H04L45/125—Shortest path evaluation based on throughput or bandwidth
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0895—Configuration of virtualised networks or elements, e.g. virtualised network function or OpenFlow elements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12—Discovery or management of network topologies
    • H04L41/122—Discovery or management of network topologies of virtualised topologies, e.g. software-defined networks [SDN] or network function virtualisation [NFV]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/40—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using virtualisation of network functions or resources, e.g. SDN or NFV entities
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/12—Shortest path evaluation
    • H04L45/124—Shortest path evaluation using a combination of metrics
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/24—Multipath
    • H04L45/247—Multipath using M:N active or standby paths
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/20—Arrangements for monitoring or testing data switching networks the monitoring system or the monitored elements being virtualised, abstracted or software-defined entities, e.g. SDN or NFV

Definitions

  • the present invention relates to, but is not limited to, the field of communications, and in particular, to a path calculation method, apparatus, processor and storage medium under multi-dimensional constraints.
  • SDN Software Defined Network
  • the SDN controller is the middle layer in the three-tier architecture. It plays a key role in the connection between the control plane and the forwarding plane. The control plane completely controls the forwarding behavior, making the device completely white.
  • the business application layer is the top layer of the SDN network. It is an application that implements security, management, and other special functions with a controller. Therefore, the application layer is the direction controlled by the SDN controller.
  • the SDN controller bundles the asset storage, service module and path calculation into the controller, and the coupling is tight.
  • the scenario is complex and the application is numerous, the amount of data is huge, and it is difficult to solve the problem of low efficiency of the controller.
  • the SDN controller still determines certain two or several constraints for the path calculation under multi-dimensional constraints, such as adding bandwidth constraints, delay constraints, prioritizing paths, etc. in the multi-dimensional constraint set. To some extent, it satisfies the business needs, but it is difficult to adapt to more and more complex business scenarios.
  • the concept of searching for the shortest path for the optimal path is not the case.
  • the product metric often occurs, for example, the reliability is the product of each point; the maximum and minimum values, for example, the remaining bandwidth is maximized.
  • the calculation of the cost of each side is still a linear combination, and the actual situation is far more complicated than the linear combination, for example, the cost of the edge is calculated according to the priority; also, the active and standby separation paths are not considered, and the shared risk link group is shared. , SRLG) shared risk link groups and other important application scenarios.
  • the service module of the SDN controller is coupled with the path calculation module and cannot be applied to the expansion and change of the new service scenario.
  • the path calculation module cannot be used as a separate component for other service scenarios, resulting in lower efficiency of the SDN controller.
  • the fixed constraints used result in bottlenecks in path computation performance.
  • the embodiments of the present invention are directed to providing a path calculation method, apparatus, processor, and storage medium under multi-dimensional constraints, so as to at least solve the coupling of the service module and the path calculation module of the SDN controller in the related art, and cannot be applied to the new one.
  • the expansion and change of the service scenario and the path calculation module cannot be used as a separate component for other business scenarios, resulting in a problem of low efficiency of the SDN controller and a problem of path computation performance bottleneck caused by fixed constraints used in the path calculation process.
  • the embodiment of the present invention provides a method for calculating a path under a multi-dimensional constraint, comprising: receiving a transaction request of a service, where the service scenario of the service includes multiple; the calculation request includes at least one of the following information: a constraint type corresponding to the service, an evaluation type corresponding to the service; matching the constraint type with a preset multi-dimensional constraint model to obtain a multi-dimensional constraint set corresponding to the service; The evaluation model is matched to obtain an evaluation function corresponding to the service; and the path calculation is performed according to the multi-dimensional constraint set and the evaluation function.
  • the preset multi-dimensional constraint model includes at least one of the following: a tenant identity ID, a multi-dimensional constraint set; the preset evaluation model includes at least one of the following: a next hop evaluation of the network topology, and a network topology.
  • the edge cost estimate for each edge includes at least one of the following: a tenant identity ID, a multi-dimensional constraint set; the preset evaluation model includes at least one of the following: a next hop evaluation of the network topology, and a network topology.
  • the method further includes: constructing a network topology model, where the network topology model includes at least one of the following information: a topology node set, a topology chain Road set, tenant set.
  • the method further includes: detecting whether the network topology information changes; and when detecting that the network topology information changes, updating the network topology model.
  • performing path calculation according to the multi-dimensional constraint set and the evaluation function includes: injecting the multi-dimensional constraint set in a process of traversing the network topology diagram model; and using the multi-dimensional constraint set for cooperation In the next hop selection process of the network topology model, the evaluation function is injected in the next hop selection process and the edge cost selection process of the network topology model to implement the path calculation.
  • the method before the path calculation is performed according to the multi-dimensional constraint set and the evaluation function, the method further includes: searching for an algorithm corresponding to the path calculation, wherein the algorithm includes at least one of the following: a multi-dimensional constraint before K The optimal path TopK algorithm, the multi-dimensional constrained shared risk link group (SRLG) algorithm, and the multi-dimensional constrained shortest active/standby separation path algorithm.
  • the algorithm includes at least one of the following: a multi-dimensional constraint before K The optimal path TopK algorithm, the multi-dimensional constrained shared risk link group (SRLG) algorithm, and the multi-dimensional constrained shortest active/standby separation path algorithm.
  • SRLG multi-dimensional constrained shared risk link group
  • the method further includes: detecting whether the service changes in real time; and re-searching for an algorithm corresponding to the path calculation when the service changes.
  • the method further includes: receiving resource information optimized by the performance analysis system, where the resource information is used for path calculation.
  • the method further includes: issuing a calculation result of the path calculation.
  • the embodiment of the invention further provides a path computing device under multi-dimensional constraints, comprising:
  • the receiving module is configured to receive a transaction request of the service, where the service scenario of the service includes multiple; the computing request includes at least one of the following information: a constraint type corresponding to the service, and the service Corresponding evaluation type; the first obtaining module is configured to match the constraint type and the preset multi-dimensional constraint model to obtain a multi-dimensional constraint set corresponding to the service; and the second obtaining module is configured to: And matching with the preset evaluation model to obtain an evaluation function corresponding to the service; and a calculation module configured to perform path calculation according to the multi-dimensional constraint set and the evaluation function.
  • the preset multi-dimensional constraint model includes at least one of the following: a tenant identity ID, a multi-dimensional constraint set; the preset evaluation model includes at least one of the following: a next hop evaluation of the network topology, and a network topology.
  • the edge cost estimate for each edge includes at least one of the following: a tenant identity ID, a multi-dimensional constraint set; the preset evaluation model includes at least one of the following: a next hop evaluation of the network topology, and a network topology.
  • the apparatus further includes: a building module configured to construct a network topology model before performing path calculation according to the multi-dimensional constraint set and the evaluation function, wherein the network topology model includes at least the following information One: topology node set, topology link set, tenant set.
  • the device further includes: a detecting module configured to detect whether the network topology information changes; and an updating module configured to update the network topology model when detecting that the network topology information changes.
  • the calculation module includes: an injection unit configured to inject the multi-dimensional constraint set in a process of traversing the network topology diagram model; and a calculation unit configured to cooperate with the multi-dimensional constraint set In the next hop selection process of the network topology model, the evaluation function is injected in the next hop selection process and the edge cost selection process of the network topology model to implement the path calculation.
  • the apparatus further includes: a first search module, configured to search for an algorithm corresponding to the path calculation before performing path calculation according to the multi-dimensional constraint set and the evaluation function, where the algorithm
  • the method includes at least one of the following: a K-optimal path TopK algorithm before the multi-dimensional constraint, a multi-dimensional constrained shared risk link group algorithm, and a multi-dimensional constrained shortest active/standby separation path algorithm.
  • the device further includes: a detecting module configured to detect whether the service changes in real time; and a second searching module configured to re-search for an algorithm corresponding to the path calculation when the service changes.
  • the apparatus further includes: a receiving module, configured to receive resource information optimized by the performance analysis system before performing path calculation according to the multi-dimensional constraint set and the evaluation function, wherein the resource information is used by the resource information For path calculation.
  • the device further includes a sending module configured to deliver the calculation result of the path calculation.
  • the embodiment of the present invention further provides a storage medium, where the storage medium includes a stored program, wherein the program is executed to execute the multi-dimensional constraint path calculation method provided by the embodiment of the present invention.
  • a processor configured to execute a program, wherein the program is executed to execute a multi-dimensional constraint path calculation method provided by an embodiment of the present invention.
  • the embodiment of the invention further provides a path computing device under multi-dimensional constraints, comprising:
  • a memory configured to store a path calculation under multidimensional constraints
  • the processor is configured to run the program, wherein the program is executed to execute the multi-dimensional constraint path calculation method provided by the embodiment of the present invention.
  • the embodiment of the present invention is configured to receive a transaction request of a service, where the service scenario of the service includes multiple; the calculation request includes at least one of the following information: a constraint type corresponding to the service, and an evaluation corresponding to the service. a type; matching the constraint type with a preset multi-dimensional constraint model to obtain a multi-dimensional constraint set corresponding to the service; matching the evaluation type with a preset evaluation model to obtain an evaluation function corresponding to the service; The multidimensional constraint set and the evaluation function perform path calculation.
  • any kind of service scenario can perform path calculation, and the service module of the SDN controller in the related technology is solved.
  • Coupling with the path calculation module cannot be applied to the expansion and change of new service scenarios, and the problem that the path calculation module cannot be used as a separate component for other business scenarios, resulting in low efficiency of the SDN controller, and through the multidimensional constraint model, Any constraint set can be constructed without being closely related to the business, which can be fully extended to meet the complex business constraint requirements in the future, solve the problem of path computation performance bottleneck, and realize the technical effect of network-wide resource optimization.
  • FIG. 1 is a flowchart of a path calculation method under multi-dimensional constraints according to an embodiment of the present invention
  • FIG. 2 is a diagram showing an interaction relationship between an SDN controller, a performance analysis system, a scheduler, and an asset library according to an embodiment of the present invention
  • FIG. 3 is a schematic structural diagram of a module of a performance analysis system according to an embodiment of the present invention.
  • FIG. 4 is a schematic structural diagram of an area optimization module of a performance analysis system according to an embodiment of the present invention.
  • FIG. 5 is a schematic diagram of a topology information collection component according to an embodiment of the present invention.
  • FIG. 6 is a schematic diagram of a topology node information kit provided by an embodiment of the present invention.
  • FIG. 7 is a schematic diagram of a topology link information suite provided by an embodiment of the present invention.
  • FIG. 8 is a schematic diagram of a topology link attribute information kit provided by an embodiment of the present invention.
  • FIG. 9 is a topological example model of a tenant city information model provided by an embodiment of the present invention.
  • FIG. 10 is a CIM network topology model according to an embodiment of the present invention.
  • FIG. 11 is a schematic diagram of a single constraint basic information suite provided by an embodiment of the present invention.
  • FIG. 12 is a schematic diagram of a constraint unit information kit according to an embodiment of the present invention.
  • FIG. 13 is a schematic diagram of a constraint set information suite provided by an embodiment of the present invention.
  • FIG. 14 is a schematic diagram of a constraint type information kit provided by an embodiment of the present invention.
  • FIG. 17 is a flowchart of a path computation workflow mechanism provided by an embodiment of the present invention.
  • FIG. 18 is a flowchart of adapting before path calculation starts according to an embodiment of the present invention.
  • FIG. 19 is a flowchart of analyzing a multi-dimensional constraint model according to an embodiment of the present invention.
  • FIG. 20 is a flowchart of a path calculation process adaptation according to an embodiment of the present invention.
  • 21 is a flowchart of adapting after the path calculation ends according to the embodiment of the present invention.
  • 22 is a flowchart of a multi-dimensional constraint path computation traversal algorithm according to an embodiment of the present invention.
  • 25 is a flowchart of a multi-dimensional constraint SRLG path algorithm according to an embodiment of the present invention.
  • 26 is a flowchart of a multi-dimensional constrained optimal active/standby separation path algorithm according to an embodiment of the present invention.
  • FIG. 27 is a structural block diagram of a multi-dimensional constraint path calculation apparatus according to an embodiment of the present invention.
  • FIG. 28 is a structural block diagram of a multi-dimensional constraint path calculation apparatus according to an embodiment of the present invention.
  • the asset library is used to store resource information, such as nodes, links, ports, tunnels, pseudowires, and related attribute information. It mainly provides an asset cache interface for the SDN controller.
  • the Performance Analysis System (PAS), a performance analysis system, provides an external policy system for the orchestrator and provides performance analysis strategies, including network analysis and network optimization.
  • the orchestrator mainly provides policy management (constrained strategy, evaluation strategy) and workflow mechanism for the SDN controller.
  • SDN controller An important function of the SDN controller is to perform path calculation and manage the path, and finally complete the path delivery.
  • FIG. 1 is a flowchart of a path calculation method under the multi-dimensional constraint provided by the embodiment of the present invention. As shown in FIG. 1 , the process includes the following steps:
  • Step S102 Receive a transaction request of the service, where the service scenario of the service includes multiple; the calculation request includes at least one of the following information: a constraint type corresponding to the service, and an evaluation type corresponding to the service;
  • Step S104 matching the constraint type with the preset multi-dimensional constraint model to obtain a multi-dimensional constraint set corresponding to the service;
  • Step S106 matching the evaluation type and the preset evaluation model to obtain an evaluation function corresponding to the service
  • Step S108 performing path calculation according to the multi-dimensional constraint set and the evaluation function.
  • the execution body of the foregoing steps may be an SDN controller or the like, but is not limited thereto.
  • the foregoing multi-dimensional constraint model includes at least one of the following: a tenant identity ID and a multi-dimensional constraint set; and the foregoing preset evaluation model includes at least one of the following: a next hop evaluation of the network topology, and a network topology.
  • a transaction request of the service where the type of the service includes a plurality of; the calculation request includes at least one of the following information: a constraint type corresponding to the service, and a correspondence corresponding to the service
  • the evaluation type is matched with the preset multi-dimensional constraint model to obtain a multi-dimensional constraint set corresponding to the service; the evaluation type is matched with the preset evaluation model to obtain an evaluation function corresponding to the service;
  • the multidimensional constraint set and the evaluation function perform path calculation.
  • any kind of service scenario can perform path calculation, and the service module of the SDN controller in the related technology is solved.
  • Coupling with the path calculation module cannot be applied to the expansion and change of new service scenarios, and the problem that the path calculation module cannot be used as a separate component for other business scenarios, resulting in low efficiency of the SDN controller, and through the multidimensional constraint model, Any constraint set can be constructed without being closely related to the business, which can be fully extended to meet the complex business constraint requirements in the future, solve the problem of path computation performance bottleneck, and realize the technical effect of network-wide resource optimization.
  • This example is based on the SDN network to provide a method for optimal path calculation strategy under multi-dimensional constraints.
  • the business application is decoupled from the path calculation algorithm.
  • the asset information storage is decoupled from the controller core, which solves the current inefficient operation of the SDN controller.
  • the performance analysis system data analysis optimization result is used as the premise of path calculation, driven by strategy, constructing multi-dimensional constraint model and evaluation model; using workflow mechanism, multi-dimensional constraint model analysis, evaluation function injection, decoupling Multi-dimensional constraint set and traversal algorithm, decoupling optimal path evaluation and traversal algorithm; at the same time, the construction of multi-dimensional constraint model solves the problem that the multi-dimensional constrained optimal path computation cannot be flexibly dealt with in complex business scenarios, and achieves the goal of network-wide resource optimization.
  • the method for optimal path calculation strategy under the multi-dimensional constraint provided by the example includes the following steps:
  • Step S11 The performance analysis system performs refined analysis and optimization on the performance data, and provides the optimization result to the orchestrator policy module and the controller path calculation module;
  • Step S12 the orchestrator provides policy management, constructs a multidimensional constraint model, and constructs an evaluation model
  • Step S13 The controller provides a context switch entry, uses a workflow mechanism to drive the path calculation process, and associates with the orchestrator; performs path calculation before the search adaptation: parsing the multi-dimensional constraint model, adapting the search algorithm, selecting the return path type, and performing scene switching. Detection
  • Step S14 The controller path calculates the scene adaptation in the search process: the path search calculation, the injection evaluation model, and the scene switching detection;
  • Step S15 The controller adapts the scene after the search is completed: the path analysis, the cache, the scene switch detection, and the path are sent to the device.
  • the performance analysis system provides the basic performance data for path calculation by analyzing and optimizing the refined resources of the service, to solve the situation that the future network is more and more complex, and the resource utilization is insufficient, and the network resources are more rationally planned;
  • the application adapts the business strategy, constructs the constraint model, constructs the evaluation model injection mode to select the optimal node for the next hop, fully satisfies the needs of the flexible and varied business scenarios, and makes the controller control the network easier;
  • the controller adapts the scenario change to perform context scenario switching, and can update the topology information in time, trigger the recalculation mechanism, and timely prevent and control the user experience caused by network oscillation;
  • the controller shortens the convergence time of the optimal path calculation by providing a general multi-constraint path search algorithm, fully satisfies the path calculation under the arbitrary constraint set, and makes the controller more convenient to plan and manage the path.
  • Step S21 Construct a network topology model, where the network topology model includes at least one of the following information: a topology node set, a topology link set, and a tenant set.
  • the above method further comprises the following steps:
  • Step S31 detecting whether the network topology information changes
  • Step S32 updating the network topology model when detecting that the network topology information changes.
  • the network topology model can be updated in time to meet the needs of the flexible and varied business scenarios.
  • performing the path calculation according to the multidimensional constraint set and the evaluation function comprises the following steps:
  • Step S41 injecting the multi-dimensional constraint set in the process of traversing the network topology map model
  • Step S42 the multi-dimensional constraint set is used in the next hop selection process of the network topology model, and the evaluation function is injected in the next hop selection process and the edge cost selection process of the network topology model to implement the Path calculation.
  • Step S51 Search for an algorithm corresponding to the path calculation, where the algorithm includes at least one of the following: a K-optimal path TopK algorithm before multi-dimensional constraint, a multi-dimensional constraint shared risk link group SRLG algorithm, and a multi-dimensional constraint shortest active/standby separation path algorithm.
  • the algorithm includes at least one of the following: a K-optimal path TopK algorithm before multi-dimensional constraint, a multi-dimensional constraint shared risk link group SRLG algorithm, and a multi-dimensional constraint shortest active/standby separation path algorithm.
  • the method further includes:
  • Step S61 detecting whether the service changes in real time
  • step S62 when the service changes, the algorithm corresponding to the path calculation is re-searched.
  • the path calculation algorithm can be updated in time according to the change of the service.
  • Step S63 Receive resource information optimized by the performance analysis system, where the resource information is used for path calculation.
  • the performance analysis system provides basic performance data for path calculation by analyzing and optimizing the refined resources of the service, so as to solve the situation that the future network is more and more complex, and the resource utilization is insufficient, and the network resources are more rationally planned.
  • the method further includes: issuing a calculation result of the path calculation.
  • the performance analysis system communicates with the asset library data through a separate interface.
  • the performance analysis system collects, calculates, analyzes, and returns the optimized data to the orchestrator and controller.
  • the orchestrator policy module communicates with the performance analysis system through a separate interface, and after obtaining the optimized data, the optimized data is applied to the policy management.
  • the controller path calculation module communicates with the performance analysis system through the independent interface, and after obtaining the optimization data, the optimization data is applied to the path calculation.
  • the controller internally caches the assets through communication between the independent interface and the asset library data, and prepares for constructing the global topology network model.
  • the controller communicates with the orchestrator policy module through a separate interface to provide a strategy for controller path calculation; the controller performs path calculation according to the policy.
  • the data acquisition module collects the network resources of the network level where the performance analysis system is located through the asset library;
  • the data calculation module aggregates and calculates the data in the time dimension and the resource dimension according to the data collected by the data acquisition module;
  • the data analysis module analyzes the data calculated by the data calculation module, and analyzes the data required by the user in the time dimension, the resource dimension, and the user-defined dimension;
  • the data display module is displayed in various dimensions according to the analysis data of the data analysis module, and supports exporting and editing;
  • the domain optimization module obtains an optimization strategy according to the analysis data of the data analysis module, and the strategy includes network analysis, network optimization, path calculation, and reports the policy to the corresponding orchestrator.
  • the composition of the domain optimization module is as shown in FIG. 4, and includes a network analysis module, a network optimization module, and a path calculation module.
  • FIG. 5 to FIG. 8 are diagrams of the kit information and the relationship between the kits in the process of creating a CIM network abstract topology model according to the embodiment.
  • Figure 5 is a topology information collection component relationship. It mainly describes the source and destination of topology information collection, caches topology information, and provides topology for path calculation. The functions of each component are as follows:
  • Asset library Provide asset data from the external interface of the asset library
  • controller asset cache component the asset data cache is performed inside the controller
  • Event distribution component Send an event to notify the current network that there is topology information change, such as node change, link change, tenant change;
  • Event parsing component parsing the current topology event, analyzing and extracting useful information to prepare for the topology cache.
  • Topology cache component Provides basic attribute information for the topology cache component by asset data, dynamically constructs a CIM abstract network topology model, and caches topology information of nodes, links, tenants and related attribute information in the topology model to provide search services. Topology model data preparation.
  • FIG. 6 is an abstract network topology node information suite provided by the embodiment, where the node abstract information includes an ID, a current version number, and a node calculation related attribute information.
  • the node ID is the globally unique node ID identifier in the whole network abstract model.
  • the version number information is mainly used to determine whether the network node has been changed. When the related node is changed (mainly the node goes online or offline), the information is sent before. It may have expired; the attribute information mainly extracts attributes related to the search calculation, such as delay information, and is mainly used for the path optimization calculation service.
  • FIG. 7 and FIG. 8 are the abstract network link information suites provided by the embodiment, which are mainly link abstract information in the model, and are as follows:
  • the topology link extraction information includes an ID, a current version number, a source node (ingress node), a destination node (outgoing node), and attribute information.
  • the link ID is the global unique node ID identifier in the whole network abstract model. If the link IDs of the same source node and the destination node are different, different links are represented; therefore, different physical ports (physical or logical) of the two physical nodes can be distinguished. Multiple links are mapped without the need to extract port information.
  • the source node and the destination node are the information extracted in Figure 6.
  • the version number information is mainly used to determine whether the network link has been changed (mainly for link discovery, link disconnection, attribute value change, etc.). After the related link is changed, the previously sent information may have expired. ;
  • Link attribute information mainly extracts attributes related to search calculation, such as bandwidth-related information (bandwidth capacity, used bandwidth, bandwidth utilization, etc.), delay, load, reliability, etc., and is mainly used for path-based computing services. .
  • FIG. 9 is an abstract network topology example model of a Common Information Modeling (CIM) actually used by a tenant according to the embodiment. Including: tenant ID, change time, sub-topology model, and other parameters.
  • tenant ID is the global unique identifier of the different tenants.
  • the change time mainly records the time when the topology is changed.
  • the sub-topology model mainly includes the subset of the entire network topology node and the subset of links, which are respectively the entire network.
  • the local reference of the node set and the link set; finally, other parameters mainly record some other information of the user, such as the number of nodes, the number of links and other reference information.
  • FIG. 10 is a CIM abstract topology model of the entire network provided by the embodiment.
  • the model mainly provides information needed for an abstract topology network, caches information, and provides a data model foundation for the topology search service and the upper layer service model. It mainly includes: node collection, link collection, and tenant collection.
  • node collection is the whole network node information set shown in FIG. 6.
  • link set is the full network link information set shown in FIG. 6.
  • the tenant collection is the set of topology instances (tenant collections) shown in Figure 9.
  • FIG. 11 through 14 are diagrams of the constraint model provided by the present invention for the organizer policy module service set. It mainly serves the construction of constraint sets under arbitrary business requirements, adapting to the complexity and flexibility of business scenarios. At the same time, a constraint resolution entry is provided for the topology search service for searching process constraint processing. The content is as follows:
  • FIG. 11 provides a single constraint basic information suite according to the embodiment; the kit mainly includes three objects: a constraint value, a constraint upper limit, and a constraint lower limit, and is mainly used to describe basic information of a single constraint.
  • all business constraints can be described by one or a combination of the three. For example, bandwidth utilization is described by upper and lower limits; hop count, delay, necessary point, etc. can be described by constraint values.
  • FIG. 12 is a constraint unit information kit provided by this embodiment.
  • the kit mainly contains constraint names and a set of constraint values.
  • a constraint unit such as a set of necessary points, will contain multiple sets of nodes; and, if necessary, a set of links, including multiple sets of links. Therefore, the constraint unit needs to be described by the constraint value set graph 11, and the constraint unit is to provide a name to identify the constraint meaning.
  • the kit mainly includes: a constraint type, a constraint unit (shown in Figure 12) collection.
  • a constraint type mainly serves the path calculation search service, and is used to determine the constraint action point identifier, for example, the Graph type constraint, which mainly acts on the path calculation process, such as the path hop limit of all nodes, the delay limit, the service application bandwidth, and the bandwidth utilization.
  • Path type constraints are mainly used for path management and optimization; for example, TopK path, SRLG active/standby path, active/standby separation path, and equivalent balanced path.
  • Node type constraints and Connection type constraints such as delay, bandwidth, load, reliability, etc., are mainly used in the node traversal optimization process.
  • FIG. 15 is a multidimensional constraint model (MRM) of the multi-dimensional constraint model provided by the embodiment.
  • the MRM model is mainly composed of a tenant ID and a multidimensional constraint set (S33).
  • S33 multidimensional constraint set
  • FIG. 14 shows.
  • the 16 is an evaluation model provided by the embodiment, an evaluation model (EM, Evaluation Model), which mainly serves an evaluation function structure under any service requirement, adapts to the complexity and flexibility of the business scenario, and is used for the evaluation function in the search process. injection.
  • the next evaluation is used to calculate the cost calculation function for each edge in the topology.
  • the calculation of the edge can be obtained by prioritizing the consideration indicators on the edge; or the measurement indicators can be arranged according to a certain coefficient ratio. Combinations; can also be constructed for other business needs.
  • the edge cost evaluation is used to construct the next hop node selection function for the topology calculation.
  • the next hop node may be the addition and the minimum, or the product may be the smallest; or the maximum; or constructed according to the actual business needs.
  • FIG. 17 is a flow chart of a path calculation search service workflow mechanism provided by the embodiment, which mainly includes a topology event resolution process, a topology local cache, an orchestrator policy module, a service constraint model, an orchestrator policy module, a service evaluation model, and a path calculation search.
  • Step S1701 Parse the topology event, construct a CIM topology model, and perform local caching.
  • Step S1702 The orchestrator policy module performs constraint set construction.
  • the orchestration policy module collects the business with the request for calculation, it constructs the constraint set according to the business needs and the constraint model construction mode.
  • Step S1703 Constructing an evaluation function model, and the path calculation search service injects an evaluation function in the search process.
  • the orchestration strategy module collects the business request
  • the evaluation function model is constructed according to the business needs
  • the path calculation search service injects the evaluation function in the search process.
  • Step S1704 Perform path calculation search adaptation conversion, and initialize the service incoming parameters.
  • Step S1705 The path calculation pre-search adaptation, as shown in FIG. 18, the flow includes steps S1801 to S1804.
  • Step S1706 The path calculation is adapted during the search process, as shown in FIG. 20, and the flow includes steps S2001 to S2007.
  • Step S1707 After the path calculation search ends, the adaptation is as shown in FIG. 21, and the flow includes step S2101 and step S2103.
  • FIG. 18 is a flow chart of the path calculation search algorithm before starting the adaptation process according to the embodiment. Implementation steps include:
  • Step S1801 Multidimensional constraint model analysis.
  • Step S1802 Determine the path return type by parsing of the multidimensional constraint set.
  • the path types mainly include: optimal path, equal-balanced path, TopK path, SRLG shared risk link group active/standby path, and optimal separated active/standby path.
  • Step S1803 The search algorithm determines.
  • the search algorithm mainly includes: general search algorithm, SRLG algorithm, and separation path algorithm.
  • Step S1804 Scene switching detection.
  • the pre-search adaptation After the pre-search adaptation is complete, you need to detect changes in the scenario and changes in the scenario. For example, the topology has changed and the service request has changed.
  • the scene change directly affects the execution actions of the subsequent steps: re-execute the pre-search adaptation work, directly end the return, and perform the subsequent process.
  • FIG. 19 is a flowchart of parsing related parameters under different constraint types according to the embodiment. It mainly includes graph type constraint parameters, path type constraint parameters, connection type constraint parameters, and node type constraint parameters.
  • the resolution of all constraints is adapted by the configuration name, and each constraint resolution resolves only the multiple constraint sets contained in the current request.
  • Multidimensional constraint model analysis mainly serves the search algorithm and injects constraints into the algorithm.
  • the implementation steps are as follows:
  • Step S1901 Parsing the chart type constraint in the constraint model.
  • Step S1902 Parsing the path type constraint in the constraint model.
  • Step S1903 Parse the connection point type constraint in the constraint model.
  • the weight, bandwidth utilization, bandwidth capacity, load, reliability, and delay are included;
  • the Connection type constraint data and the Node type constraint data in step S1904 are mainly provided by a third party such as a cache or a database, and the topology service and the service module are provided. Share third-party resources. In this way, the impact of the calculation result failure caused by the data update can be sufficiently weakened; at the same time, the service module does not need to prepare data for the topology module, and the controller operation efficiency is also improved.
  • Step S1904 Parse the node type constraint in the constraint model.
  • FIG. 20 is a flow chart of a path calculation search process adaptation process according to the embodiment. The implementation steps are as follows:
  • Step S2001 Clone a topology instance that currently has a path request.
  • the purpose of cloning is to record the currently calculated topology object and the version number of the link object and the node object, so as to facilitate subsequent comparison of the topology, and whether the current calculation result is meaningful.
  • Step S2002 Check the initialization parameters to perform abnormal protection.
  • the request source node and the target node are empty.
  • Step S2003 Evaluation function initialization.
  • This step is mainly to prepare for the search traversal algorithm, and provides the criteria for judging the nodes and links.
  • Step S2004 Traversing the algorithm search.
  • the traversal algorithm search process is performed (see step S2201-step S2206 in FIG. 22 below), the cost of the source point to all nodes is calculated, and the optimal path data is returned.
  • Step S2005 Organizing the returned result data and returning the required path type.
  • Step S2006 Comparison of the path version numbers.
  • the version number of the link and node list in the return path is compared. It is found whether the topology has a link or node change of the related path. If it is an influential change, it needs to be recalculated.
  • Step S2007 scene switching detection.
  • the search process scene is detected.
  • the corresponding result action is executed: ending the calculation, recalculating the path, and continuing the process.
  • FIG. 21 is an adaptation procedure after the path calculation search algorithm provided by the embodiment ends. Implementation steps include:
  • Step S2101 path analysis, cache.
  • the path needs to be analyzed, the path is cached, and when the number of nodes is small, the path redundancy calculation is performed.
  • Step S2102 The path is switched, converted into a transmission object, and returned to the service module.
  • the path is composed of a list of links, and the link exists in the form of a dedicated abstract object in the topology search service module. It can be either physical or logical, so a certain conversion is required. In order to reduce the transmission consumption between modules before the path is sent, the path is transmitted through the link ID identification list.
  • Step S2103 Scene switching detection.
  • the scene is detected after the search ends, and when the scene changes, the corresponding result action is performed: recalculating the path and continuing the process.
  • FIG. 22 is a flow chart of the traversal algorithm provided in the embodiment, and FIG. 22 to FIG. It mainly includes the following algorithms: multi-dimensional constrained TopK path algorithm, multi-dimensional constrained SRLG algorithm, multi-dimensional constrained shortest separation path algorithm, multi-dimensional constrained optimal path algorithm. Implementation steps include:
  • Step S2201 Search for an algorithm entry, and the constraint set is passed in.
  • the multi-dimensional constraint set parameters are provided by the above steps, and enter the search algorithm entry;
  • Step S2202 Determine whether the active/backed path.
  • the path constraint parameter is used to determine whether the active/standby path calculation is needed. If the primary and backup paths are not required, the multi-dimensional constraint TopK path algorithm flow S2204 is entered.
  • Step S2203 Determine whether there is a SRLG link group.
  • the active/standby path it is determined whether there is a SRLG shared risk link group; if there is a SRLG link group, the multi-dimensional constraint SRLG algorithm flow S2203 is entered; otherwise, the multi-dimensional constraint shortest separation path algorithm flow S2206 is entered;
  • Step S2204 Multi-dimensional constraint TopK path algorithm flow.
  • step S2401 See step S2404, step S2404 of FIG.
  • Step S2205 Multidimensional constrained SRLG algorithm flow.
  • Step S2206 The multi-dimensional constrained optimal active/standby separation path algorithm flow.
  • FIG. 23 is a flowchart of an optimal path algorithm provided by the embodiment, which mainly provides an optimal path and an equivalent balanced path calculation for other algorithm flows, and the implementation steps include:
  • Step S2301 Initial node initialization.
  • Performing traversal search preparation on the topology map for which the optimal path calculation is to be performed first preparing the current traversal node (the node traversed for the first time is the source node);
  • Step S2302 traverse all the outgoing edges corresponding to the starting node.
  • Step S2303 Perform constraint screening on the outbound node.
  • the outbound nodes of the outbound edge are constrained and filtered.
  • the outbound node mainly determines whether the node has a delay limit, whether it is a avoidance point or a necessary point; the outbound node that does not meet the constraint and the corresponding edge are added to the exclusion list; And the edge does not perform the next optimal node traversal election;
  • Step S2304 Perform constraint screening on the outgoing edge.
  • the outbound edge mainly judges whether the bandwidth, bandwidth utilization, delay, load, reliability, avoidance side list, and bound list are constrained; the outbound edge does not meet the constraint and the corresponding The exit nodes are added to the exclusion list.
  • Step S2305 Calling the outbound node cost calculation function to calculate the cost of the outbound node corresponding to the constraint condition corresponding to the outbound node.
  • the out-of-node cost calculation function (generally a linear function combination, or a priority order, which gives the user the most flexible injection method), and calculates the non-exclusive list from the origin. The cost to the current out node.
  • Step S2306 Add the outbound node that satisfies the condition to the candidate list.
  • the outbound node that satisfies the condition and the corresponding cost are added to the candidate list or the update candidate column.
  • Step S2307 Call the node optimal evaluation function to select the next traversal starting node.
  • step S2308 Call the user-injected optimal node selection evaluation function (which can be the smallest sum, the smallest product, or the maximum or minimum value, which gives the user the most flexible injection method), and choose the next time.
  • the starting node of the traversal if all the nodes have been traversed, go to step S2308; otherwise, go to step S2301.
  • Step S2308 All node traversal is completed.
  • Step S2309 According to the cost (cost) of all nodes in the candidate column, perform optimal path organization, remove the loop, and organize the equivalence balanced path according to the number of optimal path requirements, and return the path (group), the algorithm The process ends.
  • FIG. 24 is a general multi-dimensional constrained optimal path algorithm search process provided by the embodiment, that is, a multi-dimensional constrained TopK optimal path algorithm; TopK defaults to 1, that is, only one or more equal-balanced paths are returned.
  • the implementation steps are as follows:
  • Step S2401 Optimal path algorithm flow.
  • Step S2402 Whether the optimal path satisfies the hop count and the delay limit, returns the path data organization, and removes the unsatisfied path.
  • Step S2403 Determine whether the TopK is less than or equal to the number of returned optimal paths, if it is greater than the next step S2404; otherwise, the algorithm flow ends.
  • Step S2404 The KSP algorithm suboptimally traverses the search to remove a link.
  • a link is removed from the source point, and a suboptimal path traversal search is performed, that is, the topology map loop step S2401 is performed in step S2404.
  • the shared risk link group refers to all links in the same group, when one of them has a fault, the rest Will be disconnected. Therefore, when the active and standby paths are calculated, all links belonging to the same risk group appear on one link. That is, the active and standby paths do not contain links of the same risk group.
  • Step S2501 Analyze the risk link group, and allocate according to the group; and simultaneously initialize the mutually exclusive Boolean object, and use the next genetic algorithm to distinguish the active and standby path risk groups.
  • Step S2502 Genetic algorithm process (initialization group generation, fitness calculation, selection operation, cross operation, mutation operation).
  • Step S2503 The intermediate active/standby path group generates an algorithm for calling an optimal path algorithm
  • Step S2504 Whether the optimal primary and backup path groups meet the limitation of hop count, delay, reliability, etc., remove the unsatisfied path group, and return the satisfied path group;
  • Step S2505 When the optimal SRLG master and path group cannot be selected, the sub-optimal path group is selected to return. If there is a TopK constraint, the TopK active/backed path group is returned, and the algorithm flow ends.
  • FIG. 26 is a flowchart of a multi-dimensional constrained optimal active/standby separation path algorithm provided by the present invention; the implementation steps are as follows:
  • Step S2601 Calculate the equivalence equilibrium optimal path and perform final path constraint detection, and return an optimal path that satisfies the condition.
  • step S2301 - step S2309 Using the multi-dimensional constrained optimal path algorithm (step S2301 - step S2309), one or more equal-balanced optimal paths are calculated, and final path constraint detection (hop count, delay, reliability, etc.) is performed, and the most satisfying condition is returned. Excellent path (group), looping;
  • Step S2602 Reverse constructing a new topology map for the multi-dimensional constrained optimal path
  • Step S2603 Calculating the optimal path for the second time
  • the optimal path (group) is calculated for the multi-dimensional constrained optimal path algorithm process, and the loop;
  • Step S2604 removing the duplicate path twice and reconstructing the topology map
  • Step S2605 Calculate the optimal path for the third time, and perform constraint detection of the final path
  • the multi-dimensional constrained optimal path algorithm process calculates the optimal path (group) for the third time, and performs constraint detection (hop count, delay, reliability, etc.) of the final path.
  • This path (group) is the main path; loop;
  • Step S2606 Removing the main path node and the edge, and reconstructing the topology map
  • Step S2607 calculating the optimal path for the fourth time, and performing constraint detection of the final path;
  • the multi-dimensional constrained optimal path algorithm process calculates the optimal path (group) for the fourth time; and performs final path constraint detection (hop count, delay, reliability, etc.), which is a backup path and a loop;
  • Step S2608 The active/standby path is reorganized to detect the TopK requirement. When there is a TopK requirement, the TopK active/backed path is returned, and the algorithm flow ends.
  • the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, by hardware, but in many cases, the former is A better implementation.
  • the technical solution of the present invention which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a storage medium (such as ROM/RAM, disk,
  • the optical disc includes a number of instructions for causing a terminal device (which may be a cell phone, a computer, a server, or a network device, etc.) to perform the methods described in various embodiments of the present invention.
  • the embodiment of the present invention further provides a multi-dimensionally constrained path computing device, which is configured to implement the foregoing embodiments and implementation manners, and has not been described again.
  • the term "module” may implement a combination of software and/or hardware for a predetermined function.
  • the apparatus described in the following embodiments is preferably implemented in software, hardware, or a combination of software and hardware, is also possible and contemplated.
  • FIG. 27 is a structural block diagram of a multi-dimensional constraint path calculation apparatus according to an embodiment of the present invention. As shown in FIG. 27, the apparatus includes:
  • the receiving module 2702 is configured to receive a transaction request of the service, where the service scenario of the service includes multiple; the computing request includes at least one of the following information: a constraint type corresponding to the service, corresponding to the service Type of assessment;
  • the first obtaining module 2704 is configured to match the constraint type and the preset multi-dimensional constraint model to obtain a multi-dimensional constraint set corresponding to the service;
  • the second obtaining module 2706 is configured to match the evaluation type and the preset evaluation model to obtain an evaluation function corresponding to the service;
  • a calculation module 2708 configured to perform path computation based on the multi-dimensional constraint set and the evaluation function.
  • the foregoing multi-dimensional constraint model includes at least one of the following: a tenant identity ID and a multi-dimensional constraint set; and the foregoing preset evaluation model includes at least one of the following: a next hop evaluation of the network topology, and a network topology.
  • a transaction request of a service where the type of the service includes a plurality of; the calculation request includes at least one of the following information: a constraint type corresponding to the service, and an evaluation corresponding to the service a type; matching the constraint type with a preset multi-dimensional constraint model to obtain a multi-dimensional constraint set corresponding to the service; matching the evaluation type with a preset evaluation model to obtain an evaluation function corresponding to the service;
  • the multidimensional constraint set and the evaluation function perform path calculation.
  • any kind of service scenario can perform path calculation, and the service module of the SDN controller in the related technology is solved.
  • Coupling with the path calculation module cannot be applied to the expansion and change of new service scenarios, and the problem that the path calculation module cannot be used as a separate component for other business scenarios, resulting in low efficiency of the SDN controller, and through the multidimensional constraint model, Any constraint set can be constructed without being closely related to the business, which can be fully extended to meet the complex business constraint requirements in the future, solve the problem of path computation performance bottleneck, and realize the technical effect of network-wide resource optimization.
  • This example is to solve the multi-dimensional constraint path calculation also includes a multi-dimensional constraint path search service device, including the following modules:
  • Data extraction module configured to dynamically obtain topology information, cache topology information, and create a CIM network topology model.
  • the policy adaptation module is configured to parse the service scenario and start the path calculation process.
  • Constraint parsing module configured to parse a multidimensional constraint set and provide a constraint parameter set for path computation
  • Evaluation parsing module configured to parse the evaluation function to provide a method for next hop optimization and edge cost calculation;
  • Computational module configured to traverse the idea as the basic principle, injecting a multidimensional constraint set in the traversal process, allowing the multidimensional constraint set to act in the next hop selection process; and injecting the corresponding evaluation in the next hop selection and edge cost calculation process Function, the evaluation function is selected according to the business strategy, and the multi-dimensional constrained optimal path calculation is completed; and the process is extended to the multi-dimensional constrained TopK algorithm, the multi-dimensional constrained SRLG active/standby path algorithm, and the multi-dimensional constrained shortest active/standby separation path algorithm to complete different services.
  • the path management module is configured to manage the path and return the obtained path to the policy module for delivery.
  • the apparatus further includes: a building module, wherein the building module is configured to construct a network topology model before performing path calculation according to the multi-dimensional constraint set and the evaluation function, wherein the network topology model includes the following At least one of the information: a topology node set, a topology link set, and a tenant set.
  • the apparatus further includes: a detecting module configured to detect whether the network topology information changes; and an updating module configured to update the network topology model when detecting that the network topology information changes.
  • a detecting module configured to detect whether the network topology information changes
  • an updating module configured to update the network topology model when detecting that the network topology information changes.
  • FIG. 28 is a structural block diagram of a multi-dimensional constraint path calculation apparatus according to an embodiment of the present invention.
  • the calculation module 2708 includes:
  • an injection unit 2802 configured to inject the multidimensional constraint set during the traversal of the network topology map model
  • the computing unit 2804 is configured to cooperate the multi-dimensional constraint set in the next hop selection process of the network topology model, and inject the evaluation in the next hop selection process and the edge cost selection process of the network topology model Function to implement the path calculation.
  • the apparatus further includes: a first search module configured to search for an algorithm corresponding to the path calculation before performing path calculation according to the multi-dimensional constraint set and the evaluation function, where the algorithm includes at least the following One: the K-optimal path TopK algorithm before the multi-dimensional constraint, the multi-dimensional constrained shared risk link group SRLG algorithm, and the multi-dimensional constrained shortest active-standby separation path algorithm.
  • a first search module configured to search for an algorithm corresponding to the path calculation before performing path calculation according to the multi-dimensional constraint set and the evaluation function, where the algorithm includes at least the following One: the K-optimal path TopK algorithm before the multi-dimensional constraint, the multi-dimensional constrained shared risk link group SRLG algorithm, and the multi-dimensional constrained shortest active-standby separation path algorithm.
  • the apparatus further includes: a detecting module configured to detect whether the service changes in real time; and a second searching module configured to re-search for an algorithm corresponding to the path calculation when the service changes.
  • the path calculation algorithm can be updated in time according to changes in the service.
  • the apparatus further includes: a receiving module, configured to receive, after the path calculation according to the multi-dimensional constraint set and the evaluation function, the resource information optimized by the performance analysis system, wherein the resource information is used for the path Calculation.
  • the performance analysis system provides basic performance data for path calculation by analyzing and optimizing the refined resources of the service, and solves the situation that the future network is more and more complex, and the resource utilization is insufficient, and the network resources are more rationally planned.
  • the apparatus further includes: a sending module configured to deliver a calculation result of the path calculation.
  • each of the above modules may be implemented by software or hardware.
  • the foregoing may be implemented by, but not limited to, the foregoing modules are all located in the same processor; or, the above modules are in any combination.
  • the forms are located in different processors.
  • the embodiment of the present invention further provides a storage medium, where the storage medium includes a stored program, wherein the program is executed to execute the multi-dimensional constraint path calculation method provided by the embodiment of the present invention.
  • the storage medium may be configured to store program code for performing the following steps:
  • the calculation request includes at least one of the following information: a constraint type corresponding to the service, and an assessment corresponding to the service Types of;
  • the foregoing storage medium may include, but is not limited to, a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.
  • ROM read-only memory
  • RAM random access memory
  • mobile hard disk a magnetic disk
  • optical disk a variety of media that can store program code.
  • the embodiment of the present invention further provides a processor configured to run a program, where the program is executed to execute the multi-dimensional constraint path calculation method provided by the embodiment of the present invention.
  • the above program is used to perform the following steps:
  • the calculation request includes at least one of the following information: a constraint type corresponding to the service, and an assessment corresponding to the service Types of;
  • the examples in this embodiment may refer to the examples described in the foregoing embodiments and the optional embodiments, and details are not described herein again.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明公开了一种多维约束下路径计算方法、装置、处理器及存储介质，方法包括：接收业务的算路请求，其中，该业务的业务场景包括多个；算路请求中包括：与业务对应的约束类型、与业务对应的评估类型；将约束类型和预设的多维约束模型进行匹配，得到与业务对应的多维约束集合；将评估类型和预设的评估模型进行匹配，得到与业务对应的评估函数；根据多维约束集合和评估函数，进行路径计算。

Description

多维约束下路径计算方法、装置、处理器及存储介质
相关申请的交叉引用
本申请基于申请号为201810005339.9、申请日为2018年01月03日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本申请作为参考。
技术领域
本发明涉及但不限于通信领域，尤其涉及一种多维约束下路径计算方法、装置、处理器及存储介质。
背景技术
软件定义网络(Software Defined Network，SDN)是一种新型网络结构，其中SDN控制器是三层架构中的中间层，起着承上启下的关键作用，而控制器的核心是控制面与转发面分离，控制面完全控制转发行为，让设备彻底成为白盒。业务应用层是SDN网络的顶层，它是一些用控制器实现安全、管理和其他特殊功能的应用程序。因此，应用层是SDN控制器控制的方向。
相关技术中存在以下问题：
首先SDN控制器将资产存储、业务模块、路径计算捆绑内置于控制器中，耦合紧密，当场景复杂并且应用繁多，数据量庞大，难以解决控制器运行效率低下问题。
其次，SDN控制器对于多维约束下的路径计算，依然是在确定的某两个或几个约束，例如在多维约束集里面加入了带宽约束，时延约束，对路径进行优先级排序等，这些在一定程度上满足了业务需求，但是难以适应越来越复杂的业务场景。
例如，对最优路径搜索最短路径的概念上，但实际情形并非如此，复 杂业务场景下时常出现乘积度量，例如可靠性为各点乘积；最大最小值，例如剩余带宽取最大等等。对各边代价的计算依然是线性组合，而实际情形远比线性组合复杂，例如边的代价按优先级来计算；还有，没有考虑主备分离路径，共享风险链路组(Shared Risk Link Groups，SRLG)共享风险链路组等重要的应用场景。
当网络不稳定，出现震荡，抖动断链等情形时，会导致一段时间的丢包，此时路由计算的收敛时间则是用户体验好坏的关键因素。而现行约束路径计算简单粗暴，简单的说就是对所有路径进行全排，再逐一筛选满足约束的，当网络过于庞大，约束复杂时，算路效率低下问题显而易见。
综上所述，相关技术中存在的问题如下：
SDN控制器的业务模块与路径计算模块耦合，无法适用于新的业务场景的扩展和变化；路径计算模块无法作为独立组件给其他业务场景使用，导致SDN控制器运行效率较低；路径计算过程中使用的约束固定导致路径计算性能出现瓶颈。
发明内容
有鉴于此，本发明实施例期望提供一种多维约束下路径计算方法、装置、处理器及存储介质，以至少解决相关技术中SDN控制器的业务模块与路径计算模块耦合，无法适用于新的业务场景的扩展和变化以及路径计算模块无法作为独立组件给其他业务场景使用所导致的SDN控制器运行效率较低的问题以及路径计算过程中使用的约束固定所导致的路径计算性能瓶颈的问题。
本发明实施例提供了一种多维约束下路径计算方法，包括：接收业务的算路请求，其中，所述业务的业务场景包括多个；所述算路请求中包括以下信息至少之一：与所述业务对应的约束类型、与所述业务对应的评估类型；将所述约束类型和预设的多维约束模型进行匹配，得到与所述业务 对应的多维约束集合；将所述评估类型和预设的评估模型进行匹配，得到与所述业务对应的评估函数；根据所述多维约束集合和所述评估函数，进行路径计算。
上述方案中，所述预设的多维约束模型包括以下至少之一：租户身份标识ID、多维约束集合；所述预设的评估模型包括以下至少之一：网络拓扑的下一跳评估、网络拓扑中每条边的边代价评估。
上述方案中，在根据所述多维约束集合和所述评估函数，进行路径计算之前，还包括：构建网络拓扑模型，其中，所述网络拓扑模型包括以下信息至少之一：拓扑节点集、拓扑链路集、租户集。
上述方案中，所述方法还包括：检测网络拓扑信息是否发生变化；在检测到所述网络拓扑信息发生变化时，更新所述网络拓扑模型。
上述方案中，根据所述多维约束集合和所述评估函数，进行路径计算包括：在对所述网络拓扑图模型进行遍历的过程中，注入所述多维约束集合；将所述多维约束集合作用于所述网络拓扑模型的下一跳选择过程中，并在所述下一跳选择过程中和所述网络拓扑模型的边代价选择过程中注入所述评估函数，以实现所述路径计算。
上述方案中，在根据所述多维约束集合和所述评估函数，进行路径计算之前，还包括：搜索与所述路径计算对应的算法，其中，所述算法包括以下至少之一：多维约束前K条最优路径TopK算法、多维约束共享风险链路组(SRLG，Share Risk Link Group)算法、多维约束最短主备分离路径算法。
上述方案中，所述方法还包括：实时检测所述业务是否发生变化；在所述业务发生变化时，重新搜索与所述路径计算对应的算法。
上述方案中，在根据所述多维约束集合和所述评估函数，进行路径计算之前，还包括：接收性能分析***优化后的资源信息，其中，所述资源信息用于路径计算。
上述方案中，所述方法还包括：下发所述路径计算的计算结果。
本发明实施例还提供了一种多维约束下路径计算装置，包括：
接收模块，配置为接收业务的算路请求，其中，所述业务的业务场景包括多个；所述算路请求中包括以下信息至少之一：与所述业务对应的约束类型、与所述业务对应的评估类型；第一获取模块，配置为将所述约束类型和预设的多维约束模型进行匹配，得到与所述业务对应的多维约束集合；第二获取模块，配置为将所述评估类型和预设的评估模型进行匹配，得到与所述业务对应的评估函数；计算模块，配置为根据所述多维约束集合和所述评估函数，进行路径计算。
上述方案中，所述预设的多维约束模型包括以下至少之一：租户身份标识ID、多维约束集合；所述预设的评估模型包括以下至少之一：网络拓扑的下一跳评估、网络拓扑中每条边的边代价评估。
上述方案中，所述装置还包括：构建模块，配置为在根据所述多维约束集合和所述评估函数，进行路径计算之前，构建网络拓扑模型，其中，所述网络拓扑模型包括以下信息至少之一：拓扑节点集、拓扑链路集、租户集。
上述方案中，所述装置还包括：检测模块，配置为检测网络拓扑信息是否发生变化；更新模块，配置为在检测到所述网络拓扑信息发生变化时，更新所述网络拓扑模型。
上述方案中，所述计算模块包括：注入单元，配置为在对所述网络拓扑图模型进行遍历的过程中，注入所述多维约束集合；计算单元，配置为将所述多维约束集合作用于所述网络拓扑模型的下一跳选择过程中，并在所述下一跳选择过程中和所述网络拓扑模型的边代价选择过程中注入所述评估函数，以实现所述路径计算。
上述方案中，所述装置还包括：第一搜索模块，配置为在根据所述多维约束集合和所述评估函数，进行路径计算之前，搜索与所述路径计算对 应的算法，其中，所述算法包括以下至少之一：多维约束前K条最优路径TopK算法、多维约束共享风险链路组算法、多维约束最短主备分离路径算法。
上述方案中，所述装置还包括：检测模块，配置为实时检测所述业务是否发生变化；第二搜索模块，配置为在所述业务发生变化时，重新搜索与所述路径计算对应的算法。
上述方案中，所述装置还包括：接收模块，配置为在根据所述多维约束集合和所述评估函数，进行路径计算之前，接收性能分析***优化后的资源信息，其中，所述资源信息用于路径计算。
上述方案中，所述装置还包括下发模块，配置为下发所述路径计算的计算结果。
本发明实施例还提供了一种存储介质，所述存储介质包括存储的程序，其中，所述程序运行时执行本发明实施例提供的多维约束下路径计算方法。
根据本发明的又一个实施例，还提供了一种处理器，所述处理器用于运行程序，其中，所述程序运行时执行本发明实施例提供的多维约束下路径计算方法。
本发明实施例还提供了一种多维约束下路径计算装置，包括：
存储器，配置为保存多维约束下路径计算的程序；
处理器，配置为运行所述程序，其中，所述程序运行时执行本发明实施例提供的多维约束下路径计算方法。
应用本发明实施例，接收业务的算路请求，其中，该业务的业务场景包括多个；该算路请求中包括以下信息至少之一：与该业务对应的约束类型、与该业务对应的评估类型；将该约束类型和预设的多维约束模型进行匹配，得到与该业务对应的多维约束集合；将该评估类型和预设的评估模型进行匹配，得到与该业务对应的评估函数；根据该多维约束集合和该评估函数，进行路径计算。也就是说，通过将业务与路径计算解耦，即，将 业务模块从SDN控制器中单独出去，进而使得任何一种业务场景都可以进行路径计算，解决了相关技术中SDN控制器的业务模块与路径计算模块耦合，无法适用于新的业务场景的扩展和变化以及路径计算模块无法作为独立组件给其他业务场景使用所导致的SDN控制器运行效率较低的问题，同时通过多维约束模型，可以任意构造约束集，而无需与业务紧密关联，可以充分扩展，满足未来复杂业务约束需求，解决了路径计算性能瓶颈的问题，实现了全网资源优化的技术效果。
附图说明
图1是本发明实施例提供的多维约束下路径计算方法流程图；
图2是本发明实施例提供的SDN控制器与性能分析***、编排器、资产库交互关系图；
图3是本发明实施例提供的性能分析***的模块结构示意图；
图4是本发明实施例提供的性能分析***领域优化模块的结构示意图；
图5是本发明实施例提供的拓扑信息收集组件关系图；
图6是本发明实施例提供的拓扑节点信息套件示意图；
图7是本发明实施例提供的拓扑链路信息套件示意图；
图8是本发明实施例提供的拓扑链路属性信息套件示意图；
图9是本发明实施例提供的租户城市信息模型拓扑实例模型；
图10是本发明实施例提供的CIM整网拓扑模型；
图11是本发明实施例提供的单个约束基本信息套件示意图；
图12是本发明实施例提供的约束单元信息套件示意图；
图13是本发明实施例提供的约束集合信息套件示意图；
图14是本发明实施例提供的约束类型信息套件示意图；
图15是本发明实施例提供的多维约束模型；
图16是本发明实施例提供的评估模型；
图17是本发明实施例提供的路径计算工作流机制流程图；
图18是本发明实施例提供的路径计算开始前适配流程图；
图19是本发明实施例提供的多维约束模型解析流程图；
图20是本发明实施例提供的路径计算过程适配流程图；
图21是本发明实施例提供的路径计算结束后适配流程图；
图22是本发明实施例提供的多维约束路径计算遍历算法流程图；
图23是本发明实施例提供的多维约束最优路径算法流程图；
图24是本发明实施例提供的多维约束TopK最优路径算法流程图；
图25是本发明实施例提供的是多维约束SRLG路径算法流程图；
图26是本发明实施例提供的多维约束最优主备分离路径算法流程图；
图27是本发明实施例提供的多维约束下路径计算装置的结构框图；
图28是本发明实施例提供的多维约束下路径计算装置的结构框图。
具体实施方式
下文中将参考附图并结合实施例来详细说明本发明。需要说明的是，在不冲突的情况下，本申请中的实施例及实施例中的特征可以相互组合。
需要说明的是，本发明的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。
在对本发明实施例进行说明之前，首先对本发明实施例中提到的部分名词进行说明。
资产库，用来存储资源信息，如节点、链路、端口、隧道、伪线等资源以及相关属性信息，它主要为SDN控制器提供资产缓存接口。
性能分析***(Performance Analysis System，PAS)，即性能分析***，主要为编排器提供外置策略***，提供性能分析策略，包括网络分析，网络优化。
编排器，主要为SDN控制器提供策略管理(约束策略、评估策略)与工作流机制。
SDN控制器，它的一个重要功能是进行路径计算并对路径进行管理、最终完成路径下发。
在本实施例中提供了一种多维约束下路径计算方法，基于SDN网络，图1是本发明实施例提供的多维约束下路径计算方法流程图，如图1所示，该流程包括如下步骤：
步骤S102，接收业务的算路请求，其中，该业务的业务场景包括多个；该算路请求中包括以下信息至少之一：与该业务对应的约束类型、与该业务对应的评估类型；
需要说明的是，上述约束类型和上述评估类型可以为多个。
步骤S104，将该约束类型和预设的多维约束模型进行匹配，得到与该业务对应的多维约束集合；
步骤S106，将该评估类型和预设的评估模型进行匹配，得到与该业务对应的评估函数；
步骤S108，根据该多维约束集合和该评估函数，进行路径计算。
在一实施例中，上述步骤的执行主体可以为SDN控制器等，但不限于此。
需要说明的是，上述预设的多维约束模型包括以下至少之一：租户身份标识ID、多维约束集合；上述预设的评估模型包括以下至少之一：网络拓扑的下一跳评估、网络拓扑中每条边的边代价评估。
通过上述步骤S102至步骤S108，接收业务的算路请求，其中，该业务的类型包括多个；该算路请求中包括以下信息至少之一：与该业务对应的约束类型、与该业务对应的评估类型；将该约束类型和预设的多维约束模型进行匹配，得到与该业务对应的多维约束集合；将该评估类型和预设的 评估模型进行匹配，得到与该业务对应的评估函数；根据该多维约束集合和该评估函数，进行路径计算。也就是说，通过将业务与路径计算解耦，即，将业务模块从SDN控制器中单独出去，进而使得任何一种业务场景都可以进行路径计算，解决了相关技术中SDN控制器的业务模块与路径计算模块耦合，无法适用于新的业务场景的扩展和变化以及路径计算模块无法作为独立组件给其他业务场景使用所导致的SDN控制器运行效率较低的问题，同时通过多维约束模型，可以任意构造约束集，而无需与业务紧密关联，可以充分扩展，满足未来复杂业务约束需求，解决了路径计算性能瓶颈的问题，实现了全网资源优化的技术效果。
下面结合示例，对本实施例进行举例说明。
本示例基于SDN网络提供一种多维约束下最优路径计算策略的方法，将业务应用与路径计算算法进行解耦，资产信息存储与控制器核心解耦，解决了目前SDN控制器运行效率低下，路由计算性能瓶颈问题；同时将性能分析***数据分析优化结果作为路径计算的前提，由策略驱动，构建多维约束模型、评估模型；利用工作流机制，通过多维约束模型解析，评估函数注入，解耦多维约束集与遍历算法、解耦最优路径评估与遍历算法；同时多维约束模型的构建，解决了复杂业务场景下的无法灵活应对多维约束最优路径计算问题，达到全网资源优化的目的。
在一实施例中，本示例提供的多维约束下最优路径计算策略的方法，包含步骤如下：
步骤S11：性能分析***对性能数据进行精细化分析优化，并将优化结果提供给编排器策略模块和控制器路径计算模块；
步骤S12：编排器提供策略管理，构建多维约束模型，构建评估模型；
步骤S13：控制器提供上下文切换入口，利用工作流机制驱动路径计算流程，与编排器联动；做路径计算搜索前适配：解析多维约束模型、适配搜索算法、选择返回路径类型、进行场景切换检测；
步骤S14：控制器路径计算搜索过程中场景适配：路径搜索计算、注入评估模型、进行场景切换检测；
步骤S15：控制器路径计算搜索结束后场景适配：路径分析、缓存、场景切换检测、路径下发到设备。
应用本发明实施例，具备如下优点：
1)通过资产库，业务应用策略与控制器内部路径计算解耦，并利用异步远程调用，分布式等手段可以大大提升控制器的运行效率。
2)性能分析***通过对业务精细化资源分析优化，为路径计算提供基础性能数据，来解决未来网络越来越复杂，而资源利用不足普遍存在的现状，更加合理的规划网络资源；
3)应用通过业务策略适配，构造约束模型，构建评估模型注入方式进行下一跳最优节点的选择，完全满足了灵活多变的业务场景需要，让控制器对网络的控制更为轻松；
4)控制器通过适配策略应用场景变更进行上下文场景切换，可以及时对拓扑信息进行更新，触发重算机制，对网络震荡带来的用户体验不佳有了及时的预防和控制；
5)控制器通过提供一种通用的多约束路径搜索算法，缩短最优路径计算收敛时间，完全满足任意约束集下的路径计算，让控制器对路径的规划和管理更为便捷。
在一个实施方式中，在根据该多维约束集合和该评估函数，进行路径计算之前，还包括以下步骤：
步骤S21，构建网络拓扑模型，其中，该网络拓扑模型包括以下信息至少之一：拓扑节点集、拓扑链路集、租户集。
在一实施例中，上述方法还包括以下步骤：
步骤S31，检测网络拓扑信息是否发生变化；
步骤S32，在检测到该网络拓扑信息发生变化时，更新该网络拓扑模型。
通过步骤S31至步骤S32，使得网络拓扑模型能够及时得到更新，满足了灵活多变的业务场景需要。
在一个实施方式中，根据该多维约束集合和该评估函数，进行路径计算包括以下步骤：
步骤S41，在对该网络拓扑图模型进行遍历的过程中，注入该多维约束集合；
步骤S42，将该多维约束集合作用于该网络拓扑模型的下一跳选择过程中，并在该下一跳选择过程中和该网络拓扑模型的边代价选择过程中注入该评估函数，以实现该路径计算。
通过上述步骤S41至步骤S42，进一步解决了路径计算性能瓶颈的问题，实现了全网资源优化的技术效果。
在一实施例中，在根据该多维约束集合和该评估函数，进行路径计算之前，还包括以下步骤：
步骤S51，搜索与该路径计算对应的算法，其中，该算法包括以下至少之一：多维约束前K条最优路径TopK算法、多维约束共享风险链路组SRLG算法、多维约束最短主备分离路径算法。
在一实施例中，上述方法还包括：
步骤S61，实时检测该业务是否发生变化；
步骤S62，在该业务发生变化时，重新搜索与该路径计算对应的算法。
通过上述步骤S61至步骤S62，使得路径计算算法可以根据业务的变化及时更新。
在一个实施方式中，在根据该多维约束集合和该评估函数，进行路径计算之前，还包括以下步骤：
步骤S63，接收性能分析***优化后的资源信息，其中，该资源信息用于路径计算。
通过上述步骤S63，性能分析***通过对业务精细化资源分析优化，为 路径计算提供基础性能数据，来解决未来网络越来越复杂，而资源利用不足普遍存在的现状，更加合理的规划网络资源。
在一个实施方式中，上述方法还包括：下发该路径计算的计算结果。
下面结合示例，对本实施例进行举例说明。
SDN控制器与性能分析***、编排器、资产库四者交互关系如图2所示，其中包括：
1)性能分析***通过独立接口与资产库数据之间通讯，性能分析***对资产数据进行采集，计算，分析，并得到优化后的结果返回给编排器和控制器。
2)编排器策略模块通过独立接口与性能分析***之间进行通讯，获取优化数据后，将优化数据应用于策略管理。
3)控制器路径计算模块通过独立接口与性能分析***之间进行通讯，获取优化数据后，将优化数据应用于路径计算。
4)控制器通过独立接口与资产库数据之间的通讯，对资产进行内部缓存，为构造全局拓扑网络模型做准备。
5)控制器通过独立接口与编排器策略模块之间通讯，为控制器路径计算提供策略；控制器根据策略进行路径计算。
其中，性能分析***的内部结构如图3所示，包括以下模块：
1)数据采集模块，通过资产库采集性能分析***所在网络层次的网络资源；
2)数据计算模块，根据数据采集模块采集到的数据，在时间维度和资源维度上汇聚，计算数据；
3)数据分析模块，根据数据计算模块计算得到的数据，在时间维度，资源维度，用户自定义维度上分析用户所需要的数据；
4)数据展示模块，根据数据分析模块的分析数据通过图表形式在多种维度上展示，同时支持导出，可编辑；
5)领域优化模块，根据数据分析模块的分析数据，得到优化策略，策略包括网络分析，网络优化，路径计算，并将策略上报给对应的编排器。
其中，领域优化模块的组成如图4所示，包括网络分析模块、网络优化模块、路径计算模块。
图5至图8为本实施例提供的CIM整网抽象拓扑模型创建过程中各套件信息，以及套件之间的关系。
其中，图5是拓扑信息收集组件关系，它主要说明拓扑信息收集来源与去向，进行拓扑信息缓存，为路径计算提供拓扑，各组件的作用如下：
1)资产库：由资产库的对外接口提供资产数据；
2)控制器资产缓存组件：控制器内部进行资产数据缓存；
3)事件分发组件：发送事件，通知当前网络有拓扑信息变更，比如节点变更，链路变更，租户变更；
4)事件解析组件：对当前拓扑事件进行解析，分析提取有用信息，为拓扑缓存做准备。
5)拓扑缓存组件：由资产数据为拓扑缓存组件提供基本属性信息，动态构建CIM抽象网络拓扑模型，并将拓扑模型中节点、链路、租户以及相关属性信息进行拓扑信息缓存，为搜索服务提供拓扑模型数据准备。
图6为本实施例提供的抽象网络拓扑节点信息套件，其中，节点抽象信息包含ID、当前版本号、节点计算相关属性信息。其中，节点ID为全网抽象模型中全局唯一节点ID标识；版本号信息主要用于判断网络节点是否已有变更，当相关节点发生变更后(主要为节点上线、下线)，之前下发信息则可能已经失效；属性信息主要抽取了与搜索计算相关的属性，例如延时信息，主要用于路径择优计算服务。
图7和图8为本实施例提供的抽象网络链路信息套件，主要为模型中链路抽象信息，包括如下：
拓扑链路抽取信息包含ID、当前版本号、源节点(入节点)，目的节 点(出节点)、属性信息。首先，链路ID为全网抽象模型中全局唯一节点ID标识，对于相同源节点与目的节点的链路ID不同则表示不同链路；因此可以区分两个物理节点不同端口(物理或逻辑)所映射的多条链路，而无需借助端口信息的提取。其次，源节点，目的节点即为图6所抽取的信息。另外，版本号信息主要用于判断网络链路是否已有变更(主要为链路发现，链路断开、属性值变化等)，当相关链路发生变更后，之前下发信息则可能已经失效；
链路属性信息主要抽取了与搜索计算相关的属性，例如带宽相关信息(带宽容量，已用带宽，带宽利用率等)，延时，负载，可靠性等相关信息，主要用于路径择优计算服务。
图9为本实施例提供的租户实际使用的公共信息模型(CIM，Common Information Modeling)抽象网络拓扑实例模型。包括：租户ID、变更时间、子拓扑模型、其他参数。首先，租户ID是不同租户的全局唯一标识；其次，变更时间主要记录收集拓扑发生变更的时间；第三，子拓扑模型主要包含全网拓扑节点子集与链路子集，它们分别是全网节点集与链路集的局部引用；最后，其他参数，主要记录一些用户的其他信息，例如节点个数，链路条数等参考信息。
图10为本实施例提供的整网CIM抽象拓扑模型，该模型主要提供抽象拓扑网络所需要的信息，并将信息进行缓存，并为拓扑搜索服务和上层业务模型提供数据模型基础。它主要包含：节点集合、链路集合、租户集合。首先，节点集合即为图6所示的全网节点信息集。其次，链路集合即为图7所示的全网链路信息集。最后，租户集合即为图9所示的拓扑实例集(租户集合)。
图11至图14是本发明为编排器策略模块业务集提供的约束模型构造。主要服务于任意业务需求下的约束集构造，适应业务场景的复杂性与灵活多变性。同时为拓扑搜索服务提供约束解析入口，用于搜索过程约束处理。 内容如下：
图11为本实施例提供单个约束基本信息套件；该套件主要包含约束值、约束上限、约束下限三个对象，主要用于描述单个约束的基本信息。在实施时，所有业务的约束落到实处都可以由此三项中一个或其组合来描述。例如，带宽利用率由上限、下限来描述；跳数、延时、必经点等可以由约束值来描述。
图12为本实施例提供的约束单元信息套件。该套件主要包含约束名称，约束值集合。在实施时，约束单元，比如必经点集，它会包含多个节点集合；再如必经链路集，包含多条链路集合。因此，需要由约束值集合图11来描述约束单元，而约束单元要提供名称来识别约束含义。
图13至图14为本发明提供的约束集合信息套件。该套件主要包含：约束类型、约束单元(图12所示)集合。在实施时，当业务有算路请求时，必然事先了解是哪个租户(ID)的请求，该租户的算路请求中包含了哪些约束条件(约束单元集合)，并且这些约束条件都分别是什么类型的约束(约束类型，如图14所示)。其中，约束类型主要服务于路径计算搜索服务，用于判断约束作用点标识，例如Graph类型约束，主要作用于路径计算过程，如路径全部节点跳数限制、延时限制、业务申请带宽、带宽利用率限制、必经点列表、必经链路列表、避开点列表、避开链路列表等。Path类型约束，主要用于路径管理与优化；例如：TopK路径、SRLG主备路径、主备分离路径、等价平衡路径等。Node类型约束与Connection类型约束，例如，延时，带宽，负载，可靠性等，都主要作用于节点遍历择优过程。
图15为本实施例提供的多维约束模型多维约束模型(MRM，Multidimensional Restrict Model)。MRM模型主要由租户ID、多维约束集合(S33)构成。在实施时，当业务有算路请求时，必然事先了解是哪个租户(ID)的请求，该租户的算路请求中包含了哪些维度(约束类型)的约束集合，多维度(类型)约束见图14所示。
图16是本实施例提供的评估模型，评估模型(EM，Evaluation Model)，主要服务于任意业务需求下的评估函数构造，适应业务场景的复杂性与灵活多变性，用于搜索过程中评估函数注入。在实施时，下一条评估用于为拓扑中每条边的代价计算函数构建，边的计算可以是边上各考量指标按优先级来获取；也可以是各考量指标按照一定的系数比例进行排列组合；还可以是其他业务需求进行构造。边代价评估用于为拓扑计算下一跳节点选择函数的构建，下一跳节点可以是取加和最小，也可以是乘积最小；或者取最大；或者根据实际业务需要进行构造。
图17为本实施例提供的路径计算搜索服务工作流机制流程，主要包含拓扑事件解析过程、拓扑本地缓存、编排器策略模块构建业务约束模型、编排器策略模块构建业务评估模型、路径计算搜索开始前适配、路径计算搜索过程中适配、路径计算搜索结束后适配的过程，包括：
步骤S1701：解析拓扑事件，构造CIM拓扑模型，并进行本地缓存。
步骤S1702：编排器策略模块进行约束集合构建。
编排器策略模块搜集到业务有算路请求时，根据业务需要，以及约束模型构建方式，进行约束集合构建。
步骤S1703：构建评估函数模型，路径计算搜索服务在搜索过程中注入评估函数。
编排器策略模块搜集到业务有算路请求时，根据业务需要，构建评估函数模型，路径计算搜索服务在搜索过程中注入评估函数。
步骤S1704：进行路径计算搜索适配转换，将业务传入参数进行初始化。
步骤S1705：路径计算搜索前适配，见图18，流程包括步骤S1801至步骤S1804。
步骤S1706：路径计算搜索过程中适配，见图20，流程包括步骤S2001至步骤S2007。
步骤S1707：路径计算搜索结束后适配，见图21，流程包括步骤S2101 步骤S2103。
图18为本实施例提供的路径计算搜索算法开始前适配流程。实施步骤包括：
步骤S1801：多维约束模型解析。
路径计算搜索算法开始之前，先要进行搜索之前与业务请求的适配。首先，需要解析约束模型，解析过程见图19，步骤S1901至步骤S1904。
步骤S1802：通过多维约束集的解析来确定路径返回类型。
路径类型主要包含：最优路、等价平衡路、TopK路径、SRLG共享风险链路组主备路径和最优分离主备路径。
步骤S1803：搜索算法确定。
除了路径返回类型，还需要确定搜索算法的选择。搜索算法主要包含：通用搜索算法、SRLG算法、分离路径算法。
步骤S1804：场景切换检测。
当搜索前适配工作完成后，需要检测此时场景的变化，场景变化，例如，拓扑已经发生变更、业务请求已经发生变化。场景变化直接影响后续步骤的执行动作：重新执行搜索前适配工作、直接结束返回、进行后续流程。
图19为本实施例提供的不同约束类型下相关参数的解析流程。主要包含图表(Graph)类型约束参数、路径(Path)类型约束参数、连接(Connection)类型约束参数、节点(Node)类型约束参数。对所有约束的解析通过配置名称来适配，每次约束解析只解析当前请求中包含的多约束集。
多维约束模型解析主要服务于搜索算法，为算法注入约束条件。实施步骤如下：
步骤S1901：解析约束模型中的图表类型约束。
这里主要包括解析跳数限制、延时限制、请求带宽容量、带宽利用率上下限限制、必经点列表、必经链路列表、必经链路列表、避开点列表、 避开链路列表。
步骤S1902：解析约束模型中的路径类型约束。
这里主要包括解析返回路径的类型。
步骤S1903：解析约束模型中的连接点类型约束。
这里包含权值、带宽利用率、带宽容量、负载、可靠性、延时；此处Connection类型约束数据以及步骤S1904中的Node类型约束数据主要由缓存或者数据库等第三方提供，拓扑服务与业务模块共享第三方资源。这样可以充分减弱数据更新带来的计算结果失效的影响；与此同时，业务模块不需要为拓扑模块准备数据，也提升了控制器运行效率。
步骤S1904：解析约束模型中的节点类型约束。
这里只解析包含节点约束的对象，如延时；其数据由缓存或者数据库提供。
图20为本实施例提供的路径计算搜索过程适配流程。实施步骤如下：
步骤S2001：克隆当前有算路请求的拓扑实例。
克隆的目的是为了记录当前计算的拓扑对象以及链路对象和节点对象的版本号，方便后续比较拓扑有没有发生变更，当前计算结果是否有意义。
步骤S2002：对初始化参数进行检查，做异常保护。
例如算路请求源节点、目标节点是否为空等。
步骤S2003：评估函数初始化。
此步骤主要为搜索遍历算法做准备，提供节点，链路择优的评判标准。
步骤S2004：遍历算法搜索。
进行遍历算法搜索过程(见下图22中步骤S2201-步骤S2206)，计算出源点到所有节点的花费，并且组织最优路径数据返回。
步骤S2005：对返回结果数据进行组织，返回需要的路径类型。
步骤S2006：路径版本号比较。
对返回路径中链路、节点列表进行版本号比较，发现拓扑是否有相关 路径的链路或节点变更，如果是有影响的变更，则需要重算。
步骤S2007：场景切换检测。
对搜索过程场景进行检测，当场景变化时，执行相应结果动作：结束算路、重新算路、继续流程。
图21为本实施例提供的路径计算搜索算法结束之后的适配流程。实施步骤包括：
步骤S2101：路径分析，缓存。
搜索算法结束后需要对路径进行分析，对路径进行缓存，当节点个数较少时，进行路径冗余计算。
步骤S2102：路径切换，转成传输对象，返回业务模块。
路径是由链路列表组成的，而链路在拓扑搜索服务模块中是以专用的抽象对象形式存在的，它可以是物理的，也可以是逻辑的，因此需要进行一定转换。在路径下发之前为了减少模块之间的传输消耗，路径通过链路ID标识列表来进行传输。
步骤S2103：场景切换检测。
对搜索结束后场景进行检测，当场景变化时，执行相应结果动作：重新算路、继续流程。
图22为本实施例提供的遍历算法流程，见图22至图26。它主要包含以下算法：多维约束TopK路径算法、多维约束SRLG算法、多维约束最短分离路径算法、多维约束最优路径算法。实施步骤包括：
步骤S2201：搜索算法入口，约束集传入。
由上面步骤提供多维约束集参数，进入搜索算法入口；
步骤S2202：判断是否主备路径。
根据路径约束参数判断是否需要主备路径计算；不需要主备路径，则进入多维约束TopK路径算法流程S2204。
步骤S2203：判断是否有SRLG链路组。
需要主备路径，则判断是否有SRLG共享风险链路组；有SRLG链路组则进入多维约束SRLG算法流程S2203；否则，进入多维约束最短分离路径算法流程S2206；
步骤S2204：多维约束TopK路径算法流程。
见图24的步骤S2401步骤S2404。
步骤S2205：多维约束SRLG算法流程。
见图25，步骤S2501至步骤S2505。
步骤S2206：多维约束最优主备分离路径算法流程。
见图26，步骤S2601步骤S2608。
图23为本实施例提供的最优路径算法流程，主要为其他算法流程提供最优路径与等价平衡路径计算，实施步骤包括：
步骤S2301：起始节点初始化。
对要进行最优路径计算的拓扑图进行遍历搜索准备，首先对当前遍历节点进行初始化准备(第一次遍历的节点为源节点)；
步骤S2302：遍历起始节点对应的所有出边。
步骤S2303：对出边的出节点进行约束筛选。
对出边的出节点进行约束筛选，出节点主要判断节点是否有延时限制、是否是避开点、必经点；不符合约束的出节点以及对应的边都加入排斥列表；排斥列表中点和边不进行下一个最优节点遍历选举；
步骤S2304：对出边进行约束筛选。
对出边进行约束筛选，出边主要判断是否满足带宽，带宽利用率、延时、负载、可靠性、避开边列表、必经边列表这些约束条件；不符合约束条件的出边以及相应的出节点都加入排斥列表。
步骤S2305：调用出节点代价计算函数，计算符合约束条件的出边对应出节点的代价。
调用用户注入的评估函数，此处需要的是出节点代价计算函数(一般 为线性函数组合，也可以是优先级排列，此处给予用户最大程度的灵活注入方式)，计算非排斥列表中从原点到当前出节点的花费。
步骤S2306：将满足条件的出节点加入候选列表。
将满足条件的出节点以及对应的花费加入候选列表或者更新候选列。
步骤S2307：调用节点最优评估函数，选出下一次遍历起始节点。
调用用户注入的最优节点选择评估函数(可以是加和最小、也可以是乘积最小、还可以是选取最大值或者最小值等，此处给予用户最大程度的灵活注入方式)，选出下一次遍历的起始节点；若所有节点都已经遍历完成，转步骤S2308；否则转入步骤S2301。
步骤S2308：所有节点遍历完成。
步骤S2309：根据候选列中所有节点的代价(花费)，进行最优路径组织，去除环路，并根据最优路径需求的条数来组织等价平衡路径，并将路径(组)返回，算法流程结束。
图24为本实施例提供的通用多维约束最优路径算法搜索流程，即多维约束TopK最优路径算法；TopK默认为1，即只返回一条或者多条等价平衡路。实施步骤如下：
步骤S2401：最优路径算法流程。
首先搜索一条或多条等价平衡最优路，见上面步骤S2301至步骤S2309；
步骤S2402：最优路径是否满足跳数、延时限制，返回路径数据组织，去除不满足的路径。
该最优路径(组)是否满足跳数限制、延时限制、可靠性限制；去除不满足的路径，返回满足条件的路径，并对满足条件的路径进行数据重新组织；转下一步S2403；
步骤S2403：判断TopK是否小于等于返回最优路径条数，若大于转下一步S2404；否则算法流程结束。
步骤S2404：KSP算法次优路遍历搜索，去除一条链路。
利用KSP算法，对源点去掉一条链路，进行次优路径遍历搜索，即改变拓扑图循环步骤S2401步骤S2404。
图25为本实施例提供的多维约束SRLG共享风险链路组主备路径搜索算法流程图；共享风险链路组是指，在同一组中的所有链路，当有其中一条有故障时，其余的都会断开。因此，在主备路径计算时，属于同一个风险组的所有链路都出现在一条链路上。即主备路径不会包含同一风险组的链路。实施步骤：
步骤S2501：分析风险链路组，按组进行分配；同时初始化互斥布尔型对象，用于后面遗传算法过程中对主备路径风险组进行区分。
步骤S2502：遗传算法过程(初始化群体的产生、适应度计算、选择运算、交叉运算、变异运算)。
步骤S2503：中间主备路径群体产生调用最优路径算法流程；
遗传算法中间主、备路径群体产生过程调用最优路径算法流程，见图23，步骤S2301步骤S2309；
步骤S2504：最优主、备路径组是否满足跳数、延时、可靠性等的限制，去除不满足的路径组，返回满足的路径组；
步骤S2505：当不能选择出最优SRLG主、路径组时，选择次优路径组返回，有TopK约束时，返回TopK主备路径组，算法流程结束。
图26为本发明提供的多维约束最优主备分离路径算法流程；实施步骤如下：
步骤S2601：计算等价平衡最优路径并进行最终路径约束检测，返回满足条件的最优路径。
利用多维约束最优路径算法(步骤S2301-步骤S2309)，计算一条或多条等价平衡最优路径，并进行最终路径约束检测(跳数、延时、可靠性等)，返回满足条件的最优路径(组)，进行循环；
步骤S2602：对多维约束最优路径进行反向构造新的拓扑图；
这里可采用Suurballe算法；
步骤S2603：第二次计算最优路径；
对多维约束最优路径算法流程第二次计算最优路径(组)，循环；
步骤S2604：去除两次计算重复路径，重新构造拓扑图；
步骤S2605：第三次计算最优路径，并进行最终路径的约束检测；
多维约束最优路径算法流程第三次计算最优路径(组)，并进行最终路径的约束检测(跳数、延时、可靠性等)。此路径(组)为主路径；循环；
步骤S2606：去除主路径节点和边，重新构造拓扑图；
步骤S2607：第四次计算最优路径，并进行最终路径的约束检测；
多维约束最优路径算法流程第四次计算最优路径(组)；并进行最终路径约束检测(跳数、延时、可靠性等)，此路径为备路径，循环；
步骤S2608：主备路径重新组织，检测TopK需求，当有TopK需求时，返回TopK主备路径，算法流程结束。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到根据上述实施例的方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，或者网络设备等)执行本发明各个实施例所述的方法。
本发明实施例还提供了一种多维约束下路径计算装置，该装置配置为实现上述实施例及实施方式，已经进行过说明的不再赘述。如以下所使用 的，术语“模块”可以实现预定功能的软件和/或硬件的组合。尽管以下实施例所描述的装置较佳地以软件来实现，但是硬件，或者软件和硬件的组合的实现也是可能并被构想的。
图27是根据本发明实施例的多维约束下路径计算装置的结构框图，如图27所示，该装置包括：
1)接收模块2702，配置为接收业务的算路请求，其中，该业务的业务场景包括多个；该算路请求中包括以下信息至少之一：与该业务对应的约束类型、与该业务对应的评估类型；
需要说明的是，上述约束类型和上述评估类型可以为多个。
2)第一获取模块2704，配置为将该约束类型和预设的多维约束模型进行匹配，得到与该业务对应的多维约束集合；
3)第二获取模块2706，配置为将该评估类型和预设的评估模型进行匹配，得到与该业务对应的评估函数；
4)计算模块2708，配置为根据该多维约束集合和该评估函数，进行路径计算。
需要说明的是，上述预设的多维约束模型包括以下至少之一：租户身份标识ID、多维约束集合；上述预设的评估模型包括以下至少之一：网络拓扑的下一跳评估、网络拓扑中每条边的边代价评估。
通过图27所示装置，接收业务的算路请求，其中，该业务的类型包括多个；该算路请求中包括以下信息至少之一：与该业务对应的约束类型、与该业务对应的评估类型；将该约束类型和预设的多维约束模型进行匹配，得到与该业务对应的多维约束集合；将该评估类型和预设的评估模型进行匹配，得到与该业务对应的评估函数；根据该多维约束集合和该评估函数，进行路径计算。也就是说，通过将业务与路径计算解耦，即，将业务模块从SDN控制器中单独出去，进而使得任何一种业务场景都可以进行路径计算，解决了相关技术中SDN控制器的业务模块与路径计算模块耦合，无法 适用于新的业务场景的扩展和变化以及路径计算模块无法作为独立组件给其他业务场景使用所导致的SDN控制器运行效率较低的问题，同时通过多维约束模型，可以任意构造约束集，而无需与业务紧密关联，可以充分扩展，满足未来复杂业务约束需求，解决了路径计算性能瓶颈的问题，实现了全网资源优化的技术效果。
下面结合示例，对本实施例进行举例说明。
本示例为解决多维约束下路径计算还包含一种多维约束路径搜索服务装置，包含以下模块：
数据抽取模块：配置为动态获取拓扑信息，并进行拓扑信息缓存，创建CIM整网拓扑模型；
策略适配模块：配置为对业务场景进行解析，开始路径计算流程；
约束解析模块：配置为对多维约束集进行解析，为路径计算提供约束参数集；
评估解析模块：配置为对评估函数进行解析，提供下一跳择优和边代价计算的方法；
算路模块：配置为以遍历思想为基本原则，遍历过程中注入多维约束集，让多维约束集作用于下一跳选择过程中；并且在下一跳选择和边的代价计算过程中注入相应的评估函数，实现评估函数按业务策略来选择，完成多维约束最优路径计算；并将此过程扩展到多维约束TopK算法、多维约束SRLG主备路径算法、多维约束最短主备分离路径算法，完成不同业务需求下多维约束最优路径搜索；
路径管理模块：配置为对路径进行管理，将得到的路径返回给策略模块，进行下发。
在一个实施方式中，上述装置还包括：构建模块，其中，该构建模块配置为在根据该多维约束集合和该评估函数，进行路径计算之前，构建网络拓扑模型，其中，该网络拓扑模型包括以下信息至少之一：拓扑节点集、 拓扑链路集、租户集。
在一实施例中，上述装置还包括：检测模块，配置为检测网络拓扑信息是否发生变化；更新模块，配置为在检测到该网络拓扑信息发生变化时，更新该网络拓扑模型。通过该装置，使得网络拓扑模型能够及时得到更新，满足了灵活多变的业务场景需要。
图28是根据本发明实施例的多维约束下路径计算装置的结构框图，如图28所示，计算模块2708包括：
1)注入单元2802，配置为在对该网络拓扑图模型进行遍历的过程中，注入该多维约束集合；
2)计算单元2804，配置为将该多维约束集合作用于该网络拓扑模型的下一跳选择过程中，并在该下一跳选择过程中和该网络拓扑模型的边代价选择过程中注入该评估函数，以实现该路径计算。
通过图28的装置，进一步解决了路径计算性能瓶颈的问题，实现了全网资源优化的技术效果。
在一实施例中，上述装置还包括：第一搜索模块，配置为在根据该多维约束集合和该评估函数，进行路径计算之前，搜索与该路径计算对应的算法，其中，该算法包括以下至少之一：多维约束前K条最优路径TopK算法、多维约束共享风险链路组SRLG算法、多维约束最短主备分离路径算法。
在一实施例中，上述装置还包括：检测模块，配置为实时检测该业务是否发生变化；第二搜索模块，配置为在该业务发生变化时，重新搜索与该路径计算对应的算法。通过该装置，使得路径计算算法可以根据业务的变化及时更新。
在一个实施方式中，上述装置还包括：接收模块，配置为在根据该多维约束集合和该评估函数，进行路径计算之前，接收性能分析***优化后的资源信息，其中，该资源信息用于路径计算。通过该装置，性能分析系 统通过对业务精细化资源分析优化，为路径计算提供基础性能数据，来解决未来网络越来越复杂，而资源利用不足普遍存在的现状，更加合理的规划网络资源。
在一实施例中，上述装置还包括：下发模块，配置为下发该路径计算的计算结果。
需要说明的是，上述各个模块是可以通过软件或硬件来实现的，对于后者，可以通过以下方式实现，但不限于此：上述模块均位于同一处理器中；或者，上述各个模块以任意组合的形式分别位于不同的处理器中。
本发明实施例还提供了一种存储介质，该存储介质包括存储的程序，其中，上述程序运行时执行本发明实施例提供的所述多维约束下路径计算方法。
在一实施例中，上述存储介质可以被设置为存储用于执行以下步骤的程序代码：
S1，接收业务的算路请求，其中，所述业务的业务场景包括多个；所述算路请求中包括以下信息至少之一：与所述业务对应的约束类型、与所述业务对应的评估类型；
S2，将所述约束类型和预设的多维约束模型进行匹配，得到与所述业务对应的多维约束集合；
S3，将所述评估类型和预设的评估模型进行匹配，得到与所述业务对应的评估函数；
S4，根据所述多维约束集合和所述评估函数，进行路径计算。
在一实施例中，上述存储介质可以包括但不限于：U盘、只读存储器(Read-Only Memory，ROM)、随机存取存储器(Random Access Memory，RAM)、移动硬盘、磁碟或者光盘等各种可以存储程序代码的介质。
本发明实施例还提供了一种处理器，该处理器配置为运行程序，其中，该程序运行时执行本发明实施例提供的所述多维约束下路径计算方法。
在一实施例中，上述程序用于执行以下步骤：
S1，接收业务的算路请求，其中，所述业务的业务场景包括多个；所述算路请求中包括以下信息至少之一：与所述业务对应的约束类型、与所述业务对应的评估类型；
S2，将所述约束类型和预设的多维约束模型进行匹配，得到与所述业务对应的多维约束集合；
S3，将所述评估类型和预设的评估模型进行匹配，得到与所述业务对应的评估函数；
S4，根据所述多维约束集合和所述评估函数，进行路径计算。
在一实施例中，本实施例中的示例可以参考上述实施例及可选实施方式中所描述的示例，本实施例在此不再赘述。
显然，本领域的技术人员应该明白，上述的本发明的各模块或各步骤可以用通用的计算装置来实现，它们可以集中在单个的计算装置上，或者分布在多个计算装置所组成的网络上，在实际应用中，它们可以用计算装置可执行的程序代码来实现，从而，可以将它们存储在存储装置中由计算装置来执行，并且在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤，或者将它们分别制作成各个集成电路模块，或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。这样，本发明不限制于任何特定的硬件和软件结合。
以上所述仅为本发明的优选实施例而已，并不用于限制本发明，对于本领域的技术人员来说，本发明可以有各种更改和变化。凡在本发明的原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。

Claims (21)

  1. 一种多维约束下路径计算方法，包括：
    接收业务的算路请求，其中，所述业务的业务场景包括多个；所述算路请求中包括以下信息至少之一：与所述业务对应的约束类型、与所述业务对应的评估类型；
    将所述约束类型和预设的多维约束模型进行匹配，得到与所述业务对应的多维约束集合；
    将所述评估类型和预设的评估模型进行匹配，得到与所述业务对应的评估函数；
    根据所述多维约束集合和所述评估函数，进行路径计算。
  2. 根据权利要求1所述的方法，其中，
    所述预设的多维约束模型包括以下至少之一：租户身份标识ID、多维约束集合；
    所述预设的评估模型包括以下至少之一：网络拓扑的下一跳评估、网络拓扑中每条边的边代价评估。
  3. 根据权利要求2所述的方法，其中，在根据所述多维约束集合和所述评估函数，进行路径计算之前，还包括：
    构建网络拓扑模型，其中，所述网络拓扑模型包括以下信息至少之一：拓扑节点集、拓扑链路集、租户集。
  4. 根据权利要求3所述的方法，其中，还包括：
    检测网络拓扑信息是否发生变化；
    在检测到所述网络拓扑信息发生变化时，更新所述网络拓扑模型。
  5. 根据权利要求3或4所述的方法，其中，根据所述多维约束集合和所述评估函数，进行路径计算包括：
    在对所述网络拓扑图模型进行遍历的过程中，注入所述多维约束集合；
    将所述多维约束集合作用于所述网络拓扑模型的下一跳选择过程中，并在所述下一跳选择过程中和所述网络拓扑模型的边代价选择过程中注入所述评估函数，以实现所述路径计算。
  6. 根据权利要求1所述的方法，其中，在根据所述多维约束集合和所述评估函数，进行路径计算之前，还包括：
    搜索与所述路径计算对应的算法，其中，所述算法包括以下至少之一：多维约束前K条最优路径TopK算法、多维约束共享风险链路组SRLG算法、多维约束最短主备分离路径算法。
  7. 根据权利要求6所述的方法，其中，还包括：
    实时检测所述业务是否发生变化；
    在所述业务发生变化时，重新搜索与所述路径计算对应的算法。
  8. 根据权利要求1所述的方法，其中，在根据所述多维约束集合和所述评估函数，进行路径计算之前，还包括：
    接收性能分析***优化后的资源信息，其中，所述资源信息用于所述路径计算。
  9. 根据权利要求1所述的方法，其中，所述方法还包括：
    下发所述路径计算的计算结果。
  10. 一种多维约束下路径计算装置，包括：
    接收模块，配置为接收业务的算路请求，其中，所述业务的业务场景包括多个；所述算路请求中包括以下信息至少之一：与所述业务对应的约束类型、与所述业务对应的评估类型；
    第一获取模块，配置为将所述约束类型和预设的多维约束模型进行匹配，得到与所述业务对应的多维约束集合；
    第二获取模块，配置为将所述评估类型和预设的评估模型进行匹配，得到与所述业务对应的评估函数；
    计算模块，配置为根据所述多维约束集合和所述评估函数，进行路径 计算。
  11. 根据权利要求10所述的装置，其中，
    所述预设的多维约束模型包括以下至少之一：租户身份标识ID、多维约束集合；
    所述预设的评估模型包括以下至少之一：网络拓扑的下一跳评估、网络拓扑中每条边的边代价评估。
  12. 根据权利要求11所述的装置，其中，所述装置还包括：
    构建模块，配置为在根据所述多维约束集合和所述评估函数，进行路径计算之前，构建网络拓扑模型，其中，所述网络拓扑模型包括以下信息至少之一：拓扑节点集、拓扑链路集、租户集。
  13. 根据权利要求12所述的装置，其中，所述装置还包括：
    检测模块，配置为检测网络拓扑信息是否发生变化；
    更新模块，配置为在检测到所述网络拓扑信息发生变化时，更新所述网络拓扑模型。
  14. 根据权利要求12或13所述的装置，其中，所述计算模块包括：
    注入单元，配置为在对所述网络拓扑图模型进行遍历的过程中，注入所述多维约束集合；
    计算单元，配置为将所述多维约束集合作用于所述网络拓扑模型的下一跳选择过程中，并在所述下一跳选择过程中和所述网络拓扑模型的边代价选择过程中注入所述评估函数，以实现所述路径计算。
  15. 根据权利要求10所述的装置，其中，还包括：
    第一搜索模块，配置为在根据所述多维约束集合和所述评估函数，进行路径计算之前，搜索与所述路径计算对应的算法，其中，所述算法包括以下至少之一：多维约束前K条最优路径TopK算法、多维约束共享风险链路组SRLG算法、多维约束最短主备分离路径算法。
  16. 根据权利要求15所述的装置，其中，还包括：
    检测模块，配置为实时检测所述业务是否发生变化；
    第二搜索模块，配置为在所述业务发生变化时，重新搜索与所述路径计算对应的算法。
  17. 根据权利要求10所述的装置，其中，所述装置还包括：
    接收模块，配置为在根据所述多维约束集合和所述评估函数，进行路径计算之前，接收性能分析***优化后的资源信息，其中，所述资源信息用于所述路径计算。
  18. 根据权利要求10所述的装置，其中，所述装置还包括：
    下发模块，配置为下发所述路径计算的计算结果。
  19. 一种存储介质，所述存储介质包括存储的程序，其中，所述程序运行时执行权利要求1至9中任一项所述的多维约束下路径计算方法。
  20. 一种处理器，所述处理器配置为运行程序，其中，所述程序运行时执行权利要求1至9中任一项所述的多维约束下路径计算方法。
  21. 一种多维约束下路径计算装置，包括：
    存储器，配置为保存多维约束下路径计算的程序；
    处理器，配置为运行所述程序，其中，所述程序运行时执行权利要求1至9中任一项所述的多维约束下路径计算方法。
PCT/CN2018/120350 2018-01-03 2018-12-11 多维约束下路径计算方法、装置、处理器及存储介质 WO2019134483A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP18898081.7A EP3720064A4 (en) 2018-01-03 2018-12-11 METHOD AND DEVICE FOR TRAVELING UNDER MULTI-DIMENSIONAL RESTRICTIONS, PROCESSOR AND STORAGE MEDIUM

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810005339.9A CN109995650B (zh) 2018-01-03 2018-01-03 基于sdn网络的多维约束下路径计算方法及装置
CN201810005339.9 2018-01-03

Publications (1)

Publication Number Publication Date
WO2019134483A1 true WO2019134483A1 (zh) 2019-07-11

Family

ID=67128470

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/120350 WO2019134483A1 (zh) 2018-01-03 2018-12-11 多维约束下路径计算方法、装置、处理器及存储介质

Country Status (3)

Country Link
EP (1) EP3720064A4 (zh)
CN (1) CN109995650B (zh)
WO (1) WO2019134483A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112491619A (zh) * 2020-11-25 2021-03-12 东北大学 一种基于sdn的服务定制网络资源自适应分配技术
CN113904974A (zh) * 2021-10-09 2022-01-07 咪咕文化科技有限公司 智能路由方法、装置及设备
CN116939058A (zh) * 2023-09-15 2023-10-24 工业云制造(四川)创新中心有限公司 基于云制造模式的多协议设备接入方法及***

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112448886B (zh) * 2019-08-30 2023-08-01 中兴通讯股份有限公司 最短路径计算方法、路由获取装置及服务器
CN114390626A (zh) * 2020-10-20 2022-04-22 ***通信集团设计院有限公司 网络路由确定方法、装置及可读存储介质
CN112714062B (zh) * 2020-12-07 2022-04-29 山东省计算中心（国家超级计算济南中心） 一种面向超算用户体验质量的多路径路由方法和装置
CN112965738B (zh) * 2021-02-02 2022-07-08 烽火通信科技股份有限公司 一种信息模型版本差异比对方法及装置
CN113179154B (zh) * 2021-03-25 2022-07-01 北京邮电大学 量子密钥分发物联网中资源联合分配方法及相关设备
CN113207138B (zh) * 2021-05-11 2022-08-02 网络通信与安全紫金山实验室 基于多质量属性的sr方法、***、设备和介质
CN114419897A (zh) * 2022-03-30 2022-04-29 成都柔水科技有限公司 一种基于v2x技术的城市交通cim***及其展示方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101610433A (zh) * 2009-07-10 2009-12-23 北京邮电大学 一种支持策略解析的多约束条件路由选择方法
CN101753425A (zh) * 2008-12-01 2010-06-23 北京航空航天大学 在多约束下求取网络中多条最短简单路径的启发式方法
CN103873364A (zh) * 2012-12-11 2014-06-18 清华大学 域间多路径路由的实现方法
CN104158753A (zh) * 2014-06-12 2014-11-19 南京工程学院 基于软件定义网络的动态流调度方法及***
CN104253749A (zh) * 2014-09-18 2014-12-31 华南理工大学 一种基于软件定义网络架构的用户端分布式路径计算方法
CN104702502A (zh) * 2013-12-09 2015-06-10 中兴通讯股份有限公司 网络路径计算方法及装置

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9634867B2 (en) * 2014-05-02 2017-04-25 Futurewei Technologies, Inc. Computing service chain-aware paths
US10715447B2 (en) * 2015-10-15 2020-07-14 Futurewei Technologies, Inc. Framework for temporal label switched path tunnel services

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101753425A (zh) * 2008-12-01 2010-06-23 北京航空航天大学 在多约束下求取网络中多条最短简单路径的启发式方法
CN101610433A (zh) * 2009-07-10 2009-12-23 北京邮电大学 一种支持策略解析的多约束条件路由选择方法
CN103873364A (zh) * 2012-12-11 2014-06-18 清华大学 域间多路径路由的实现方法
CN104702502A (zh) * 2013-12-09 2015-06-10 中兴通讯股份有限公司 网络路径计算方法及装置
CN104158753A (zh) * 2014-06-12 2014-11-19 南京工程学院 基于软件定义网络的动态流调度方法及***
CN104253749A (zh) * 2014-09-18 2014-12-31 华南理工大学 一种基于软件定义网络架构的用户端分布式路径计算方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3720064A4 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112491619A (zh) * 2020-11-25 2021-03-12 东北大学 一种基于sdn的服务定制网络资源自适应分配技术
CN113904974A (zh) * 2021-10-09 2022-01-07 咪咕文化科技有限公司 智能路由方法、装置及设备
CN113904974B (zh) * 2021-10-09 2023-08-15 咪咕文化科技有限公司 智能路由方法、装置及设备
CN116939058A (zh) * 2023-09-15 2023-10-24 工业云制造(四川)创新中心有限公司 基于云制造模式的多协议设备接入方法及***
CN116939058B (zh) * 2023-09-15 2023-11-28 工业云制造(四川)创新中心有限公司 基于云制造模式的多协议设备接入方法及***

Also Published As

Publication number Publication date
EP3720064A1 (en) 2020-10-07
CN109995650A (zh) 2019-07-09
EP3720064A4 (en) 2021-01-20
CN109995650B (zh) 2022-02-22

Similar Documents

Publication Publication Date Title
WO2019134483A1 (zh) 多维约束下路径计算方法、装置、处理器及存储介质
CN104903894B (zh) 用于分布式数据库查询引擎的***和方法
CN109582667A (zh) 一种基于电力调控大数据的多数据库混合存储方法及***
CN108228724A (zh) 基于图数据库的电网gis拓扑分析方法及存储介质
Mongiovi et al. Efficient multicasting for delay tolerant networks using graph indexing
US20090097418A1 (en) System and method for network service path analysis
CN106156328B (zh) 一种银行运行日志数据监控方法及***
CN110825769A (zh) 一种数据指标异常的查询方法和***
US8812645B2 (en) Query optimization in a parallel computer system with multiple networks
CN106126629B (zh) 一种基于直播行业的主数据管理方法及***
CN108737213A (zh) 一种基于fpga的高并行大吞吐量渗透测试***及方法
CN111737483A (zh) 智能电网大数据知识图谱的构建方法
CN109753401A (zh) 监控方法、采集端、管控端、监控***及装置
CN107491463A (zh) 数据查询的优化方法和***
CN101741717A (zh) 一种基于条件风险分离度的多域网路由方法
US9195710B2 (en) Query optimization in a parallel computer system to reduce network traffic
CN115297048B (zh) 一种基于光纤网络的路由路径生成方法及装置
WO2012174882A1 (zh) 一种电信网管性能数据输出的方法及装置
CN106209978B (zh) 一种联盟关系服务组合选择***及选择方法
Imawan et al. Road traffic analytic query processing based on a timeline modeling
Chiky et al. Summarizing distributed data streams for storage in data warehouses
WO2015154641A1 (zh) 一种业务并发性预测方法与预测***
Lastine et al. A fault-tolerant multipoint cycle routing algorithm (mcra)
CN110297827A (zh) 一种集合Lucene和Hbase的电力数据检索***
US8037184B2 (en) Query governor with network monitoring in a parallel computer system

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18898081

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2018898081

Country of ref document: EP

Effective date: 20200630