| 摘要: |
| 针对优化服务功能链(Service Function Chain,SFC)可靠性时资源利用率低的问题，提出了特征关联的服务功能链可靠性优化算法。首先，从底层节点的资源可用率、周边资源密集度、全网中心性评价其作为虚拟网络功能（Virtual Network Function，VNF）实例备选节点的优势，从VNF实例的节点度数与带宽流量评价其作为备选对象的优势。其次，逐个对优越性最高的VNF实例选择备份底层节点，直到所有SFC的可靠性满足客户要求。在为VNF实例选择备份底层节点时，提出了基于资源特征关联感知的备份节点选择算法，将当前VNF实例所映射的底层节点与其他待备份VNF实例的底层节点的邻接节点进行关联，为当前VNF实例选择最优备份底层节点的同时，减少为其他待备份VNF实例选择备份底层节点时链路的开销。仿真结果表明，相比于已有算法，该SFC可靠性优化算法下备份实例数降低了约29%，备份带宽资源消耗降低了约42%，SFC接受率提升了约12%。 |
| 关键词: 网络功能虚拟化(VNF) 服务功能链(SFC) 服务可靠性 资源备份 特征关联 |
| DOI：10.20079/j.issn.1001-893x.220823003 |
|
| 基金项目:国家重点研发计划(2018YFB1402704)；信息网络安全公安部重点实验室开放课题(C20612) |
|
| Reliability Optimization Algorithm of Service Function Chain Based on Feature Association |
| ZHANG Shunli,SHAO Sujie |
| (1.Department of Information Technology and Engineering,Jinzhong University,Jinzhong 030619,China;2.School of Computer Science,Beijing University of Posts and Telecommunications,Beijing 100876,China) |
| Abstract: |
| For the problem of low resource utilization when optimizing the reliability of service function chain(SFC),a reliability optimization algorithm for SFC based on feature association is proposed.First,the advantages of the underlying node as a candidate node for virtual network function(VNF) instances are evaluated from the resource availability rate,the density of surrounding resources,and the centrality of the entire network,and the advantages of VNF instance as an alternative object are evaluated from the node degree and bandwidth traffic of VNF instance.Second,backup underlying nodes for the VNF instance with the highest superiority are selected one by one until the reliability of all SFCs meets customer requirements.When selecting backup underlying nodes for VNF instances,a backup node selection algorithm based on resource feature association awareness is proposed.The algorithm associates the underlying node mapped by the current VNF instance with the adjacent nodes of the underlying node of other VNF instances to be backed up.While selecting the optimal backup underlying node for the current VNF instance,it reduces the link overhead when selecting the backup underlying node for other VNF instances to be backed up.The simulation results show that compared with that of the existing algorithms,the number of backup instances is reduced by about 29%,the backup bandwidth resource consumption is reduced by about 42%,and the SFC acceptance rate is improved by about 12%. |
| Key words: network function virtualization(NFV) service function chain(SFC) service reliability resource backup feature association |
