论文部分内容阅读
网络功能虚拟化技术的提出,为减少服务提供商的投资成本与管理开销,实现网络功能的灵活组合、新服务的快速部署提供了解决方案,使多样化、可扩展的服务供应成为可能.网络功能虚拟化中的研究难点是服务链映射问题.现有以最小化端到端时延为目标的映射算法受制于资源瓶颈的限制,导致有效服务路径不能被映射,影响服务链的映射成功率.针对这一问题,提出一种求解服务链映射问题的粒子群优化算法(SIFC-SCMPSO),并以路径跳数为约束设计了初始化可行解与位置更新策略.实验结果表明,该算法有效地减少了资源用量,提高了有效路径被发现的概率,在服务链映射成功率、长期平均收益、平均开销等评价指标上均取得了较好的结果.
The proposed virtualization of network functions has provided solutions to reduce the investment costs and overhead of service providers, enable flexible combinations of network functions and provide rapid deployment of new services, making it possible to diversify and expand service offerings. The research difficulty in functional virtualization is the service chain mapping problem.At present, the mapping algorithm that aims to minimize the end-to-end delay is limited by the resource bottleneck, which leads to that the effective service path can not be mapped and the success rate of service chain mapping To solve this problem, a particle swarm optimization algorithm (SIFC-SCMPSO) is proposed to solve the service chain mapping problem, and the initial feasible solution and location updating strategy are designed based on the path hop count. The experimental results show that this algorithm is effective Reduce the amount of resources and improve the probability of finding effective paths. Good results have been obtained on the evaluation indexes such as the service chain mapping success rate, long-term average return and average cost.