摘要
随着设备间通信数量的增多,网络的稳定性和资源的利用率成为评价网络性能的两个重要指标.现有的可生存性算法使用备用资源来提高网络的稳定性,但资源浪费问题严重.本文提出了一种支持路径切割和资源共享的可生存性算法(GSBKE-Pc-Yshare),此算法支持虚拟请求的不同链路共享同一条物理链路来减少备用资源的使用量.由于现实中物理资源有限,任意两个物理节点之间很难找到一条资源数量能完全满足虚拟网络请求,同时与主用流不交叉的备用路径,因此网络接受率较低.本文中提出的GSBKE-Pc-Yshare算法支持路径切割,允许两条物理路径同时为虚拟链路提供备用资源,即通过降低对备用路径资源数量的限制来提高网络的接受率.此外,本文用群搜索优化算法(Group Search Optimizer,GSO)来优化目标函数以获得较好的映射方案.仿真实验表明,与现有的可生存性算法相比,GSBKE-Pc-Yshare算法的接受率提高30%,网络平均收益提高50%,网络成本降低超过30%.
With the increasing number of communications between devices,the stability of the network and the utilization of resources become two important indicators to evaluate network performance. Most of existing survivability algorithms uses backup resources to improve the stability of the network,however the waste of resources is a serious problem. We propose a survivability algorithm ( GS- BKE-Pc-Yshare) in this paper,its feature is to support path splitting and resources sharing. This algorithm supports the same virtual link requests to share the same physical link which reduces backup resource usage. Due to the limited physical resources, it is difficult to find a backup flow path which the number of resources can satisfy virtual requests and do not cross with the main flow path between any two physical nodes. Therefore,the acceptance rate of network is low. The proposed GSBKE-Pc-Yshare algorithm in this paper sup- ports the path splitting,allowing two physical paths to provide backup resources for the virtual links. That is by releasing the constraint of required resource of backup path to improve the acceptance rate of the network. Meanwhile, in this paper, we use ( Group Search Optimizer, GSO ) GSO to optimize the objective function to get better mapping schemes. Simulation results show that comparing to the existing survivability algorithms, GSBKE-Pc-Yshare can obtain 30% improvement for acceptance rate, the average revenue improves 50% and network cost reduces over 30%.
出处
《小型微型计算机系统》
CSCD
北大核心
2016年第8期1755-1759,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61373149
61272094
61472232
61402270)资助