覆盖网络中多服务静态部署算法

发布时间：2018-03-20 21:58

本文选题：覆盖网络　切入点：服务部署　出处：《西安电子科技大学学报》2014年04期 　论文类型：期刊论文

【摘要】：针对因特网的覆盖网络中多服务在不同服务节点的部署问题,提出了一种保证平均请求转发延迟满足服务质量要求,以最小化服务部署规模为目标的服务部署模型.该模型在传统的单服务部署问题的基础上,增加了多服务的分配任务;为了合理均衡利用服务节点的服务器资源,引入并发上限限制单节点的并发数目.证明了该模型属于非确定性多项式时间完全问题,提出了两种贪婪启发式算法,两种算法可以在多项式时间内求解.实验结果表明,所提出模型和启发式方法能够大大降低服务部署规模,分别将服务部署规模降低为原始规模的41%和47.8%.
[Abstract]:According to the deployment of network coverage in the Internet service in different service nodes, presents a guaranteed average request forwarding delay to meet quality of service requirements, service deployment model to minimize the scale of service deployment target. In this model, the traditional single service deployment problem based on the increase of task allocation of multi service; in order to use reasonable and balanced service node number of concurrent server resources, the introduction of concurrency limit of single node. It is proved that the model is non deterministic polynomial time problem, propose two greedy heuristic algorithm, the two algorithms can be solved in polynomial time. The experimental results show that the proposed model and heuristic methods can greatly reduce the scale of service deployment, which will reduce the scale of service deployment to the original size of 41% and 47.8%.

【作者单位】：中国科学院大学理学院;中国科学院声学研究所国家网络新媒体工程技术研究中心;
【基金】：国家科技支撑计划资助项目(2011BAH11B04) 国家高技术研究发展计划(863)资助项目(2011AA01A102) 中国科学院先导专项资助项目(XDA6030500)
【分类号】：TP393.02

【参考文献】