Peer-to-Peer流量缓存部署建模  

Modeling of peer-to-peer traffic caching deployment

作　　者：叶明江[1] 吴建平[1] 徐恪[1] 

机构地区：[1]清华大学计算机科学与技术系,北京100084

出　　处：《清华大学学报（自然科学版）》2009年第1期110-113,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家"九七三"重点基础研究项目(2003CB314801);国家自然科学基金资助项目(60473082);国家自然基金委和香港RGC联合研究资助项目(20731160014)

摘　　要：为了在网络中更好地部署对等网络(P2P)流量缓存,从而优化网络的性能,提出了一个评价方法量化在不同链路上部署P2P缓存的收益。基于该方法建立了最优化模型:在给定部署开销的约束下最大化收益。分析了问题复杂度,提出了优先选择收益开销比的链路进行部署的启发式算法求解。对实际校园网络的仿真结果表明:在10%的链路上部署P2P流量缓存,能够降低30%的平均链路利用率。该文提出的P2P流量缓存部署模型和算法可以有效地指导P2P缓存的部署,优化网络性能。The paper provides a method to optimize network performance by deploying cache devices across the network. The deployment problem was modeled as an optimization problem that aims to maximize the benefit of the cache deployment while limiting the total cost. A heuristic algorithm developed to solve the problem selects the link with the maximum benefit per cost in each round until no candidate link remains. Experiments on a real campus network show that deploying cached devices on just less than 10% of the links can reduce the average link utilization by 30%. This algorithm can effectively manage the deployment of peer-to-peer caches to optimize network performance.

关 键 词：计算机网络管理 对等网络(P2P) 流量缓存 流量管理 

分 类 号：TP393.07[自动化与计算机技术—计算机应用技术]