一种新的基于连续哈希函数的合作式Web缓存系统  

A new cooperative Web caching system based on consistent hashing

作　　者：周刚[1] 周建国[1] 晏蒲柳[1] 

机构地区：[1]武汉大学电子信息学院,湖北武汉430079

出　　处：《计算机应用》2006年第3期733-735,738,共4页journal of Computer Applications

摘　　要：提出了一种新的基于连续哈希函数的合作式缓存系统。针对传统合作式缓存系统中多级转发造成的高时延和多重哈希计算问题,设计了一种高效的Web对象定位和路由模式,保证任意Web请求只需计算一次哈希且至多经过一次转发就可到达目标节点。采用失效-触发的策略来解决路由表一致性维护的问题,减少了网络开销,提高了系统的可扩展性和可靠性。仿真实验表明,该系统性能优于基于互联网缓存协议和缓存阵列路由协议的系统。A new cooperative Web caching system （CHRCS） based on consistent hashing function was proposed. Aiming at reducing the multilevel latency in traditional cooperative Web caching systems, a efficient location and routing schema of Web objects was designed, through which a Web request can reach the target cache by only one hash computation and through less than one transfer. The strategy of invalidation - triggering was assumed to solve the problem of routing consistency, which reduced network traffic and made system more robust and scalable. Simulation results show that the performance of CHRCS is much better than that of systems of Internet cache protocol（ICP） and cache array routing protocol（CARP）.

关 键 词：合作缓存 代理 连续哈希函数 互联网缓存协议 缓存阵列路由协议 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]