基于非精确网络状态信息的多服务器选择

服务器复制经常被用来作为改进处理大量客户请求的服务能力的方法．在采用服务器复制时．一个关键的问题是客户如何确定哪一个服务器是最好的,通过一个简易的服务器模型，研究了在延时非精确的网络情形下，通过采用智能的票据探测技术，同时搜索延时较小的网络路径并探测目的服务器性能来进行多服务器选择的算法．仿真结果显示，与其它几种多服务器选择算法比较，由于综合考虑了服务器延时和网络延时的最优，我们的算法可以取得最好的性能．

Multi-Server Selection with Imprecise State Information

Server replication is an approach often used to improve ability of a service to handle a large number of clients. When replication is used, the primary concern is how a client may discover which server is the best to use. This paper adopts a simple server model and investigates an efficient kind of algorithms for multi-server selection under the condition of imprecise network delay, using approach of ticket-based probing to search the path with short delay and probe the server capability at the same time. Compared with several other multi-server selection algorithms, the simulation results demonstrate that since our algorithms try to optimal the server and the path delay at the same time, they can achieve the best performances.