服务器池的高可用N+1冗余结构模型  

High Availability N+1 Redundancy Structure Model of Server Pool

作　　者：丁晓光[1] 古志民[1] 

机构地区：[1]北京理工大学计算机科学技术学院,北京100081

出　　处：《计算机工程》2007年第15期112-114,117,共4页Computer Engineering

摘　　要：Linux virtual server是典型的服务器机群结构,通过双前端负载分配器和后端服务器池提供高可用服务。当后端服务器池发生故障时,故障服务器的工作量将通过负载分配算法重新分配到其他可用的服务器上,结果导致服务器池负担变重,进而可用服务器的故障概率增加。该文设计并实现了高可用N+1冗余结构模型系统,改进了典型的服务器池的上述缺陷。实验表明模型有着极低的资源占用率和非常短的故障切换时间。Linux virtual server is a typical server cluster, which provides high availability through dual load balancer and server pool. When server pool goes wrong, the workload of failure server will be redistrubuted to other available servers by using load distribute algorithm, as a result that server pool has heavy burden, and failure probability of available server increases greatly. High availability N＋I redundancy structure model of server pool is designed and implemented, in order to solve the problems. Experimental results present that the model costs a few system resources and little time in a period of one time takeover.

关 键 词：高可用性 服务器池 服务切换 心跳技术 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]