一种网络进程故障恢复机制的设计与实现  

Design and implementation of a mechanism to recover from failed network processes

作　　者：李海山[1] 代征[2] 杨升春[2] 袁由光[2] 

机构地区：[1]哈尔滨工程大学计算机科学与技术学院,黑龙江哈尔滨150001 [2]中国船舶重工集团公司第七九研究所,湖北武汉430074

出　　处：《计算机应用》2007年第1期242-245,共4页journal of Computer Applications

摘　　要：以商用PC/Linux为平台,设计并实现了一种在C/S工作模式下服务器网络进程故障时基于检查点的卷回恢复技术,利用该技术可以保证网络通信连接在恢复前后的一致性和连续性,从而保证网络通信的不间断、不停顿,能广泛应用于永久故障恢复、集群系统、负载均衡等领域。Based on the commercial PC/Linux platform, a new method combining rollback fault recovery technique with checkpointing technique was put forward and validated to resume failed server network processes when working in C/S model. The method was able to guarantee the consistency and persistency of connections of network communication, i.e., to guarantee that there are no stops and no stalls with network communications. The method can be used in permanent fault recovery, duster system, load balancing field, etc.

关 键 词：检查点 卷回恢复 网络进程 

分 类 号：TP302.8[自动化与计算机技术—计算机系统结构] TP306[自动化与计算机技术—计算机科学与技术]