论文部分内容阅读
This thesis designs and implements a Network-based Monitoring System,which is used to monitor the nodes of Information Analysis System of Large-scale Network.The Monitoring System monitors the application state,the network connection state,the application data and the system information of all of the nodes.It also has the ability to judge the aging degree of the software system running on the nodes,and forecast the degressive trend of the software performance.
The notion of software aging was proposed in 1995 by the researchers of AT&T Bell Laboratory.It refers to the phenomenon that the state ofthe software system degrades with time.This phenomenon does great harm to the software system.It eventually leads to seriously performance degradation of the software and even crash failure at last.Some people proposed software rejuvenation technology to deal with software aging.Its a concept of gracefully terminating an application and immediately restarting it at a clean internal state.
This thesis proposes a measurement-based approach to judge the degree of software aging and forecast the degressive trend of the software performance.This approach studies the relationship between OS resource and performance,then analyse the law of the OS resource,and forecast the situation of it by time-series modeling,and forecast the situation of the system performance at last. After discussing the approach that proposed in this thesis particularly,the author brings forward the further improving plan for the study ofsoftware rejuvenation,and this lays a solid foundation for future work.