| Since the twenty-first century,computer technology,information technology and network technology in China have been rapid development.It makes that a large number of artificial working modes in the society are replaced by the automatic working modes with the computer systems.In this process,those computer systems for customer service extend their running time continuously in order to better meet customer demand whenever and wherever possible.They gradually develop into 24 hour uninterrupted operation systems.However,the computer systems running continuously for a long time,especially their important software systems would cause the continued degradation of the internal state and the software performance decline,increased loss of efficiency and a series of problems because of reasons of calculation error and data corruption,lots of memory leak.This phenomenon is the software aging.For the safety critical systems such as the aerospace,defense,energy and finance systems,the internal software aging not only will cause serious economic losses,but also seriously threaten the personal safety of people.Therefore,for this kind of application system,in order to minimize or avoid the huge losses of software aging problems,it is urgent to reasonably and accurately assess the aging condition of current software systems.Then appropriate intervention measures can be taken in a timely manner to ensure that the whole system can reliably for long time.Aiming at the problem of state estimation software aging,this dissertation firstly presents the influence factors of software aging and their classification by combining with the existing research results based on the comprehensive analysis of the running process and environment of aging software with long running time.Secondly,for the problem that using all factors to evaluate the state of software aging will result in unacceptable complexity and efficiency,this dissertation presents the attribute reduction algorithm that can be used to handle both continuous and discrete variables and the method to determine the weights of all influence factors based on rough sets theory and attribute reduction technology.Thirdly,aiming at the uncertainty problems and qualitative analysis process in the analysis and evaluation of software aging,this dissertation presents a multidimensional weighted method to determine membership function for the accurate description of fuzzy sets in software aging evaluation based on the fuzzy theory and the characteristics and the weights of impact factors after the attribute reduction.It elaborates the implementation process of the fuzzy evaluation method of software aging on the basis of the forward method.Finally,this dissertation selects the streaming media Helix server as an example to verify the environment and the research object,and gives the experimental verification of the method.The experimental results show that the fuzzy evaluation method proposed in this dissertation is better than the existing evaluation methods.The research results of this dissertation reveal the essence of software changes and contribute to a better accurate assessment for aging state of a software system.It is also the necessary prerequisite for the research and implementation of the control technology to software rejuvenation and finally lays a solid foundation for the target of long time reliable operation of the software systems including safety critical systems,and so it has good social and economic benefits. |
