Research And Implementation Of Policy-based Software Self-adaptation Mechanism

The progress and integration of Personal Computing and Network Computing midwifery the era of pevesive computing.Pevesive computing software should be able to sensor the variation in it's environment and adapt itself correspondly.There are two common problems of pevesive computing software self-adaptation.On the one hand, a consistent pattern for the software is necessary to acquire the environment's states of both the physical space and the information space. On the other hand, along with the increase in complexity of pevesive computing software, flexible self-adaptive mechanisms should be integrated into the software infrastructure in order to enhance the efficiency of software development and reduce the cost.To address those challenges, this thesis proposes a policy-based self-adaptive model based on modeling the context consistently in the physical spaces and the information spaces,and introduces a XML-based policy language and the design of a policy-driven engine which is integrated with StarUAP, a pevesive computing software platform .The content of this thesis can be generalized as follows.(1) Analyzing correlative pevesive computing software self-adaptation technical, and the shortage of them.(2) Introducing a policy-based self-adaptive model. The model has three layers, including context layer, policy expression layer and self-adaptation layer. Context layer and policy expression layer are formally defined. And a detaled analysis of this model is presented.(3) Based on the self-adaptive model,a XML-based policy language is introduced and a policy-driven engine is implemented , which is one part of StarUAP and supports the self-adaptation of software.(4) Presenting the method of customizing policy and verifiying the model by applications..