Research On Service Migration For OSGi Based Pervasive Cloud

With the development of computing technologies and popularity of smart phones,cloud computing,mobile computing and pervasive computing have been new computing paradigms which to meet the requirement of computing and service capabilities.But we can see that they all have disadvantages respectively.In this regard,the integration of the three paradigms is the trend and pervasive cloud computing comes into being.Pervasive cloud computing is actually a further extension of mobile cloud computing.It integrates mobility of mobile computing,powerful computing and storage capacity of cloud computing into the pervasive computing concept,making full use of not only the lightweight and convenience of mobile devices,but also powerful capacity of cloud computing,to provide more high quality and convenient service.Pervasive cloud computing now plays an important role in national policy,the development of enterprises and people's daily life.But at the same time it is more difficult to be managed because of its complexity cause by it distributed and heterogeneous characteristics.To this problem,migration and scheduling of services and data in pervasive cloud environment is a feasible solution.With migration,pervasive cloud environment can provide services with high availability,stability,flexibility and performance.OSGi,as the Java standard modular technology,builds systems in the form of components and bundles.It has been widely supported and applied since proposed.At the same time,because of the characteristics of its service registration and discovery mechanism,OSGi is often used to develop a dynamic and deployable system,so it has a good adaptability to the research on service migration in pervasive cloud environment.In this paper,a method for service migration in an OSGi based pervasive cloud is proposed.Firstly we build a pervasive cloud service migration infrastructure based on OSGi,which combines D-OSGi and R-OSGi distributed OSGi deployment mechanism in order to adapt to the PC environment and intelligent mobile platform.And kinds of monitoring technologies are used to monitor the pervasive cloud environment status and obtain performance indicators of OSGi services running in it.After that a deep learning network LSTM is used to predict following resource status and we establish a service migration mechanism based on monitoring and prediction of resources in order to optimize the system and services execution.And then service migration is easily carried out according to the migration decision making use of the characteristics of modularity and loosely coupled of OSGi.We test and analyze the migration method of OSGi services in the cloud environment including the accuracy of monitoring and prediction and effectiveness of the migration decision method.The experimental results show that the proposed method has good applicability in the pervasive cloud environment.