A Method Of QoS-aware Service Components Composition For Pervasive Computing Environments

Currently, pervasive computing is a hot research area. The theory and technology of pervasive computing focus on how to make efficient use of these available resources to satisfy functional and nonfunctional requirements of user tasks. Today, pervasive computing technologies and the associated software are being developed to facilitate such applications as telemedicine, education, manufacturing and military for all the time and everywhere use. These applications demand automated,continual, QoS(Quality of Service)-guaranteed services in pervasive computing environments. QoS-aware service components composition has the practical value for those application areas. Thus, a method of QoS-aware service components composition is proposed in this paper. The main points of this dissertation are described as follows: (1) It designs a QoS model for service components in pervasive computingenvironments, named QuCOM(Quality Component).Due to advantages of component technology, such as reusability, configurablity and adaptability, it is a efficient and feasible method to compose service with components,called component-based service in this paper,in dynamic and heterogeneous environments. However, the first challenge of how to design QoS-aware service components composition method is that the application cannot receive any QoS guarantees from existing component technology. Therefore, firstly this paper presents the component-based service model, then describes QuCOM model for service components. Not only are the features of service components supported in QuCOM model, but also developers are allowed to define different QoS modes for a component using QuCOM in order to make component-based services adapt to dynamic heterogeneous environments. The general service component model and QuCOM are the foundation of the composition method presented in this paper.(2) A method of QoS-awarc service components composition is put forward based on QuCOM.Currently, there are many service composition technologies. However, these QoS-aware service composition methods are limited. And they are not suit for pervasive computing application developed with QuCOM model. Firstly, most of them are designed for enterprise computing. Their assumptions are not practical for services in pervasive computing. Second, the dynamic and heterogeneous features of pervasive computing environments aren't considered in most methods. At last, most previous solutions can only support linear composition structure with fixed composition order, which limit the flexibility and efficiency of service composition. Thus, in order to solve those problems, a method of QoS-aware service components composition is put forward based on QoS negotiation model of QuCOM. And relative algorithms and their experiments are described in this paper.(3) It proposes an adaptation method of service components based on resource container in pervasive computing environments.To make service adapt to dynamic heterogeneous environments, this paper presents service components adaptation method for composited service based on resource container concept. Using this method, service components can be deployed dynamically during service setup, which overcomes resource limitations of embedded devices in pervasive computing environments and provides better QoS for users; And QoS contract of service components is controlled by adapting its resource reservation based on feedback model, which guarantees QoS of service components in dynamic environments.(4) The QuCOM prototype for QoS-aware services composition, called LiquidQ, is designed and implemented.In order to apply the QoS-aware service components composition method in real world and test the theories and approaches presented, the QuCOM prototype for QoS-aware services composition, called LiquidQ, is designed and implemented based on CORBA component model. In this paper, we describe the details of LiquidQ: its architecture, the relative interfaces of service component and some relative system modules including service manager, service planner, component...