Dynamic Algorithms for Self-Deployment and Self-Configuration of Pervasive Service Components

Pervasive services have attracted numerous attentions from both industries and academia. Its very high adaptability demand makes autonomic computing (AC), and in particular, personal autonomic computing (PAC), an ideal enabling technology. Some generic AC framework has been proposed in the literature but without considering the dynamic deployment and re-configuration of pervasive service components in a self-controlled manner. This paper makes some preliminary attempts towards these two issues. In particular, a service component self-deployment algorithm using partitioning techniques is proposed and evaluated. A simple service reconfiguration algorithm is also presented. In addition, an AC environment is proposed and its major building blocks being discussed. This AC environment serves as the container of the proposed service component self-deployment and self-configuration mechanisms. A prototype of the system has been designed and implemented. Experiment results have shown the effectiveness of the proposed mechanisms.