Reliability-Oriented Service Discovery Mechanism In Ubiquitous Stub Environment

"U-Calculation" or "Ubiquitous Calculation" or "Pervasive Calculation" is a type of technology which has overstepped traditional desktop calculation."Ubiquitous Calculation" has two characteristics:the first characteristic is devices can access to information at anytime and anywhere. The second characteristic is the services are invisible and users can enjoy services unawareness."Ubiquitous network" is a new kind of network based on "Ubiquitous calculation" technology. Ubiquitous network’s target is provide4A services (at anytime, at anywhere, anyone can use and anything). Ubiquitous network’s characteristics are the services in ubiquitous network are omnipresent, all-powerful and all-embracing. Along with the expansion of network size, the number of network edge device is increasing. The heterogeneous terminal equipment connect to each other through P2P and not through access node AP. MANET is a way to organizing ubiquitous network. There are no infrastructure to store devices’position and service information. The network can self-organize and self-cure. Through devices’cooperation and services’sharing, MANET provide rich services. In ubiquitous stub environment, devices can move dynamic and supply energy by battery. Hence the cooperation service may interrupt. Service interruption will cause long response delay, a waste of network resource and device resource. The goals of reliability-oriented service discovery mechanism are selecting service provider which has minimum service interruption times and improve users’experience. Therefore it is important to research reliability-oriented service discovery mechanism.To solve the problems in service discovery, we first analyze the reasons of unreliable service. Then, at the basis of analysis we propose the Reliability-Oriented Service Discovery Mechanism (ROD) and divide ROD into three parts:Predicting Mobility Model (PMM), Maximum Reliability Clustering Algorithm (MRCA) and Reliability-Predicted Finding&Selecting (RPFS). In PMM, predict devices’velocity. In MRCA, considering devices’remain battery energy, CPU calculating power, degree and velocity, we select best node as cluster node. RPFS’s goal is improve the successful service finding rate and decrease the failed service interruption rate. We select the provider and path with maximum reliability. The simulation result shows ROD can reduce the failure rate of searching service and the interrupt rate of execution, prolong nodes’time and decrease energy loss. It proves that ROD can improve the reliability of service discovery mechanism and reduce the energy loss.