A Self-Adaptive Service Recovery Mechanism In Ubiquitous Stub Environments

Ubiquitous networks have the basic characteristics of omnipresent, omnipotent and all-encompassing, for the goal of achieving4A communication smoothly, in which, users can enjoy ubiquitous services. Therefore, ubiquitous networks have attracted great attention from the public and the industry at home and abroad. With the devices increasingly isomerization and the networks higher speed, there are a large number of heterogeneous devices which can interconnect by heterogeneous network access mode in users’surroundings, which significantly enlarge the stub of the network. The cooperation between these stub environments and devices extend the distributed service ability to form intelligent environments, such as intelligence campus, smart home, vehicle network, personal network, opportunistic network and other new network forms.Mobile Ad hoc network becomes the typical networking mode because of its unique advantages, namely, never relying on fixed infrastructure, multi-hop and autonomy, organizing network flexibly, rapid deployment, high anti-destroying ability. Multiple device synergy and service sharing can support reliable and high quality ubiquitous service in mobile ad hoc network. However, due to limited sources, heterogeneous, device unbalanced ability, low reliability, large-scale network, dynamic change, un-centralized control and other characteristics, ubiquitous services cannot rely on single device to provide, which needs to compose atomic services provided by heterogeneous devices to support. But at the same time, devices’limited different ability and network dynamic topology caused by mobility, the network cannot ensure persistent, high quality, personalized ubiquitous service. Therefore, how to design service recovery mechanism to repair service interruption, further to improve reliability, sustainability, and quality becomes the key research in ubiquitous networks, which has high research value in ubiquitous stub environments.In order to solve the problem of service interruption, we study the service recovery mechanism from two aspects, namely, service recovery index function and service recovery strategy. To ensure the service quality, we take individual capabilities and cooperative capabilities together into consideration when designing the service recovery index function. In addition, mobility has a great influence on service in mobile ad hoc network, so mobility will be the factor in the recovery index function. In this way, the recovery index function can not only give preference to individual high ability and good coordination ability, which guarantees the quality of service; and preferentially select the device that has the biggest probability to provide service, which guarantees the reliability of service.In order to ensure the sustainability of service, provide users with seamless experience, service recovery mechanism needs to design a better recovery strategy; we propose a heuristic-based adaptive backup recovery strategy in this thesis. Using mobility prediction to select device having bigger probability preferentially, adopting dynamic update to ensure the real-time performance of the network information, backup in advance to reduce response time, applying environment perception to aware the changes of environments and device context, using cache to avoid redundant execution, sorting by priority to select optimal device preferentially; further, based on these measures, we design local recovery algorithm and global recovery algorithm according to different application scenarios, which can reduce the number of interruption and response time of service.