Enhancing Reliability and Availability Through Redundancy in Vehicular Clouds

摘要: The past eight years have seen the emergence of vehicular clouds as a topic research in its own right. Vehicular were inspired by insight that presentday vehicles feature an impressive array on-board compute, storage and sensing capabilities. These capabilities are vast untapped resource that, at moment, is wasted. One defining ways which differ from conventional volatility. As enter leave cloud, new compute resources become available while others depart, creating volatile environment where tasks enhancing reliability availability very challenging. It intuitively clear longer more predictable vehicle residency times cloud are, easier it to ensure system availability. In this work we look with short unpredictable times. We propose enhance these types through family redundancy-based job assignment strategies attempt mitigate effect offer theoretical prediction Mean Time To Failure (MTTF) strategies. also show how fine-tune granularity redundancy order meet QoS requirements specified terms minimum MTTF for given user job. Extensive simulations, using data derived shopping mall statistics, confirmed accuracy our analytical predictions.