Enhancing dependability through redundancy in military vehicular clouds

The phenomenal success of cloud computing has inspired researchers to introduce the concept of a vehicular cloud (VC). In this work we introduce the concept of a military vehicular cloud (MVC) that specializes VCs to low/controlled mobility situations. Specifically, we envision a MVC involving either parked vehicles or vehicles that move in formation so that the relative distance between neighboring vehicles is, essentially, the same. This is a scenario that applies to military units deployed in support of a tactical mission. We anticipate that the vehicles in the MVC are provided with a wireless connection to an (often mobile) access point which, in turn, is connected in some fashion to a central server. A key difference between MVCs and conventional clouds lies, in addition to mobility, in the unpredictable availability of computational resources. As vehicles enter the MVC, new computational resources become available; when vehicles depart, their onboard resources depart as well, creating a volatile environment where ensuring dependability is a major challenge. Our main contribution is to enhance the dependability of MVCs through a family of redundancy-based job assignment strategies that mitigates the effect of resource volatility in MVCs. We offer a theoretical prediction of the mean time to failure (MTTF) of these strategies. Extensive simulations have confirmed the accuracy of our predictions.