Abstract
To enable the evolving trend of ubiquitous computing and multi-service support of next generation ad hoc and sensor networks this paper proposes a new concept of static and dynamic application-based heterogeneity and introduces STEPS; a mathematically-based distributed multi-rate storage and data management mechanism that dynamically responds to the stochastic temporal networks' disconnectivity pattern in order to preserve the network's functionality and data under challenged conditions. Through a probabilistic analysis of collaborative storage provisioning the paper explores the performance of STEPS by investigating storage outage, weighted storage outage and storage traffic patterns. Through both analytical formulations and simulation results STEPS is confirmed to perform better than other mechanisms that have been recently proposed.