Opportunistic networks are (ad hoc) networks created dynamically by exploiting contacts between pairs of mobile devices that come within communication range. This networking paradigm overcomes main limitations of conventional MANETs, related to the fact that, due to mobility and energy conservation issues, it is often not practical to maintain connected multihop paths among nodes. While forwarding in opportunistic networking has been explored, investigations into asynchronous service provisioning are unique contributions of this paper. Mobile devices are typically heterogeneous, possess disparate physical resources, and can provide a variety of services. During opportunistic contacts, the pairing peers can cooperatively provide (avail of) its (other peeru27s) services. This service provisioning paradigm is a key feature of the emerging opportunistic computing paradigm. We develop an analytical model to study the behaviors of service seeking nodes (seekers) and service providing nodes (providers) that spawn and execute service requests, respectively. The model considers the case in which seekers can spawn parallel executions on multiple providers for any given request, and determines: i) the delays at different stages of service provisioning; and ii) the optimal number of parallel executions that minimizes the expected execution time without saturating providersu27 resources. The analytical model is validated through simulations, and exploited to investigate the performance of service provisioning over a wide range of parameters.
展开▼
