Efficient collaboration in opportunistic networks

Abstract

Content sharing, dissemination and collaboration in open environments require the ability to send and combine data packets across network even in the absence of an end-to-end connected path [1]. User mobility plays a fundamental role in selection of intermediate relays for multi-hop communication [2]. Current schemes do not take into account the effect of power law distribution of inter-contact times, flight lengths and pause duration between flights that are shown to be statistically significant in real traces [3]. We analyze how these properties can be utilized for forwarding data and composition of services available at devices in an open environment (parks, malls, streets in a city etc.) where citizens interact. Contributions of this dissertation include: i) an analytical model to infer user's location over time; ii) a distributed algorithm with low overhead and complexity to infer network topology; iii) a scheme for efficient route discovery between mobile users; and iv) a scheme for scheduling requests to compose services available across the network to provide enhanced functionality in minimal time. Initial simulations on simple forwarding in real environments show improved delivery ratios and reduced delays in comparison to existing schemes.