Resource constrained randomized coverage strategies for unstructured networks

Abstract

Most of the information management algorithms in large-scale unstructured networks, require one to maximize the coverage (C) i.e. expected number of distinct visited nodes. An unstructured system lacks an index. Hence, to maximize coverage, flooding-based strategies are used when there is an abundance of bandwidth (B) and single random walker is used in abundance of time (T). However, there exists an inherent tradeoff between coverage speed and bandwidth utilization. Hence, in practical scenarios, when the amount of resource, both bandwidth B and time T are finite, it is nontrivial to design strategies that maximizes network coverage. This chapter defines the extended network coverage problem under constrained resources, reveals the underlying challenge and discusses some of the recent works that design randomized strategies to maximize coverage C(B, T). Specifically, the chapter explains how the understanding of K-random walk dynamics has been used to develop uniform and non-uniform proliferating random walk strategies to achieve the goal. These coverage strategies may be useful in designing efficient services, e.g. search, gathering and routing for large scale networks like sensors, peer-to-peer, computing grids, etc.