Output details

Biased random walks on resource network graphs for load balancing

<05> The increasing complexity and distribution of computing applications such as the emerging Internet of Things/Services require novel methods to optimise and protect them. This paper proposes a novel algorithm based on a biased random walk whereby tasks or requests, emanating from outside a system, are handled in the most efficient manner: at the least loaded most appropriate part of the system. The performance of this algorithm is rigorously analysed theoretically, showing an almost optimal load balancing. This benefit is replicated in practical evaluation and the results have been used, internationally, for load balancing in newly emerging cloud computing environments.