AntOM: Constructing multi-layer overlays for pervasive environments

Pervasive environments are becoming increasingly more popular due to the benefits of context-aware, user-centric service provisioning, despite their inherent challenges, i.e. dynamicity, heterogeneity and complexity. In such environments, multiple applications with distinct network requirements run simultaneously over the same underlying networks. Serving as abstractions to the latter, P2P overlays enable applications to operate over virtual networks that adhere to specified requirements. In this respect, we propose here AntOM, a topology optimisation algorithm for multi-layer P2P overlays on top of pervasive environments. The algorithm is inspired by Ant Colony Optimisation in regards to the network exploration and neighborhood optimisation. By utilising different ant families, multiple overlay layers, each one optimised for a different property, can be concurrently maintained at a low cost as proven by our experimental results. Therefore, applications with diverse requirements can be served at the same time by exploiting the same network infrastructure.