Self-organising Clusters in Edge Computing

Computation-intensive applications generally require a high computing capacity for data processing and storage that cannot be easily offered by a single Internet of Things (IoT) device. Such limitations can be successfully addressed by offloading processing and storage from resource-constrained devices to more powerful ones. In this context, edge computing is emerging as a valuable approach, since it allows data to be stored and processed closer to where it is created instead of sending it across long routes to data centres or clouds. We are interested in supporting spontaneous and opportunistic behaviour in this new dynamic environment, where computational power and storage capacity can be offered from the edge, with low latency and high bandwidth, by enabling cooperation between a subset of available edge nodes. We argue that cluster formation is necessary when a single node cannot execute a specific service fulfilling the imposed nonfunctional requirements, and it may also be beneficial when groups perform more efficiently when compared to a single's node performance.