Probing Based Topology Inference for Large Scale Community Networks

Traditional research in on-demand topological solutions is gathered around the two main clusters, - wireless ad-hoc networks and fixed overlay networks. Both are very different in nature, but they both deal with the same problem which is to create a topology out of an arbitrary set of nodes. This paper considers the case of an arbitrary set of mixed-technology nodes which are to be joined in a topology based on end-to-end delay measurements among nodes. The core of the proposal is topology inference based on triangular inequality of end-to-end delay which is finalized in form of an algorithm that allows for efficient detection of a logical topology of a network with no initial topology. The algorithm is scalable and could be a practical solution for many scenarios involving community services created on-demand and intended for a short lifespan.