Scale-free Networking without Routing Tables

Summary form only given. The traditional approach to routing in ad hoc networks consists of maintaining routing tables listing entries for all or selected network destinations. Unfortunately, because the identifiers assigned to nodes (e.g., IP addresses or MAC addresses) in a network with mobile nodes have nothing to do with the topology of the network, network-wide dissemination of updates or queries must be used to maintain such tables, which renders the signaling of the routing protocols unscalable. This talk describes a new approach for scale-free routing that is called SURF (Scale-free Untethered Routing Framework). With SURF, the network itself assigns identifiers to nodes and updates these identifiers as the nodes move. The identifiers define a total ordering in the network with respect to one or multiple root nodes, which means that one or multiple routes from any source to any destination are defined automatically by the identifiers of the two nodes. To allow the sources to learn the identifiers of the destinations, a publish-subscribe distributed directory service is provided, such that a destination publishes its existence at an anchor node and a destination subscribes to destinations by contacting the proper anchors. The talk will show how SURF can be applied to different types of specialized networks, including those in which relay nodes have severe constraints (size, processing power, energy consumption).