AlanBrownphd
Former PhD Student at the University of Stirlingabout.my research
At Stirling we are interested in researching enabling technologies for Service Discovery in Wide Area Networks (WANs). Jointly with the Princeton Panasonic Research Laboratory (PPRL) in the USA, we are exploring a number of techniques to provide a large range of devices with the means to discover services. Services are functions or features available to a networked device or computer. For example, a networked device may be able to locate and use a printer or scanner service, or access Web Services on the Internet. A more futuristic view could see travelling vehicles accessing real-time traffic statistics for their journey through inter-vehicular communications.
Service Discovery techniques can be distributed or provided centrally. Traditionally, centralised approaches are discouraged due to poor scalability, central points of failure and inflexibility. Alternatively, distributed approaches such as Peer-to-Peer (P2P) systems have been proven to be scalable, fault tolerant and flexible. Subsequently much of our research at Stirling is focused around Service Discovery in P2P systems.
Although P2P systems have potential, they are not without their faults. Even though highly scalable, many existing designs suffer from inefficiencies; thus getting good performance from P2P systems can be expensive. In our work, we have looked at several methods to improve P2P efficiency, lowering the barrier of entry for modern, mobile networked devices. For example, we have recently shown through simulation that bandwidth savings of at least 30% can be achieved by implementing multi-destination multicast routing in traditional P2P systems. We have recenrly designed a new, variable-hop overlay system called Chameleon, which offers performance according to a participating nodes available bandwidth and resources; a unique system design which aims to open WAN Service Discovery up to a vast range of mobile networked devices.