Structured Peer-to-Peer (P2P) networks enable a participating peer to efficiently route a key to its responsible peer among the current set of participants. Such deployments are characterised by churn, the continuous process of peer arrivals and departures. Churn inflicts two problems on lookup performance: 1) it results in stale routing entries, causing timeouts, which lead to the increase of lookup latency, and 2) as stale entries are evicted, routing state shrinks, causing each lookup to potentially contact multiple peers, which also increases lookup latency. To maintain low latency lookups under churn, a peer must evict stale routing entries and repopulate its routing state with new entries.
I investigate the challenges of efficient low latency lookup in structured P2P networks. I hypothesise that an adaptive approach towards dissemination of network state knowledge, where parameters such as the size, content and recipient of a message are chosen adaptively, is beneficial to the efficiency and performance of routing.
I have the privilege of being supervised by Dr. Graham Kirby. My research is funded by Scottish Informatics and Computer Science Alliance (SICSA).