Enhancing DHT-based object naming service architectures with geographic-awareness

Existing Object Naming Service (ONS) architectures that are based on Distributed Hash Tables (DHT) are built on top of chord-like DHT networks, which are DHT P2P networks constructed by projecting the network nodes on a ring network and then adding long edges to each node to improve the lookup performance. A main weakness in these architectures is the lack of geographic awareness at the nodes. This paper proposes the enhancement of these architectures with geographic awareness using a technique, called Geographic-Aware Content Addressable Network (GCAN), that runs on top of any chord-like DHT network. GCAN uses the procedures of chord-like DHT networks as black-boxes. Thus, it requires only minimum additions to existing DHT-based ONS architectures. As a result, it inherits the scalability, reliability, and the maturity of chord-like DHT networks. DHT-based ONS architectures that are built with GCAN are guaranteed to have routing, join, leave complexities in O(log n), while the routing table size is also in Θ(log n) on average, where n is the number of the network nodes.