Date Added: Nov 2011
In this paper, the authors present an Information Service designed for maintaining resource information in Cloud datacenters. They employ a P2P cluster of super-peers that share the resource information and load of related activities in a Distributed Hash Table (DHT) based manner. Their DHT does not employ the standard keyspace range-partitioning, but implements more complicated algorithm to enable better distribution and fault-tolerance in the Cloud datacenter context. They implement a prototype of the proposed system and conduct comparative measurements that illustrate its scalable and fault-tolerant capabilities.