Combination of Data Deduplication and Redundancy Techniques in Distributed Systems

Inhalt:
In this paper a data deduplication algorithm is extended to be executed on a group of computer nodes. The deduplication algorithm detects equal blocks on different storage nodes and uses this information for deduplication, but also to keep block replicas in the group. We describe the protocol and report on experiments that demonstrate the feasibility of the approach, and quantify the cost of coordination.