Analysis of the Repair Time in Distributed Storage Systems

Overview & Synopsis

Distributed or peer-to-peer storage systems introduce redundancy to preserve the data in case of peer failures or departures. To ensure long-term fault tolerance, the storage system must have a self-repair service that continuously reconstructs lost fragments of redundancy. The speed of this reconstruction process is crucial for the data survival. This speed is mainly determined by available bandwidth, a critical resource of such systems. We propose a new analytical framework that takes into account the correlation of concurrent repairs when estimating the repair time and the probability of data loss. Mainly, we intro- duce queuing models in which reconstructions are served by peers at a rate that depends on the available bandwidth. The models and schemes proposed are validated by mathematical analysis, extensive set of simulations, and experimentation using the Grid'5000 test-bed platform.