Fault-tolerant mechanism combined with replication and error correcting code for cloud file systems
Yang Dongr · Journal of Tsinghua University(Science and Technology) · 2014
Fault tolerant is important for the reliability of cloud storage file systems.This paper analyzes the reliability of typical cloud file systems with a central metadata server and proposes a fault-tolerant mechanism that combines replication schemes with error correcting codes for storage node reliability guarantee as well as a hot-standby scheme for the metadata server reliability guarantee. Experimental results demonstrate that the mechanism improves the reliability of current cloud storage file systems and at the same time improves the storage utilizations compared with replication schemes.