Reliability Assessment Of Erasure Coded Systems

Replication is widely used to enhance the reliability of storage systems and protect data from device failures. The effectiveness of the replication scheme has been evaluated based on the Mean Time to Data Loss (MTTDL) and the Expected Annual Fraction of Data Loss (EAFDL) metrics. To provide high data reliability at high storage efficiency, modern systems employ advanced erasure coding redundancy and recovering schemes. This article presents a general methodology for obtaining the EAFDL and MTTDL of erasure coded systems analytically for the symmetric, clustered and declustered data placement schemes. Our analysis establishes that the declustered placement scheme offers superior reliability in terms of both metrics. The analytical results obtained enable the derivation of the optimal codeword lengths that maximize the MTTDL and minimize the EAFDL.