PROJECT TITLE :

Network-Aware Feasible Repairs for Erasure-Coded Storage - 2018

ABSTRACT:

A important amount of analysis on using erasure coding for distributed storage has focused on reducing the quantity of information that should be transferred to replace failed nodes. This continues to be a vigorous topic as the introduction of faster storage devices appearance to put an even larger strain on the network. But, with some notable exceptions, most published work assumes a flat, static network topology between the nodes of the system. We tend to propose a general framework to seek out the lowest price feasible repairs during a more realistic, heterogeneous and dynamic network, and examine how the amount of repair ways to consider will be reduced for three distinct erasure codes. We devote a important half of the paper to determining the set of possible repairs for random linear network coding (RLNC) and describe a system of economical checks using techniques from the arsenal of dynamic programming. Our answer involves decomposing the matter into smaller steps, memorizing, and then reusing intermediate results. All computationally intensive operations are performed prior to the failure of a node to confirm that the repair can start with minimal delay, primarily based on up-to-date network info. We tend to show that each one 3 codes benefit from being network aware and notice that the extra computations needed for RLNC can be reduced to a viable level for a big selection of parameter values.