New Applications Of Data Redundancy Schemes In Cloud And Datacenter Systems

Abstract

Data redundancy techniques such as replication and erasure coding have been studied in the context of distributed systems for almost four decades. This thesis discusses new uses of erasure coding and replication in the context of cloud computing and datacenter systems. It shows how erasure coding can be used to protect against vendor lock-in and how the interplay between replication and data sharding can mitigate the cost of configuration management.