Optimizing synchronization of cloud storage services: Combining benchmark monitoring and learning-based framework

Abstract

With data storage moving further away from locally based storage and into an age of cloud storage, users are going to need their data on the go, and they're going to need it to be fast, and they're going to need it to be accurate. Eventual consistency is the theoretical guarantee that, provided no new updates to an entity are made, all reads of the entity return the last updated value. In this paper, a comparison is made on cloud-based studies to formulate an idea as to how to improve the synchronization of cloud storage systems by first benchmarking the eventual consistency and then utilizing a framework that dynamically learns the characteristics of a storage node so that eventually consistency is achieved quickly.