Approach to Big Data Applications on Cloud System
Jin‐Hong Kim, Sung-Tae Hwang · Advanced science and technology letters · 2015
Consistency management within cloud storage systems is of high importance. The Consistency–Performance tradeoff is, arguably, the main tradeoff. Many static consistency solutions fail in reaching an efficient equilibrium between consistency and performance for dynamic cloud workloads. In this context, more opportunistic, adaptive consistency models are needed in order to meet the requirements of Big Data Applications (BDAs). However, most of the existing adaptive policies either lack automation or fail to apprehend and include the specific consistency requirements of the application outside its access pattern. Hereafter, in this paper, we tackle this specific issue of consistency management for dynamic workloads in the cloud. Accordingly, we provide an adaptive model that tunes the consistency level at runtime in order to provide consistency when needed and performance when possible.