Dynamic Energy and Expenditure Aware Data Replication Strategy

Abstract

Nowadays, data are generated and accessed from all over the world. Applications and companies store these data on geo-distributed cloud providers that have to be profitable while reducing their environmental impact. As Cloud providers aim to satisfy their service level objectives in terms of availability and response time, they often rely on data replication. In this paper, we propose a dynamic data replication strategy (DE2ARS) that adapts the number of replicas according to the workload and addresses energy consumption and expenditure issues. It follows an initial placement and is triggered by a Control Chart. We first compare different parameter choices in order to provide better analysis for the proposed strategy. We compare DE2ARS with strategies from the literature. Results highlight the fact that DE2ARS reaches its goal to reduce both energy consumption and expenditure while having good performance in a strong workload context.

Review on data replication strategies in single vs. interconnected cloud systems: Focus on data correlation‐aware strategies

“…To ensure a reduced response time, some replication strategies rely on a definition of thresholds. In this respect, the replication strategies are triggered in accordance with the values of these thresholds 20,30,39,40,64,68,83,89,91,118,128,129 …”

“…These strategies rely on prediction models and they are triggered based on an estimated value of response time. While the strategies presented in References 20,64,91,118, and 129 consider the replication only if the estimated response time of the tenant task is greater than the response time threshold specified in the SLA. This estimation is performed before the tenants' task execution.…”

“…In this respect, the replication strategies are triggered in accordance with the values of these thresholds. 20,30,39,40,64,68,83,89,91,118,128,129 Some strategies consider response time as the SLO to be achieved for each tenant. 68,[87][88][89] These strategies rely on prediction models and they are triggered based on an estimated value of response time.…”

“…On their side, other strategies focus more on the impact of energy consumption when replicating data. 59,63,129…”

“…Therefore, the majority of the strategies are evaluated and implemented based on simulation tools while considering real-world cloud parameters. 20,30,32,35,36,40,[59][60][61]63,64,72,73,75,77,83,85,109,121,122,129 Simulation offers many advantages, such as re-producibility, cost-effectiveness and flexibility. 134 Indeed, it enables the management of setup settings as well as the design of test scenarios that can be repeated several times.…”