SLA-aware resource reservation management in cloud workflows
Huifang Li, Xiaochen Gao, Yanjiao Di · 2015
As more and more applications are deployed on the cloud platform, resource reservation guarantees the resources will be available at the needed execution time while meeting the Quality of Service (QoS) constraints. However, large proportion of resource reservation requests could degrade system performance significantly. This paper focuses on improving system performance of resource reservation on the background of cloud workflows. First, this paper argues that only critical tasks in the workflow need to be reserved in advance. Then it proposes a reservation request model with relaxed completion time and creates a fee model of the task which is specified in Service Level Agreement (SLA). Finally, we propose an admission control algorithm called Revenue Optimization Resource Reservation algorithm (RORR). It implements the control of reservation request with relaxed completion time while maximizing the revenue of cloud providers. Simulation results validate that the proposed algorithm has a progress in system utilization, requests admission ratio and cumulative revenue compared with the existing algorithm.