QARPF: A QoS-Aware Active Resource Provisioning Framework Based on OpenStack

On-demand resource allocation is a crucial feature of Infrastructure as a Service (IaaS) in cloud computing environment. Aiming at the problem that existing cloud platforms cannot conduct active resource provisioning according to the Quality of Service (QoS) requirements of tenants' tasks, this paper designs and implements a resource actively provisioning framework based on OpenStack, which takes into account both the resource utilization efficiency of data center and the QoS requirements of task. The proposed framework consists of three stages: (1) tasks are encapsulated and forwarded according to resource requirements and time characteristics of tasks; (2) workload is dynamically analyzed based on priority-aware task scheduling model; (3) and efficient mapping between tasks and resources is accomplished based on QoS-aware resource provisioning model. Experimental evaluation on CloudSim simulation platform shows that the proposed framework allows tenants to directly submit task requirements to the cloud platform, and actively realize resource-efficient provisioning, which effectively improves the resource utilization efficiency of data center, reduces the task completion time and virtual machine (VM) rental cost of tenant.