Scheduling Service with SLA assurance for private cloud systems

Abstract

The Cloud systems performance consists in its agile response to the user needs and in the Quality of Service provided. In this context, one of the biggest challenges facing cloud computing nowadays is the Service Level Agreement (SLA) assurance, which will increase the user's confidence in the system. We present in this paper a model for a scheduling service not only to respect the Service Level Agreement (SLA) constraints but also to minimize the total cost of resource utilization. Based on this model, we outline a scheduling application architecture for a system that will receive SLAs and will provide the allocation schema for the existing hosts. The application is implemented as an Haizea add-on. Haizea is virtual machine-based lease management system built on the top of OpenNebula environment. We tested and validated the proposed model using the simulator offered by Haizea.