Service Provisioning Based on Advanced Reservation and Deadline Sensitive Leases in Iaas Cloud

Abstract

The need for computational resources are increasing day by day and has become a fundamental requirement in industry. IaaS cloud provides computing resources and storage as standardized services for users over the network. Computational resource provisioning has three dimensions: hardware resource, software stack available on those resources, and the time during which those resources are available to users. Lease based approaches and job based approaches are two different approaches used for provisioning computational resources. Virtual machines, a lease based approach can be used to provision resources across these dimensions and therefore considered as an appealing vehicle for computational resource provisioning. Different possible types of service request for IaaS cloud are advanced reservation, immediate, deadline sensitive and best-effort. A service provisioning mechanism which is a lease based approach is proposed that will schedule the different types of service requests in a IaaS cloud. The proposed algorithm is implemented in Java that also enacts the schedule by issuing OpenNebula commands to launch, terminate, suspend, resume the virtual machines and manage the computational resources. Experiments are conducted and results show that the proposed algorithm can schedule more service requests than existing approaches and the system utilization is improved because of reduced resource fragmentation.