Effective Resource Management in Clouds using Advance Reservation

Abstract

A Cloud is a type of parallel and distributed system consisting of a collection of interconnected and virtualized computers that are dynamically provisioned and presented as one or more uni ed computing resources based on service-level agreements established through negotiation between the service provider and consumers.Advance Reservation (AR) for Clouds is now a signi cant research focus as it allows consumers to gain synchronized access for their applications to be executed in parallel, and assures the availability of resources at speci ed future times. Clouds could be leased by the faculties of the college, who want to demonstrate the execution of the parallel programs on a distributed system during the lab hours or if a rm wants to lease the web server for months or years, a dedicated server in a data center may be from Amazon EC2 . The main objective of this work is to develop a Resource provisioning model for Eucalyptus Cloud. The Resource provisioning model proposed in this tshesis provides a support for best eff ort provisioning and AR,suitable for the above speci ed scenarios. AR of Resources o ers abetter Quality of Service (QoS) for time critical applications. An e ort is also made to discuss the possible states the Reservation made by a consumer may be in. This report also depicts how the ARM (Advance Resource Management) module can be embedded in the general architecture of the Eucalyptus Cloud. This thesis also confers a detailed discussion on Data Structures used for AR named as Eucalyptus Cloud Advance Reservation Queue (ECARQ) a terminology conied by me. Algorithms which would operate on the ECARQ are also discussed. Ubuntu Enterprise Cloud (UEC) which is the only Server with Eucalyptus cloud built in it, is used to test this Resource Provisioning Model.