A Resource Allocation Model for Desktop Clouds

Abstract

Cloud computing is a new paradigm that promises to move IT a step further towards utility computing, in which computing services are delivered as a utility service. Traditionally, Cloud employs dedicated resources located in one or more data centres in order to provide services to clients. Desktop Cloud computing is a new type of Cloud computing that aims at providing Cloud capabilities at low or no cost. Desktop Clouds harness non dedicated and idle resources in order to provide Cloud services. However, the nature of such resources can be problematic because they are prone to failure at any time without prior notice. This research focuses on the resource allocation mechanism in Desktop Clouds.The contributions of this chapter are threefold. Firstly, it defines and explains Desktop Clouds by comparing them with both Traditional Clouds and Desktop Grids. Secondly, the paper discusses various research issues in Desktop Clouds. Thirdly, it proposes a resource allocation model that is able to handle node failures.

Evaluation Metrics for VM Allocation Mechanisms in Desktop Clouds

“…DesktopCloudSim enables researchers to simulate failure events happening within the infrastructure level of a Cloud (i.e., enabling Cloud nodes to fail during run time). In order to simulate a Desktop Cloud, data of a Desktop Grid system retrieved from Failure Trace Archive was used to simulate both the infrastructure of a Desktop Cloud since both Desktop Cloud and Desktop Grid use infrastructure similar to each other (Alwabel et al, 2015a). Secondly, the archive provides name of the machine that fails along with the time of failure.…”

“…Node failure rates in Desktop Cloud are reported to be quite high and can affect the performance of Desktop Clouds (Abdulelah Alwabel et al, 2014b). It is proposed that a Virtual Machine (VM) allocation mechanism can play an important role in order to reduce the negative effect of node failures (Alwabel et al, 2015a). This paper proposes metrics that can be used to evaluate the behaviour of a VM allocation mechanism.…”