A Hybrid Heuristic for QoS Aware Matching of User's Job and Virtual Machines in Cloud Environment

A Hybrid Heuristic for QoS Aware Matching of User's Job and Virtual Machines in Cloud Environment

Devki Nandan Jha (Newcastle University, Newcastle upon Tyne, UK) and Deo Prakash Vidyarthi (Jawaharlal Nehru University, New Delhi, India)
Copyright: © 2018 |Pages: 22
DOI: 10.4018/JITR.2018040106

Abstract

Cloud computing is a technological advancement that provides services in the form of utility on a pay-per-use basis. As the cloud market is expanding, numerous service providers are joining the cloud platform with their services. This creates an indecision amongst the users to choose an appropriate service provider especially when the cloud provider provisions diverse type of virtual machines. The problem becomes more challenging when the user has different jobs requiring specific quality of service. To address the aforementioned problem, this article applies a hybrid heuristic using College Admission Problem and Analytical Hierarchical Process for stable matching of the users' job with the cloud's virtual machines. The case study depicts the effectiveness of the proposed model.
Article Preview

1. Introduction

Cloud computing provides computing as a utility and allows its ubiquitous access to anyone from anywhere at any time. It offers enormous benefits including pooled resources, easy scalability, easy accessibility, measured service, etc. that attracts the wide users and the enterprises. Amazon, Google, Rackspace, Soft layer, Azure, and many more are some common cloud service providers offering their services. There remains some issues in cloud regarding resource usage that hinders its proper expansion (Armbrust et al., 2009, 2010). Many organizations e.g. IDC1, Gartner2, Forrster3, etc. are specifically involved in identifying such issues. Some of the common issues, identified by all these organizations, are security and privacy, interoperability, costing/pricing and effective resource provisioning. Security is considered to be the most affecting issue as after uploading the data to the cloud, user somehow loses control over it. Before accepting the services from the cloud, the user needs to be convinced by the service provider that the requested services are properly met without any data loss. In fact, security is one of the important reason dithering the common users to hook up to cloud services.

Virtualization is the core concept of cloud computing as it creates multiple virtual resources, in the form of virtual machines, onto the limited physical resources that can easily be allocated to any user’s job (Armbrust et al., 2010). A cloud service provider (CSP) contains multiple VMs with varying capability (computational speed, memory, storage, bandwidth, etc.). The VM prices also vary depending upon its configuration. Also, all the CSPs advertise their services to be the best making it difficult for a user to select the appropriate one. The cloud offers and attracts a number of users to utilize its services. In order to satisfy the request of a user, the services must be provided in accordance with the desired level of Quality of Service (QoS) (Abdelmaboud, Jawawi, Ghani, Elsafi, & Kitchenham, 2015). CSP insures that sufficient resources are allocated to the users in order to meet its QoS requirements. To avoid any QoS violation, a written service level agreement (SLA) is signed between a user and the cloud service provider. Considering that the demand of the user may surge to some extent in future, a CSP is chosen that can easily be scaled to the increasing needs of the user. Thus, matching a users’ demand of the cloud resources and available VMs is a non-trivial problem.

The proposed work defines a matching between the user’s job and provider’s VM with the aim that both are mutually benefitted with optimal resource utilization. First, the users’ jobs and VMs are ranked separately using Analytical Hierarchical Process (AHP) (Satty, 1986). Next, matching of the users’ job and VM is done using College Admission Problem (Gale & Shapley, 1962), a variant of stable marriage problem (McVitie & Wilson, 1971), that can easily be used for matching of VMs and user’s job. The result gives us a stable match between the jobs and VMs.

The organization of this paper is as follows. The problem, addressed in this work, has been stated in section 2. It also discusses the background necessary to understand the proposed work. Section 3 outlines the proposed model for matching of user’s job and VMs describing the methodology used. The complexity of the proposed model is also evaluated and presented in this section. A case study, to show the effectiveness of the proposed model, is given in section 4. Section 5 discusses some recent related works and its comparison to the proposed work. Finally, the conclusion with some future research directions is presented in section 6.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 12: 4 Issues (2019): Forthcoming, Available for Pre-Order
Volume 11: 4 Issues (2018)
Volume 10: 4 Issues (2017)
Volume 9: 4 Issues (2016)
Volume 8: 4 Issues (2015)
Volume 7: 4 Issues (2014)
Volume 6: 4 Issues (2013)
Volume 5: 4 Issues (2012)
Volume 4: 4 Issues (2011)
Volume 3: 4 Issues (2010)
Volume 2: 4 Issues (2009)
Volume 1: 4 Issues (2008)
View Complete Journal Contents Listing