Open Access Open Access  Restricted Access Subscription Access

A Fairness Framework for Resource Allocation in Cloud Computing


Affiliations
1 Department of Computer Science & Information Technology, Central University of Jammu, Jammu and Kashmir, India
 

Cloud computing is a model for empowering request to arrange access to a common pool of configurable registering resource. In distributed computing frameworks the resource pool is built from an extensive number of heterogeneous servers. The multi-resource distribution component, called DRFH is a Predominant Resource Reasonableness (DRF) from a solitary server to various heterogeneous servers. The DRFH has various profoundly attractive properties. With DRFH, no client lean towards the allotment of another client; nobody can enhance its portion without diminishing that of the others; and all the more imperatively, no client has an impetus to lie about its resource request. As a direct application, we outline a straightforward heuristic that actualizes DRFH in true frameworks. Expansive scale reenactments driven by Google group follows demonstrate that DRFH altogether outflanks the conventional opening based scheduler, prompting substantially higher asset/resource use with considerably shorter occupation finish times.
User
Notifications
Font Size


  • A Fairness Framework for Resource Allocation in Cloud Computing

Abstract Views: 395  |  PDF Views: 0

Authors

Imtiaz Ahmed
Department of Computer Science & Information Technology, Central University of Jammu, Jammu and Kashmir, India
Yashwant Singh
Department of Computer Science & Information Technology, Central University of Jammu, Jammu and Kashmir, India

Abstract


Cloud computing is a model for empowering request to arrange access to a common pool of configurable registering resource. In distributed computing frameworks the resource pool is built from an extensive number of heterogeneous servers. The multi-resource distribution component, called DRFH is a Predominant Resource Reasonableness (DRF) from a solitary server to various heterogeneous servers. The DRFH has various profoundly attractive properties. With DRFH, no client lean towards the allotment of another client; nobody can enhance its portion without diminishing that of the others; and all the more imperatively, no client has an impetus to lie about its resource request. As a direct application, we outline a straightforward heuristic that actualizes DRFH in true frameworks. Expansive scale reenactments driven by Google group follows demonstrate that DRFH altogether outflanks the conventional opening based scheduler, prompting substantially higher asset/resource use with considerably shorter occupation finish times.

References