Chapter

Cloud Computing and Services Science

Part of the series Service Science: Research and Innovations in the Service Economy pp 315-332

Date:

A Performance Evaluation of Block I/O Paravirtualization and Virtual Machine Images

  • Django ArmstrongAffiliated withSchool of Computing, The University of Leeds Email author 
  • , Karim DjemameAffiliated withSchool of Computing, The University of Leeds

* Final gross prices may vary according to local VAT.

Get Access

Abstract

Cloud Computing continues to be a rapidly evolving and prevalent paradigm where Quality of Service (QoS) has a pivotal role to play in guaranteeing performance and provisioning of resources on-demand in a timely fashion. Improvements to the performance of Cloud technology ensure provider profitability and an increased number of applications that can make use of a Cloud where overheads would have otherwise limited usage. This paper presents the results of a quantitative evaluation into the performance overheads of accessing images, via a Hypervisor’s virtual block I/O device using various image types. Two Virtual Machine Managers (VMMs) are evaluated: XEN and KVM alongside the image types they support. Benchmark results demonstrate the performance of XEN exceeding KVM in a greater number of benchmark tests. Conclusions are drawn from the results on the suitability of these technologies for data intensive applications and applications requiring highly dynamic resource sets, where making an uninformed decision on what technology or image type to use could prevent an application reaching its full potential.

Keywords

Cloud computing Performance evaluation Paravirtualization Virtual machine images