Predicting application performance for multi-vendor clouds using dwarf benchmarks

Authors:
Vegard Engen;Juri Papay;Stephen C. Phillips;Michael Boniface
Affiliations:
IT Innovation Centre, University of Southampton, U.K.;IT Innovation Centre, University of Southampton, U.K.;IT Innovation Centre, University of Southampton, U.K.;IT Innovation Centre, University of Southampton, U.K.
Venue:
WISE'12 Proceedings of the 13th international conference on Web Information Systems Engineering
Year:
2012

Citing 4
Cited 1

The Case for Application-Specific Benchmarking

HOTOS '99 Proceedings of the The Seventh Workshop on Hot Topics in Operating Systems
Application-specific benchmarking

Application-specific benchmarking
A view of the parallel computing landscape

Communications of the ACM - A View of Parallel Computing
Snow White Clouds and the Seven Dwarfs

CLOUDCOM '11 Proceedings of the 2011 IEEE Third International Conference on Cloud Computing Technology and Science

A Walking Dwarf on the Clouds

UCC '13 Proceedings of the 2013 IEEE/ACM 6th International Conference on Utility and Cloud Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Future Internet applications are becoming increasingly dynamic and can be composed of a wide range of services controlled and hosted by different stakeholders. This paper addresses the challenge of resource provisioning for applications that have specific Quality of Service (QoS) requirements and where consumers of Cloud resources want to avoid lock-in to any specific Infrastructure-as-a-Service (IaaS) provider. Application modelling can be used to predict performance of applications given certain resources, workload and configuration. However, application modelling is a significant challenge for Cloud consumers due to the limited and varying information IaaS providers disclose about infrastructure resources. We demonstrate in this paper how Dwarf benchmarks can be used as a uniform and informative way of characterising compute resources, which is successful for application modelling, achieving high prediction accuracy on a range of applications.