Cloud@Home: performance management components
Euro-Par 2010 Proceedings of the 2010 conference on Parallel processing
Security and performance trade-off in PerfCloud
Euro-Par 2010 Proceedings of the 2010 conference on Parallel processing
Empirical evaluation of cloud-based testing techniques: a systematic review
ACM SIGSOFT Software Engineering Notes
Enhancing an autonomic cloud architecture with mobile agents
Euro-Par'11 Proceedings of the 2011 international conference on Parallel Processing
The CloudGrid approach: Security analysis and performance evaluation
Future Generation Computer Systems
| Hi-index | 0.00 |
Cloud Computing, born in the e-business context, and GRID computing, originated in the e-science context, are two different but similar paradigms for managing large sets of distributed computing resources. In the last few years there have been many efforts that aim at integrating them. One of the main problems linked to the use of the virtualization techniques adopted in clouds in the GRID and High Performance Computing context derives from the overheads in the virtualization layer. This paper introduces "PerfCloud", a cloud environment built on the top of a GRID system, which lets the user to instantiate Virtual Clusters (VCs) that become part of the starting GRID. The architecture proposed integrates a set of services able to predict the performance (response time) of user applications on the newly generated VC taking into account the actual amount of computing and communication resources allocated to the VC, as well as the presence of the virtualization layer. The obtained predictions let the user to evaluate on-the-fly if the created VC is compatible with his performance expectations or not.