SpeQulos: A Framework for QoS in Unreliable Distributed Computing Infrastructures using Cloud Resources.
Abstract
Hybrid Distributed Computing Infrastructures (DCIs) allow users to combine Grids, Desktop Grids, Clouds, etc. to obtain for their users even larger computing capabilities. In this project, we deploy an experimental study of the SpeQuloS framework which aims at providing QoS to best effort computing infrastructures, such as Desktop Grid, by provisioning on-demand Cloud resources, using Nimbus over Future Grid.
Intellectual Merit
Our research activities aim to enhance the experience of DCI utilization available to scientists, both by increasing Quality of Service delivered by those infrastructure and by enlarging the kind of resources available to them.
Broader Impact
By using Nimbus and Future Grid, we hope to strengthen the collaboration between the Grid5000 and Future Grid communities. We already deploy some experimentation of the SpeQuloS framework in Grid5000. Proceeding to similar experimentation with Nimbus over Future Grid could lead to highlight the similarities and differences between these two environments.
Use of FutureGrid
In this project, we will use the Nimbus feature of FutureGrid to allocate some Cloud VM with SpeQuloS to support best effort DCIs.
Scale Of Use
Our primary goal is to validate the capacity of SpeQuloS to use a Nimbus based Cloud, so few resources (few VMs) are needed.
Further collaboration may lead to a larger experimentation, in that case more resources could be needed.
Futuregrid is a resource provider for