GROUPAbout us
|
Grid Applications Performance Prediction ToolStart date: 01/10/2008 Duration: 12 months This is an EPSRC funded project that follows two previous successful ones namely DAME (Distributed Aircraft Maintenance Environment) and BROADEN (Business Resource Optimisation for Aftermarket & Design on Engineering Network). The e-Science DAME project saw the development of a Grid resource broker that secured resources on demand for timely critical applications. The DTI funded BROADEN project extended the broker and produced a prototype performance prediction tool which has been proved for use on a particular Grid (owned by Rolls-Royce) and for aircraft engine fault diagnosis applications. The prototype not only predicts the application overall performance (e.g. runtime, data output) but quantifies the accuracy of different prediction algorithms and the consistency of observed data from an operational Grid application. It is aimed at end-users/providers of large scale computational applications, in which processing is done on a Grid. Such large computations can take several hours or days to complete, with the time critically dependent on which assets are used and the parameters associated with the computation. This project will focus on undertaking further scientific and technical implementation of Performance Prediction. The technical development will mean that it can be used in a Grid environment to optimise the usage of Grid resources over multiple administrative domains and ensure QoS provision for a variety of applications. This is essential in building effective Grid systems in a commercial environment. The main commercial partners for this project are Rolls-Royce and EDS (Electronic Data Systems). Software Demonstration Please click one of the below links:
|
(C) UNIVERSITY OF LEEDS 2011 |
|
