Interactivity on the Grid: Limitations and Opportunities

Grid computing has the intrinsic disadvantages of a batch system: the undetermined delay between the time a job is submitted and the time it is completed. Our approach to this problem is iGrid, a framework that proposes to change the Grid from a batch system to a more interactive distributed platform. Agents are deployed on nodes and wait to be contacted by an interactive job to start the computation, thus bypassing the regular Grid scheduler. A CPU fairshare mechanism allows users to get their fair iGrid machine share over a long period. Our prototype implementation scales to all 400 Grid nodes we were granted executing on top a software for data intensive parallel computing.


Published in:
Hpdc 11: Proceedings Of The 20Th International Symposium On High Performance Distributed Computing, 273-274
Presented at:
20th International Symposium on High Performance Distributed Computing, San Jose, CA, Jun 08-11, 2011
Year:
2011
Publisher:
Acm Order Department, P O Box 64145, Baltimore, Md 21264 Usa
ISBN:
978-1-4503-0552-5
Keywords:
Laboratories:




 Record created 2011-12-16, last modified 2018-09-13


Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)