---------- Forwarded message ----------
Date: Fri Jun  6 16:59:18 2003 BST
From: AHM2003 - CyberChair <anna@nesc.ac.uk>
To: d.cameron@physics.gla.ac.uk
Cc: r.perrott@qub.ac.uk, anna@nesc.ac.uk
Subject: AHM2003 Abstract Submission Form Feedback

Dear David,

You have submitted the following data for the AHM2003 Conference:

-=-=-=-=-= Beginning of submission data (step 1)=-=-=-=-=-


Contact Person:
	Name: David Cameron
	Address:
		Department of Physics and Astronomy
		University of Glasgow
		Glasgow
		G12 8QQ
		UK
	Phone: 0141 339 8855 x0093
	Fax: 0141 330 5881
	Email: d.cameron@physics.gla.ac.uk

Title: UK Grid Simulation Benchmarks with OptorSim

Author 1:
	Name: David Cameron
	Affiliation: University of Glasgow

Author2:
	Name: Ruben Carvajal-Schiaffino
	Affiliation: ITC-irst

Author3:
	Name: Paul Millar
	Affiliation: University of Glasgow

Author4:
	Name: Caitriana Nicholson
	Affiliation: University of Glasgow

Author5:
	Name: Kurt Stockinger
	Affiliation: CERN

Author6:
	Name: Floriano Zini
	Affiliation: ITC-irst

Author is PC Member: No

Main Fields:
	1. Middleware
	6. Networks


Other Main Fields:
Grid simulation

Abstract:
keywords: grid simulation, data replication

As the computational and data handling requirements of large
scientific collaborations grow, Grid computing is rapidly emerging as
the solution to these requirements. Optimisation of Grid resources is
crucial, however, and replication of data across the Grid is a key
step in this process. To evaluate potential optimisation
strategies, it is important to simulate them as realistically as
possible before they are deployed on the Grid.

To this end, we have developed the Grid simulator OptorSim and used it
to test several performance metrics which are of interest,
identifying them as benchmarks for future simulations. These metrics
include mean job time and usage of computing, storage, and
network resources. Basing our simulations on the UK Grid for Particle
Physics (GridPP), the effects of different
scheduling algorithms, replica optimisation strategies including an economic
model of file replication, and access patterns
were considered. With
respect to previous work, a new version of the economic model was
tested and realism improved by the addition of background
(non-Grid) traffic. The use of these different metrics highlights the
strengths and weaknesses of each strategy and allows us to choose the
most effective for overall Grid optimisation.


Remarks:
[None Given]

-=-=-=-=-=-= End of abstract submission data =-=-=-=-=-=-


IMPORTANT: To submit your paper you should revisit the
submission site at http://cyberchair.nesc.ac.uk/AHM2003/, go to
step 2 and at the appropriate places fill in the following:

Login: d.cameron@physics.gla.ac.uk
Password: XXX

You will then be able to upload your paper.


If you have questions about this message, please contact Anna Kenway
(anna@nesc.ac.uk).

--- End of this message ---