Hi Roger,
Please find below a portfolio of Sheffield HEP Grid expertise drawn up by
Paul Hodgson and Matt Robinson. Note that this does *not* include
experiment specific application development expertise (please let me know
if you need this as well). There is a lot of experience with Athena/Gaudi
reconstruction software within the ATLAS and HARP groups in this area.
Cheers,
Dan.
____________________________________________________________________________
___
Daniel R. Tovey, E-mail: [log in to unmask]
Department of Physics and Astronomy, Tel. +44 (0)114 222 3546
University of Sheffield, Fax. +44 (0)114 272 8079
Hounsfield Road,
Sheffield, S3 7RH, UK.
____________________________________________________________________________
___
Sheffield HEP Grid Expertise
1. Software support/expertise
. Familiar with installation and configuring of Globus toolkit (GT) GT 2.4
and GT 3.0.
. We administer a WRG node (titania) with a fully working GT 2.4 setup,
grid-map file, GIIS, myproxy server etc.
. We have a WRG GT 3.0 test setup where we are developing grid portals
using java based tools (a really hot subject).
. We have had a working EDG 1.4 install (now defunct) and are in the
process (next two weeks) of upgrading to EDG 2.0. Have obtained our own
hep group subnet from CICS as this facilitates the EDG install. Hope to
offer full HEPGRID cluster (100 cpus) for GRIDPP use over next few months.
Requires integration of existing cluster into GRIDPP.
. Examining use of Condor for cycle stealing type jobs on managed XP
machines.
. Paul Hodgson is RA for issuing eScience grid certificates at Sheffield.
2. Accounting
. Use Ganglia cluster monitoring tool to watch both HEPGRID and TITANIA
clusters. Have customised ganglia to suit our requirements.
. Are developing custom java web based query tools that show accounting
information for WRG. Software gives TITANIA CPU usage by user, department
and (eventually) institution. This is part of effort to develop a genuinly
useful GRID portal- a utility access point for the WRG (again a hot topic)
. Could easily extent/adapt this kind of portal to NORTHGRID use with
appropriate support/funding.
3. Hardware expertise
A long list:
. Custom build Linux boxes on shelves. Full service from component
assembly to fully configured grid aware cluster. Currently completing 100
cpu hepgrid (dedicated HEP) cluster using this expertise.
. Construction and maintainance of large (1 Tb) raid disk servers, very
useful for cheap grid SE's and so cheap can have fully mirrored spares for
failover.
. Are developing a server side tape library access/control application to
use with 3.5 Tb tape library with java GUI client for portability.
. Expertise with Sun V880's (Titania WRG cluster) and installation of GT
2.4 on Solaris 8 systems.
. Soon to be applying expertise to new 400 cpu SRIF2 WRG cluster (part
HEP).
|