Mandate for starting a Regional Centre SIG in HEPiX

Background

There is an emerging consensus that an important part of the analysis of LHC
data will be performed in "Regional Computing Centres", closely integrated
with each other and the CERN facility to provide as far as possible a single
computing environment. The total combined cost of these facilities is
expected to be several hundred million Swiss francs, with running costs in
the tens of millions.

CERN  assumes that the centres themselves will be built from large clusters
of PCs integrated with inexpensive disks, using models that have been
developed in HEP over the past decade. There is a growing degree of
agreement that Linux is the operating system of choice for these farms.
However, the scale of these facilities will be very large by comparison with
today's installations, even taking account of "Moore's Law". Novel
techniques will be required to automate the management of these clusters and
the associated data, but at present it is not clear that such tools will be
available from industrial sources. In the meantime HEP institutes will
continue to develop their own solutions. It is clearly of interest to HEP
that the various institutes share such developments, agree wherever possible
on the choice of third party tools, and standardise on the details of the
Linux environment.

There is at present some expectation that "Grid Technology" will develop
over the next few years to provide a practical wide-area distributed
infrastructure for scientific computing. Whether these or other techniques
are used, it will be essential for the collaborations that the computing
facilities employed in LHC data analysis work together to provide uniform
access to data and computing facilities.

Proposal

It is proposed that we start within HEPIX a "special interest group" on
Large Scale Cluster Management to share ideas and experience between the
labs involved in regional centre computing, with a view to minimising the
number of overlapping developments and maximising the degree of
standardisation of the environment. HEPIX is an appropriate forum for this
activity as it encompasses the systems and support experts from most HEP
institutes, covering existing and future experiments. This group would
complement the work on the LHC computing model undertaken by MONARC.

Implementation

Your role would be to initiate such a group, taking whatever measures you
feel necessary to obtain the support of the main labs - including in
particular PDP group at CERN. The current situation with Regional Centres is
as follows:

LHC-
 Definite:
  France  - CCPN in Lyon
  USA -  FNAL (CMS),  BNL (ATLAS)
 Probable:
  UK - RAL
  NL - NIKHEF
  Italy - some collection of INFN centres including CNAF and Padova
  USA - a proposal is being prepared for the NSF to fund a large number of
Tier 2 centres
            at non-DoE labs and universities (coordinator Paul Avery of the
University of Florida)

Babar-
  CCPN Lyon
  INFN (not sure how this is going to work)
  RAL
  NIKHEF