Hi Oxana et al.,
On Mon, 17 Jan 2005, Oxana Smirnova wrote:
> I think ATLAS in future will try to avoid non-Tier1 sites for data
> writing. This might also imply that processing will be mostly
> concentrated at Tier1s or similar sites, to optimize w.r.t. transfer.
It would be unfortunate to waste the resources available for processing at
Tier-2 sites. We were given the impression that T2 sites would be used for
processing, with the results being stored temporarily on the local SE and
then moved on a longer timescale (days or weeks rather than years) to a T1
centre with more long-term storage capacity.
I don't think it was the intention that every T2 site would have to
provide infinite amounts of permanent storage, and I don't think many T2
sites will be able to do this. If this means that our processing power
will remain unused, then there seems little point in having a second tier
at all.
> Basically, we're forced to develop our own workload management and
> especially data management systems.
This is the real problem, of course, and these systems should of course be
provided as part of the generic middleware rather than being reinvented by
each VO. When that will be the case I don't know.
However, I think it would be worth spending some effort on managing data
so that smaller SEs at T2 sites (small as in ~1 TB, not ~10 GB) don't
fill up with data that is in any case replicated elsewhere. If you can
make full use of the Grid, including T2s, in the short term then it is
more likely that there will actually be an extensive Grid to use in the
long term.
Cheers,
Ben
--
Dr Ben Waugh Tel. +44 (0)20 7679 3783
Dept of Physics and Astronomy Internal: 33783
University College London
London WC1E 6BT
|