LHC Computer Grid - Rollout
> [mailto:[log in to unmask]] On Behalf Of Jeff Templon
said:
> I was at the 'hotly debated' meeting and indeed there was lots of
> resistance; I would have liked to have said no but I saw no other
> alternative, both from the HEP side (where is the generic
> software that
> will do what their VO software is going to do??) and from the Tier-1
> side (will anyone send us jobs if we don't install a VO box?)
I'd like to add my 2 cents' worth, as someone who has not so far been
involved in any of the discussions. To me it seems rather strange to see
that "a VO box" is a vital baseline service - as I understand it these
are not services, but containers for services. It seems to me that it
would be good to know which specific services are involved, why they are
vital, why they can't be done with the existing middleware and what the
prospects are for getting the middleware developers to produce a generic
solution.
What I actually see in the BSWG document is this:
"These agents will perform activities on behalf of the experiment and
its applications, such as submitting jobs to a CE, monitoring those
jobs, scheduling file transfers, or scheduling database updates."
At least the first three of those are on the face of it supported by the
existing middleware (I'm not really sure what the last one involves).
Even if there are specific needs which are not supported by the current
services, e.g. for job monitoring, is it really impossible to add them?
I can imagine that there are indeed some things which need a specialised
service right now, but I'm not sure why we have to deal with that by
giving VOs carte blanche to run anything they like.
Another thing, this seems to be an LCG discussion, but what happens in
the EGEE context - does any EGEE VO get to have a VO box, or are the LCG
experiments specially privileged?
Stephen
|