Hi Chris,
On Fri, May 06, 2005 at 07:39:32PM +0100, Brew, CAJ (Chris) wrote:
> I've been persused to give this talk at hepix next week.
>
> You can see the first draft at:
>
> http://hepwww.rl.ac.uk/ppdstaff/brew/public/BaBarGrid/Talks/RAL-Xrootd.p
> pt
>
> Since as far as I'm aware xrootd hasn't been presented at hepix before
> there's a section at the start with a short discription (Andy might
> recognise some of the phrasing) followed by a description of how it's
> deployed here and our experience.
It looks like an excellent talk. When you have a final version, we
can also put it in the xrootd area and link it from the xrootd web page if
you like.
> Is there anything you'd like me to add, remove or clarify? I was
> thinking possibly of a slide at the end of the description with some
> performance/scaleability numbers.
You could probably add something like the single daemon I/O performance
numbers Andy and Bill put together a couple of months ago:
http://xrootd.slac.stanford.edu/presentations/Xrootd_Performance.htm
> Any other comments are welcome.
Slide 3 - xrootd is distributed with ROOT (and just as importantly the
client is there out-of-the-box) since ROOT 4.01-02 which (IIRC)
came out early last fall (Sep2004, IIRC).
Slide 13 - probably worth making clear that the 75TB is at RAL alone
(Did we really manage 75TB with the single soft-link namespace
thing and NFS? I thought that you only did that through about
20TB or so at RAL. Which was already a lot!)
Slide 18 - there is a word "condrestarts" that I guess should be "restarts"
Slide 21 - Perhaps it is worth listing also the other big sites running
xrootd for BaBar and their sizes:
SLAC (160TB+)
In2p3 (???TB)
FZK (???TB)
CNAF (???TB)
Padova (???TB)
plus the fact that number of smaller university sites are
using it. There is of course some testing and/or use by other
experiments (AMS, Indra, CLEO, STAR, CMS, Alice, ...), too.
[I just wrote the others about their space numbers in private
since I'm curious to have these, too.]
Pete
-------------------------------------------------------------------------
Peter Elmer E-mail: [log in to unmask] Phone: +41 (22) 767-4644
Address: CERN Division PPE, Bat. 32 2C-14, CH-1211 Geneva 23, Switzerland
-------------------------------------------------------------------------
