[GSM-WG] Notes from GSM-WG at OGF27
Jens Jensen
j.jensen.ral at googlemail.com
Mon Oct 12 18:27:07 CDT 2009
Dear GSM-WG members (and data area directors).
Here is a writeup of the GSM-WG meeting at OGF27. Freshly typed,
ink is still wet. Comments welcome.
* It was announced that Arie, after a long stint with many significant
contributions to the group, has decided to retire as chair but will
stay involved in the group. Likewise Alex as secretary. Timur
Perelmutov from FNAL has been appointed co-chair (so we have one
from Europe and one from the US).
* We had an introduction to SRM and a demo of using SRM within gLite,
but with a lot of questions from the audience, so it took 2/3 of the
time (1hr), covering at a high level the protocol, and how it fits into
the wider context of the (gLite) grid.
* In the remaining 30 minutes, about 25 were spent discussing the
roadmap of the group: following up from the decision at Chapel Hill,
we summarised the discussions from Chapel Hill and since then. When
the group is rechartered, the new charter needs to cover what's in
scope and what is not. Numerous examples were given. One of them
is quite recent: implementations are getting improved support for
checksumming, so clients are starting to make use of this. Also,
the dCache decision to make ls asynchronous was discussed. These
are typically changes well within the protocol, because the protocol
was designed to be very flexible, but the implications of the change
still need to be openly discussed. Openness is important to ensure
that all communities have a chance to be heard, and to ensure that
the process is not hijacked by a specific community.
* Another thing which is considered important is the use of SRM,
particularly outside the traditional HEP communities. As an
example, we looked at a Biomed use case.
* Also important is the role of the GSM-WG in relation to other OGF
groups. Storage Elements with an SRM interface depend directly on
GridFTP and the GLUE schema. And there are other less direct
relations to other OGF groups (which we did not have time to explore
further in this session.)
* In the remaining 5 minutes, we looked very briefly at the comments
from the reviewer on GFD.129 and GFD.154. The comments were related
to something discussed during the "introduction" part of the
session, so we could close it fairly quickly. The group will write
a response which will be published on gridforge.
Thanks
--jens
More information about the gsm-wg
mailing list