[GSM-WG] Agendas and issues for the OGF

[Flavia Donno]

Jens,
concerning SRM and Digital Repositories and integration of SRM and 
SNIA/XAM, I have published the following paper that can be a good input 
for discussions:
http://www2.computer.org/portal/web/csdl/doi/10.1109/HICSS.2009.435
I won't be attending OGF, unfortunately.

Flavia

Jensen, J (Jens) wrote:
> Hi all,
>
> (apols if you get 2 copies)
>
> I am currently trying to revise the documentation, as usual in the 
> last minute, before the GSM sessions at the next OGF.
>
> SESSION I
>
> The first session will look at the specification itself.  The 
> document, GFD.129, is currently a draft spec. We still need a document 
> for the changes from the spec itself, where the implementations have 
> deviated from it, and most importantly to document the two test suites.
>
> This is the current status:
> 1. GFD.129 metadocumentation:
>
> 1.1. Documenting the test suites.
>
> 1.2. Collating the implementation issues with the current spec.
> This is so a new developer can pick up the spec. and actually expect
> to build an interoperating SRM.  If the ASGC guys attend, it would be
> interesting to hear more about their experiences.
>
> Examples include documenting implementation's implicit behaviour
> (eg what's the size and duration of implicit reservations and pins),
> as well as things that were unclear in the spec ("what is a space?")
>
> Another example will be different behaviours on srmLs on / (or other
> root path)
>
> 2. Issues
> See https://forge.gridforum.org/sf/go/doc15320?nav=1
>
> These can be split into three groups:
> a) things that can be fixed "within" the spec
> b) things postponed for version 2.3,
> c) things for 3.0.
>
> As an example of a), look at performance problems with SOAP over GSI.
> Flavia suggested REST, and I have previously suggested SSL caching
> or just reusing the socket to increase the security layer's performance.
> We need to decide how to make progress on this.
>
> SESSION II
>
> The second session will look at uses of SRM outside circle of usual
> suspects, outside physics.  While the physics community has been good
> at driving the specification and implementations, other communities
> have expressed interest in the protocol, and we should listen to them.
>
> Related to that, there is an overall topic of digital libraries and
> repositories for this OGF.  It would be useful if we could advocate
> SRM as an option for this.  To me, this may mean adding other
> capabilities to spaces.  Also metadata becomes an issue.
>
> Secondly, this session will look at interoperation.  I have a small
> extending previous work on interoperation between SRM and SRB.
> Focus is on grid interoperation now! and specifically using gLite
> as a cornerstone of data management.
>
>
> As has been pointed out before, we never know who attends the OGF
> sessions, techies, or newcomers, or a mixture of both.  Thus it is
> hard to predict exactly what we will be able to achieve at the OGF.
>
> I am aware, however, of numerous discussions going on between the
> SRM developers, client developers, and with other communities.
> I believe we are not good enough at keeping each other informed
> about this - we should also aim to catch up on current activities
> - at least the outcome of them.
>
> Comments and suggestions welcome.
>
> Hope to see you at OGF
> --jens