Hi Laurence, That was exactly what I wanted to know. We also try to define a minimal set of attributes and I am going to follow the discussion regarding that in GIN. Thanks, Timo Laurence Field wrote:
Hi Timo,
I don't exactly understand what you mean.
However, there was quite a bit of work done related to this in the gin-info group. Two approaches were presented, one was to try and agree on a minmal set off attributes and the other was to try an translate one information schema into another.
The conclusion from this work as that we need to have a common vocabably and agree on the fundamental entities that we are dealing with. You can't map something that doesn't exist in one model into the other model and if we have different semantics for something with the same name it can get very confusing if you try and have a conversation.
The other problem highlighted was that what is important for one use case is irrelevant for another. As infrastructures only need to interoperate to meet the demand of their users, we have to look at their use cases to understand what is important. This work is currently being investigate in the GIN group. One example is the WN profile which gives recommendation for the configuration of environment variables.
I am sure that GIN-CG will define a profile for Glue once it is deployed across the infrastructures.
Laurence
As we are _rudimentary and for a prototype_ trying to do exactly that, to map a basic set of attributes from the real technical information schemata used in MDS4, Unicore, BDII into GLUE 2.0, I wondered if somebody is already doing work on interoperability profiles ?
_______________________________________________ glue-wg mailing list glue-wg@ogf.org http://www.ogf.org/mailman/listinfo/glue-wg
-- Dipl-Inf. Timo Baur Leibniz Rechenzentrum Kommunikationsnetze/Netzplanung Boltzmannstr. 1 D-85748 Garching Telefon +49 89 35831-8729 Fax +49 89 35831-5729 timo.baur@lrz-muenchen.de