On 8/13/07, Sean Hefty <[EMAIL PROTECTED]> wrote:
> > If all you want is to answer a query with the redirect to the source
> > LID, it's probably quite easy to extend opensm in this way.
>
> It's not quite that simple. The actual data eventually needs to come
> from the master SA, so opensm can't simply redirect every query.
>
> I believe as soon as the master SA becomes involved in the redirection,
> it would need to run some distributed SA protocol. The master SA would
> need to know that the requested data can be found at the redirected
> location.
Also, when there are multiple SAs available, it would "need" some
basis (e.g. policy) on which to base its redirection ("nearest"
"approx equal distributed load" SA).
> - Sean
>
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
