>Why can't it operate below the existing ib_sa interface? >Does any client really care whether the record came from the SA >or from the cache?
It can, and I don't think so. Integrating the caching into the ib_sa is possible, but I didn't want to go this route without agreement first. One benefit this approach has is that the cache could store paths only after they were requested. (This would need to be a new feature, so would require more work.) I'd like to get more feedback before making this sort of change though. - 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
