Per-Entity Metadata Working Group

["Cantor, Scott" <>]

> This may come as a surprise but the apache server on md.incommon.org
> is not configured for HTTP Compression. Ops has discussed this
> occasionally over the years but the case for HTTP Compression has
> never been compelling so we let it drop each time. Maybe it's time to
> revisit.

I think we all recognized that conditional GET >>>> compression, for 
aggregates certainly.

> I don't know what metadata clients support HTTP Compression. I assume
> Shibboleth does but I don't know that for a fact.

For aggregates, the SP does. I assume the IdP supports it uniformly, but I 
don't know. The SP's code for MDQ happens to not currently turn on 
compression.

> Assuming Shibboleth handles HTTP Compression just fine, I wouldn't
> expect too many problems if we enabled apache but a few deployments
> will experience problems, I'm sure. The fallback aggregate would come
> in handy here. Does anyone know how to configure apache such that the
> fallback aggregate is NOT compressed on the fly (yet everything else
> is)?

Isn't compression just negotiated? You make the resources available in that 
form and Apache will serve up whatever the client handles. Shouldn't be any 
breakage with it.

-- Scott