[mdx] requesting all entities

[Nick Roy]

On 1/17/17 1:21 PM, Walter Forbes Hoehn (wassa) wrote:
>> On Jan 14, 2017, at 9:46 AM, Tom Scavo <trscavo at gmail.com> wrote:
>>
>> As long as we're tweaking the spec, let me suggest we take another
>> look at section 3.2.2 (Request All Entities). I'll observe that there
>> is no normative language in that section.
>>
>> First let me give a data point. The InCommon MDQ server will not serve
>> a full aggregate since that would compete with the aggregates
>> currently served from md.incommon.org.
> I’m not yet certain which side of the fence I’m on with respect to what the MDQ spec should require, but this argument that InC can’t or shouldn’t serve a full aggregate via MDQ has never resonated with me. We have the data, what’s the downside to making it available via alternate location?

Not much other than that it's yet another long-term commitment to
maintain an endpoint.  That said, there are tools and techniques for
moving our existing metadata aggregates over to use the same
infrastructure as the per-entity service, over time.  I'm not opposed to
maintaining duplicate aggregates for some period of time, with, say, a
goal of directing long-term production aggregate requests to endpoints
that are served up by the same infrastructure as per-entity.

Nick

>
> -WFH
> _______________________________________________
> mdx mailing list
> mdx at lists.iay.org.uk
> http://lists.iay.org.uk/listinfo.cgi/mdx-iay.org.uk