Per-Entity Metadata Working Group

[Tom Scavo <>]

On Wed, Sep 14, 2016 at 1:45 PM, Cantor, Scott 
<>
 wrote:
> On 9/14/16, 12:20 PM, 
> "
>  on behalf of Tom Scavo" 
> <
>  on behalf of 
> >
>  wrote:
>
>> 5) What is our failover strategy?
>
> Can you explain what you mean specifically?

Yes, I'll try. First, factor out the CDN. It doesn't have anything to
do with what I'm about to say.

The best way to motivate the question is to review our failover
strategy for aggregate distribution (md.incommon.org). There are two
servers behind md.incommon.org: a primary (here in Ann Arbor) and a
hot spare (in Indiana). We manually fail over to the hot spare when
necessary. Clients never notice when we fail over since that is the
nature of aggregate distribution.

For per-entity metadata, it's a totally different story. Fail over
needs to be automatic, and the various servers need to be
strategically located in geographic regions such that latency is
minimized (or at least acceptable).

Does that help?

Tom