InCommon metadata support

[Tom Scavo <>]

On Tue, Mar 24, 2015 at 4:14 PM, Cantor, Scott 
<>
 wrote:
> On 3/24/15, 11:24 AM, "Tom Scavo" 
> <>
>  wrote:
>
>>Will the implementation
>>take advantage of this, and if so, how?
>
> There's nothing much to take advantage of. DNS entries often stay the same 
> for years, but you still don't cache them that long. You have to ask every 
> so often.

Sure, I understand that the client should check back often (in case
something *does* change) but if nothing's changed, why bother
refreshing what's in the cache?

>>I understand the per-entity metadata server will refresh it sources
>>regularly and often, but why should that force a refresh at the
>>client? Can and will we use HTTP Conditional GET to eliminate (what
>>seems like) unnecessary metadata refreshes?
>
> That's really up to the client.

Actually, I'm asking about the server. (I'll get back to Ian's tome in
a minute :) I'm wondering how the server goes about refreshing *its*
sources. A smart server will only update an etag when the metadata has
actually changed. (I realize it's not as simple as that.)

Tom