Per-Entity Metadata Working Group

[Jorj Bauer <>]

This all sounds a whole lot like reinventing DNS. 

Sent from my iPhone

On Aug 11, 2016, at 15:57, Cantor, Scott 
<>
 wrote:

>> So far, I’ve only heard option “a” discussed. What this buys you is simply 
>> a
>> cache that spans restarts for the entities already “seen” by the node.
> 
> Right, and our expectation is that entities may still carry multi-day 
> validity windows if they need to, particularly in the early going.
> 
> We would resolve a cached instance to memory and only then be within the 
> cacheDuration the document itself is advising us to use. On disk is 
> basically a backstop to transparently fall into as though it came from the 
> real server.
> 
> -- Scott
>