Re: Persistent or opportunistic IDs

[Brian E Carpenter <brc@zurich.ibm.com>]

Iljitsch van Beijnum wrote:
> Yesterday at the RIPE meeting Geoff talked about the difference between 
> doing address agility using long-lived identifiers vs pragmatic, 
> short-lived identifiers. This is indeed a useful way to classify multi6 
> solutions. Pragmatic identifiers would generally be regular (locator) IP 
> address. However, selecting such a pragmatic approach doesn't 
> necessarily close the door on persistent identifiers, as it's possible 
> to "hide" such a persistent identifier in the lower 64 bits of a PA IPv6 
> address. This is basically what I propose in my crypto-based host 
> identifier draft.
>
> Two questions:
>
> - do we agree that going the pragmatic route first and adding persistent 
> identifiers later is indeed possible?

We need a proof-of-concept proposal before we can agree or disagree.
I'm not sure your draft really provides that. NOID for example is a
proof of concept for the pure pragmatic scenario.

> - if this is the case, does this mean such is also the preferable way 
> forward?

Too soon for that question. We need an agreed architectural analysis...

   Brian