[ppml] [address-policy-wg] Those pesky ULAs again

[David Conrad]

> Frankly, I'm baffled by the hard architectural boundary at /64.

"Stateless auto-configuration" (whether you want it or not).

> Perhaps I'm limited in
> vision, but it's difficult to imagine a new layer-2 technology that  
> wouldn't
> be entirely overwhelmed by a /64 subnet that wasn't extremely sparse.

The assumption was the lower 64 bits would be _extremely_ sparse.

> It's almost as if we're back in the early 80s, and trying to find the
> same potholes to step in.  Didn't we learn anything from IPv4?

Yes, but we chose not to apply our lessons.

> Apparently not.  We still have the EIG/RID split problem, with no
> solution apparent on the horizon, and we've simply made a new network
> protocol with more address space and incompatible headers.

Yes. We've placed ourselves into a situation where we need to create  
a new, alternate Internet, and migrate everyone over to that new  
Internet, yet have provided no real incentive to do so.  It is  
somewhat odd -- we're stuck with the same constraints we have with  
IPv4 in terms of routing architecture and implementation and naming,  
yet every application must still change (but not in a good way) and  
there is no backwards compatibility at all.

For a short time, I thought there might be a way of using the  
advantages of an EID/RID split as a way of giving end users something  
they can't easily get with IPv4 (namely, easy multi-homing,  
transparent "renumbering", and trivial mobility), but that effort has  
run into the IETF swamp and appears to have gotten a bit bogged down.  
I have lost hope of anything productive happening in any sort of  
reasonable timeframe.

So, since there isn't a carrot (at least one that people care about),  
we're left with the stick of running out of the IPv4 free pool.  Of  
course, that isn't really a stick since, as people have pointed out,  
the Internet isn't going to stop when the free pool is exhausted,  
rather people will simply adjust.

Ah well, maybe NAT isn't so bad...

Rgds,
-drc