[ppml] 2005-1 alternatives

[Geoff Huston]

At 06:33 AM 20/04/2005, Steve Feldman wrote:
>So there wasn't overwhelming support this morning for 2005-01.
>Why did people vote against it?  Because the proposed criteria
>for allowing PI assignment were bad, because PI space is a bad
>idea in general, or something else?
>
>If it's the criteria, what would work better?
>Maybe possession of PI v4 space?  That _and_ an AS number?


I suspect that the underlying issue is that we've managed to surprise 
ourselves a couple of times when the routing table has made a few leaps 
upward at a rate that looked like it would overwhelm the capability of the 
deployed routing infrastructure, and the reaction has been one of viewing 
impacts on routing in a conservative manner by those who got worried at the 
time. In some ways scaling routing is not exactly a well understood topic, 
and there is some doubting of the wisdom of the optimistic view of "well 
we'll solve that routing explosion problem when it 's clearly obvious that 
its about to go bang!". If we understood routing scaling and the dynamics 
of routing at both a technology and at a business level maybe we'd have a 
more coherent common view of what is good housekeeping of routing and the 
associated topic of what is good housekeeping of address blocks that make 
their way into routing. As it stands what I saw today is best expressed as 
an unclear picture of where the best long term interests lie here for the 
network itself.

But then the duality and implicit tensions of routing scaleability and 
addresses utility goes back a very long way - the Routing and Addressing 
Group of the IETF in the early 1990s was an early incarnation of the same 
set of tensions relating to what makes routing scale vs what makes 
addresses truly useful and convenient to use.


regards,

   Geoff