[address-policy-wg] 2006-05 New Policy Proposal (PI Assignment Size)

On Fri, 2006-09-15 at 21:01 +0200, Gert Doering wrote:
> What I am really worried about is people getting "lots and lots" of PI,
> and using multiple routing table slots, instead of getting a reasonable
> chunk of addresses (however named), and announcing only *one* route. 

This is a critical concern.  As it stands, 2006-05 will do little but
encourage rampant use of PI space by removing the psychological barrier
of lying to RIPE.  This is a bad thing: routing tables have once again
grown to the extent that they have burst commonly used equipment - the
current problem is with dual-stacked SUP720 pfc3a/b modules which now do
not have enough tcam to deal with > 192K ipv4 prefixes in their default
configuration (although this can be manually tweaked at the
inconvenience of a reboot).

But we all know this already.

I've previously expressed opinions on address-policy-wg about the
necessity of charging for PI space, both as a means for discouraging its
assignment and for providing a legitimate means to reclaim dead IP
space.

I have to be honest here and say that this issue needs to be addressed
rather more urgently than the proposals in 2006-05.  We can continue to
lie to RIPE on PI application forms, and they will continue to pretend
to believe us so long as the figures add up correctly.  But in the
interim, PI address space is being lost from the global v4 address pool
like a bad memory leak which everyone knows about but no-one wants to
fix.

Nick