[arin-ppml] IPv4 Fragment Managemnt policy proposal

michael.dillon at bt.com michael.dillon at bt.com
Thu Apr 29 12:01:32 EDT 2010

> While that is true, it does not need to remain true.  ARIN 
> allocates IP resources according to policy.  If policy states 
> that ARIN will allocate bit-aligned prefixes, then, ARIN will 
> begin allocating in that manner.

That would be dumb

1. I ask for some addresses.
2. ARIN says, you can have a /18. ARIN also reserves the entire /16 in
the expectation that I will come back.
3. Next time, I ask for and justify a /16
4. ARIN give me a /16 and the reserved addresses go into limbo, hoping
that my network growth slows down sometime before armageddon so that can
allocate it in /18 blocks.

Even though it is odd to get non-CIDR blocks, I think that ARIN did the
right thing and that policy ahould not specify bit-aligned prefixes.

> > It might be a good idea to think this through and see 
> whether or not 
> > it can be simplified to say that ARIN will never allocate 
> more than 4 
> > blocks to satisfy a request, and leave off the bit about 
> when there is 
> > insufficient free space.
> Having thought that through, I believe it would complicate 
> the policy and reduce clarity. Brevity is not always simplicity.

I don't want YOU to think it through, I want US to think it through.

I think that it should be possible to come up with a change to current
policy that would not negatively impact current activities and that
would also prevent breadcrumb sweeping without requiring a specific
trigger event for the policy.

--Michael Dillon

More information about the ARIN-PPML mailing list