Independent space from ARIN
jeffm at iglou.com
Mon Apr 14 00:58:09 UTC 2003
Also Sprach jlewis at lewis.org
>On Sun, 13 Apr 2003, Stephen Sprunk wrote:
>>Please explain how somebody with more than 4096 hosts in PA space is
>>supposed to renumber into a /20 of PI space.
>>I fear you propose that he move the first 3276.8 hosts, request a
>>second block, move another 3276.8 hosts, request a third block, etc.
>>until he's got a dozen new allocations which can't be aggregated.
>>Perhaps this explains the explosive growth in the routing tables since
>>ARIN took over.
>Perhaps the poster who mentioned they didn't get enough space to
>renumber should have started, filled the allocation, requested another,
>and finished the renumbering.
Ignoring, for the moment, that absolute absurdity of that type of
procedure...you forget what I've now said twice...that ARIN said as
clarification after I got the first block that renumbering wasn't a
consideration, full stop.
Either ARIN's policies are screwed up beyond even what I thought to
begin with, or their communications with customers/ISPs/whatever is
absolutely pitiful. Most likely, both.
>In your request, did you mention any sort of projected timeline for
>renumbering into the block you requested?
During the first request, we proposed a timeline of 6 months to a year
for renumber, if I remember correctly. And please don't even *think* of
suggesting that we should have done it in 3 months...that's just
>I don't know anyone who's actually followed this, but I haven't
>communicated with many ARIN members about this sort of thing lately.
>Is this policy being enforced consistently now? I know in the past,
>ARIN has had their own policies (at least for initial and at one time
>for second allocations) that pretty much ignored this. Once upon a
>time, you could request a /20 from a reserved /19 as long as you were
>multi-homed and could justify a /21. Fill the first /20 in 18 months
>or less, and you get the other half, and have a /19. I think the
>rationale for this at the time was routing filters, as you were allowed
>to announce the /19 even before the second half of it was officially
>yours. Now, the ARIN tune seems to be "we only assign numbers,
>routability is your problem".
FWIW, the second block that we got just a short time ago, was an
extension of the previous /20, to make it a /19...not that this is
relevant, in any way, to any of the issues raised. We still haven't
never received from ARIN, a sufficiently large block to be able to
renumber out of the currently utilized space as was offered for the
first request, and strongly requested for the second; and the
communications that we received after the first request was a flat out
lie about the consideration of renumbering in allocations.
There is no was for ARIN to get out of this one smelling like
roses...they screwed up...probably twice, depending on your opinions
about policies...but at least once in the lie about renumbering
>I don't claim to have an easy solution for this. If every idiot with a
>business plan could request and receive a /16, there'd be an awful lot
>of wasted space.
FWIW, the first request we made was for a /19, which would have been the
smallest single block that could have been allocated to us to allow us
to renumber into; and the second request was for an /18, with the same
reasoning. We got /20's both times (with the second /20 being the
second half of the /19 of the first /20).
>But if you've been around for most of the past decade and have
>continued to grow, should you really be issued new non-agregable blocks
>every several months?
IgLou has been in the Internet providing business for the better part of
>Somebody must have a better idea.
Here's a radical thought. Use some common sense and critical thinking
skills in deciding what the allocation should be. It certainly seems to
be lacking at the moment.
Jeff McAdams Email: jeffm at iglou.com
Head Network Administrator Voice: (502) 966-3848
IgLou Internet Services (800) 436-4456
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 197 bytes
Desc: not available
More information about the NANOG