[ppml] Re: Independent space from ARIN

Sweeting, John John.Sweeting at teleglobe.com
Tue Apr 22 22:40:39 EDT 2003

I can sympathize with everything you said below as I have been in the same
position on a few occasions so I rolled up the sleeves, worked with ARIN (or
RIPE or APNIC) to identify our shortcomings, corrected them and got on with
life and having my requests honored because we could prove what we were
doing with our allocated/assigned address space. I am sorry but all the
problems you listed below are not ARIN's problems, they are the company that
let their records get so corrupted and out of date. I know for a fact that
ARIN goes out of its way to help its customers solve their issues, I have
witnessed it on several occasions. I can also tell you that there is no
slack cut just because you are one of the larger providers.....ARIN will and
does check the /29 assignments just the same.

-----Original Message-----
From: Leo Bicknell
To: ARIN Policy
Sent: 4/22/03 6:32 PM
Subject: Re: [ppml] Re: Independent space from ARIN

In a message written on Tue, Apr 22, 2003 at 05:15:50PM -0400, Sweeting,
John wrote:
> To each their own opinions....I have been dealing with ARIN since its
> inception and can only say that they have been consistent in their
> for information on their templates as well as the evidence that they
> for justification. I have worked for a few of the larger Tier 1
> and requesting address space from ARIN was one of the easier tasks of
> managing IP address space. If you manage your address space and can
> them how it is used then there is no problem in acquiring additional
> addresses when required.

>From my own personal talking to people about ARIN problems, I'll
note the two ends do not seem to be the problem.  That is, if you're
really small, going for the minimum block (ie, your first block)
the steps are clear, simple, and implemented well.  If you're really
big (eg have at least a couple of /16's) things are also fairly
simple.  I'm not quite sure why in the latter case, but I assume
there is some amount of "trust" that builds up, and/or perhaps just
an overwhelming amount of work (who's going to check to the /29
level in 5 /16's before giving out another one?).

Most ISP's don't fall into either of these camps.  They have a
couple /small routes.  These days they likely formed by merging
two or more companies IP space.  Contacts are out of date and wrong,
names are out of date and wrong.  Old usage data is bogus.

You can see how this quickly becomes a problem.  ARIN's old records
(contacts, how things were being used) don't match reality any more.
the companies have changed how they do business, and many blocks
have probably been recarved in new and different ways.  They probably
have some space from an upstream making things more difficult.  The
person who has the job of getting more space probably had nothing
to do with any of the original applications.

I won't say that ARIN is wrong in how they deal with these people,
as there are very real problems on both sides of the table with how
to provide information, and how to evaluate it.  However, I will say
that it often turns out being a much larger effort for ARIN and the
provider than it should be, and often the process seems to prevent
the right things from happening.

For instance, it should be easy to go to ARIN and say I have 20
blocks, they add up to 75% of a /18, plus I have growth.  If you
give me a /18 I'll renumber, and give all 20 back.  It's better
for all of us.  In practice though, doing that simple exercise
seems to be a nightmare.  I don't really understand why it is so
hard, but it always seems to me like death of a thousand paper
cuts.  No individual question is inappropriate, no individual
response bad, but put them all together and you have a multi-year
process for many people just to get some IP space sorted out.

       Leo Bicknell - bicknell at ufp.org - CCIE 3440
        PGP keys at http://www.ufp.org/~bicknell/
Read TMBG List - tmbg-list-request at tmbg.org, www.tmbg.org

More information about the ARIN-PPML mailing list