Numbering Really Big Networks

[Howard C. Berkowitz]

At 5:04 PM -0800 1/22/97, Kent Crispin wrote:
>Howard C. Berkowitz allegedly said:
>>
>[...]
>>
>> An issue that concerns many people on the ARIN list is being "locked in" to
>> a single upstream provider, assuming the general case is provider-based
>> allocation.  One concern here is the disincentive for people to go with
>> small providers, because they may need to renumber as the small provider
>> grows.
>
>This is not really the issue that I am concerned about, though, of course,
>it is an issue.  There are two important places where your telephone
>company analogy breaks down -- first, the Telco's are a regulated
>monopoly, and there is political(public) oversight.  Second, there
>aren't a hundred small direct competitors of Pacific Bell here in the SF Bay
>area, all getting their telephone numbers *and routing service* from
>PacBell.
>
>A small ISP can be in *direct* competition with its upstream
>provider.

No argument here.

>
>> There are two ways to look at this.  One is that renumbering is
>> anticompetitive and must not happen.  The other says there are technical
>> reasons to renumber, and probably good ones.  Let's focus on making
>> renumbering less painful, because it will be a fact of life.  Let's also
>> realize there are significant real world examples where even large end
>> users might face renumbering of their Internet interfaces, but would have
>> to renumber a very small portion of their users.
>
>I appreciate the technical issues, really I do.  But it is a simple
>fact that small ISPs have to worry about renumbering, and big one's
>essentially don't.  Small ISPs have the burden of an  undeniable
>competitive disadvantage.  If they have to renumber all their
>customers do as well -- oh, wait -- the customers wouldn't have to
>renumber, if they changed ISPs.

It is my contention that customers who do not move toward a
flexible-to-renumber environment are ill-informed or fools.  I do respect
the need for preserving legacy investments.  But large enterprises are
tending to firewall and use private address space.
>
>So we have this technical problem.  The solution is to set up a
>structure that will most likely, in the long run, get rid of small
>ISPs.

A personal opinion only -- I think that in fact may very well be what
happens, if we take a narrow view of an ISP that it's a routing provider.
While I think this is getting a bit far afield, I think there are market
niches where the smaller ISPs can dominate.  But routing inherently has
economies of scale.  Where I think small ISPs can grow and prosper is in
providing access services, hosting web servers and other end user services,
firewall services, etc.

Most of these services aren't, or should not be if implemented with good
practices, fantastically dependent on static addresses.  Domain names, yes.
But dialup users can get dynamically assigned addresses.  Firewalls can
provide address translation and need registered address space on the
outside.

For personal perspective, this is my own email account, not a corporate
one.  It's with ClarkNet, a growing ISP in the Washington metropolitan
area.  Now, I deal routinely with the Big Guys, but I have not chosen to
get my IP access from them.  I like an environment where I know the
operations and engineering staff.  Sure, I know routing architects at most
of the major carriers, but I wouldn't call them for support.

Nontechnical end users may have different perspectives, but as we see the
withdrawal from that market of PSI and Netcom, and the problems AOL is
having, it may very well be that is a market unattractive to large
companies.
>
>Well, perhaps that's alarmist.  But I really don't think so.  Being
>large doesn't make a company non-aggressive.
>
Again no argument.  But I think there are natural niches for small and
large companies, and creative engineering and marketing can make small
companies attractive when renumbering may occur -- if there are other
benefits to customers, and if the ISP takes a proactive stance in making it
easy to renumber.