Re: SIP Addressing Limitations

[Frank Kastenholz <[e-mail]>]

Bill Simpson writes:

 > Why are we wasting our time with theoretical examples of providers that
 > aren't connected to the internet?

In response to a note providing a hypothetical network configuration
and asking how SIP addressing deals with it.


The answer to your question, Bill, is extraordinarily simple.
Because next decade, somone may build a network that has a
configuration just like the one offered (or very similar) and IPng
must be able to deal with it. The whole purpose of IPng is to allow
the Internet to grow and evolve in different directions than it is
currently doing.

The problem is that we do not know in fact what will happen to the
network over the next 10-20 years, so we have to come up with guesses
and ideas of things that _may_ happen. Some of these may seem
far-fetched by today's standards, but then I have a paper written
about 10 years ago, by Jon Postel (and others) showing that we only
need a 1-byte network number.  At the time that paper was written,
computers were still big expensive machines with lots of terminals --
and the original designers of IP4 assumed that the economics of
computing would stay as they were in the 70's and early 80s.  There
would be a few very large networks, each providing service to large
numbers of expensive machines.  The original designers did not count
on the fundamental change in computing economics to be brought about
by PCs and workstations and LANs.

And these were all technologies that were well known at the time.

What new technologies will be developed and commercialized over the
next 10 years? How will they affect (or be affected by) the Internet?



--
Frank Kastenholz
FTP Software
2 High Street
North Andover, Mass. USA 01845
[phone]