You are here: Home > Participate > Join a Discussion > Mailman Archives
<<< Chronological >>> Author Index    Subject Index <<< Threads >>>

Re: IAB's proposals in response to the work of the ROAD group

  • To: Daniel Karrenberg < >
  • From:
  • Date: Thu, 9 Jul 1992 14:14:36 +0100
  • Cc:

>Since there has been only one (positive) reaction so far, may I conclude
>that RIPE as a whole is supporting the IAB position? I am wondering
>how many people have read it and thought "If I don't oppose it is clear
>that I agree". So please if you are in favour or of RIPE supporting this,
>send me a personal mail message. I'll count them and report back to the list.
>
>Daniel

Daniel,
        I don't think that I can either support or oppose the IAB position.
I have been following the Big-Internet discussion for a few weeks now, and
I still don't believe that I sufficiently understand all the issues to give
a considered opinion. This applies both to "purely technical" issues, and
"mainly operational/political" issues, and all issues in between. (I'm not
even sure this is a one-dimensional space :-)). 

There does not even seem to be any consensus on Big-Internet on how
urgently "a solution" is required. Although the need for a solution to the
class-B shortage is widely recognized, there is a wide diversity of opinion
about how much time interim solutions like CIDR would buy, and thus exactly
how urgent the, a, next stage is.

The IAB proposals produced a deluge of opinion and dissent on Big-Internet
and Ietf  over the last weekend. Discounting the bigotry :-), some hard
issues emerged. For example, of the two urgent matters, address exhaustion
and routing table explosion, there is nothing in the IAB proposal about
routing table explosion (i.e. routing aggregation, etc), and Jon Crowcroft
has pointed out that if certain accepted NSAP allocation schemes are
assumed, then only 5 bytes of CLNS addressing are left. Some believe that
adopting CLNP to get 5 addressing bytes rather than 4, and no solution to
routing table explosion doesn't look like a good way forwards.

This has lead onto a discussion of exactly what the IAB proposal actually
means. One extreme is "CLNP standard exactly, together with accepted NSAP
allocation", another extreme is "well take IP as a starting point and CLNP
is a step forward. Where do we go from there", and, needless to say, all
points in between. There have been various messages trying to clear up this
confusion, and discussing the implication of various points in this
spectrum of interpretation.

The above is not deliberately my opinion, but an attempt to summarise my
understanding of a very active list. The discussion archives of a little
while ago contained 884474 bytes, and that is just for this month. Most of
it must be the message storm over the IAB proposal. I must have
missinterpreted something if not many things. Please refer back to
Big-Internet for the actual discussion.


I believe that it is extremely important and urgent that a short-term fix
be put into place - presumably CIDR. I believe that it is also extremely
important and urgent to decide on the next steps. It is urgent enough that
we will probably have to choose a pragmatic "interim" next step that might
fall short of a "theoretical best" (though some believe that the next
change to the Internet will have to be the last).

However, I believe that it is *not* yet clear that "adopting CLNP" -
whatever that actually means, is the best choice. It might actually be
completely irrelevant. I would hope that the undoubted discussion at next
week's IETF will generate more light than heat. (After all, our Internet
community prides itself in that doesn't it? That's why we are more
successful than others, isn't it?). Only coming OUT of the IETF can we hope
to make a decision on this. While it would be nice to be able to take a
recommendation in, I don't believe we should.

        "If you can keep your head when all about are losing their's..."

                                Kipling.

        Sorry for the long message.

                Denis.

PS Wish I was going to the IETF - sounds like it will be very interesting.
:-(
 
Denis Russell                   email:  Denis.Russell@localhost
Computing Service               Tel:    (+44) 91 222 8243
The University                  Fax:    (+44) 91 222 8232
Claremont Road                  Telex:  53 65 4  UNINEW G
Newcastle upon Tyne  NE1 7RU
ENGLAND




  • Post To The List:
<<< Chronological >>> Author    Subject <<< Threads >>>