[e2e] Lost Layer?
detlef.bosau at web.de
Mon Jan 13 09:25:27 PST 2014
Am 13.01.2014 16:09, schrieb Ivancic, William D. (GRC-RHN0):
> I strongly suggest reading John Day's book "Patterns in Network
> Architecture: A Return to Fundamentals." It is not and easy read and has
> a lot of history that you may or may not find useful.
It is - with all due respect towards John - the "history" which is often
difficult to read in John's publications.
Admittedly, I'm not interested in history, when a book contains to much
of it, I simply ignore it.
Science is, as DPR said, about asking questions and finding answers. And
a useful approach to asking questions is the
- what is the question?
- why is it important, with particular attention to: If it is important,
why isn't it solved or why are existing solutions not satisfactory?
- how do I want to solve the problem.
I had quite some off list discussions with John - and frankly spoken, I
eventually resiled because I couldn't find a way to communicate with
John in concise questions and selecting adequate solutions. This may be
my fault, I'm perhaps a somewhat "difficult" person and perhaps I have a
very individual way of thinking.
> If you only look at
> one portion, I suggest finding the short discussion of how the phone
> number evolved from a location identifier to an application identifier -
That's another point of interest. John is quite often interested into
numbering and addressing problems. I'm, primarily, interested in
congestion control. There are numbers. That's it. How is it called in
the HHTG? An SEP, IIRC.
(Not to be misunderstood. Wrong numbering can lead to huge problems. But
I'm not a religious person, so I'm no strong believer of
> (I think it is in the mobility section, but I don't have my copy handy).
> . In fact, you might want to jump here first then go read from the
> beginning. It should give you some insight as to why the current Internet
> Architecture has so much difficulty with mobility and multi-homing which
> is what originally drew me to this work.
And afterwards, you are that much into the existing architecture that
you fail to think otherwise.
William, during the past 14 years I was strongly into VJCC, simulations.
the e2e principle, or perhaps the first 13 of the past 14 years.
About 2013 I started to put the whole thing in question.
And the result was: I put the e2e principle in question and the more I
think about it, and I got harsh reactions when I mentioned this
to some people, the more I'm convinced: VJCC and offsprings was a
wonderful work around for the congestion collapse problem and it
deserves huge respect that it works until these days - however, the
problem itself remains still unsolved and we need a crank back to about
1989 and a clean slate approach with the knowledge about networks of
2014 in mind.
> In a nutshell, Naming and Addressing needs to be done right and,
Yes. But this is not the reason for congestion collapses to happen.
> IMHO, we
> still haven't gotten it quite right. Sometimes it looks like we are
> getting close and then we stumble. DTN sort of had a chance and missed.
> Maybe Information Centric Networking work will help move naming and
> addressing along.
Delay tolerant networking may provide wonderful insights.
However, the very first insight which I got into TCP (it was the first
insight because after this insight I intentionally abolished anything I
ever heard about TCP before) is that TCP is an asynchronous protocol.
And where TCP is not asynchronous, it is wrong.
After this sort of "garbage collection" you can start again with the
question: How can we convey a flow of bytes from one location to another -
thereby taking advantage of some existing network underneath.
And I intentionally forgot about all this self clocking and self
scheduling nonsense - even about discussions of stability.
In a sense, I'm better now :-)
(Reminds me of a tooth which caused problems for me for certainly more
than a decade. Eventually the upper half broke away and the lower half
was dug out by an oral surgeon - however: The problems were gone :-))
Perhaps, it's at least in part a matter of age. I turned 50 last year,
be it a case of mid life crisis or reasonable for whatever reasons, on
some occasions it is a good idea to clear out one's brain and start
thinking from scratch.
> William D. Ivancic
> Phone 216-433-3494
> Fax 216-433-8705
> Networking Lab 216-433-2620
> Mobile 440-503-4892
> On 1/10/14 2:34 PM, "Detlef Bosau" <detlef.bosau at web.de> wrote:
>> I would like to discuss the talk
>> given by John Day.
>> What do you think, e.g., of the claim
>>> TCP was split in the Wrong Direction!
>>> € It is one layer, not two.
>>> IP was a bad idea.
>> Detlef Bosau
>> Galileistraße 30
>> 70565 Stuttgart Tel.: +49 711 5208031
>> mobile: +49 172 6819937
>> skype: detlef.bosau
>> ICQ: 566129673
>> detlef.bosau at web.de http://www.detlef-bosau.de
70565 Stuttgart Tel.: +49 711 5208031
mobile: +49 172 6819937
detlef.bosau at web.de http://www.detlef-bosau.de
More information about the end2end-interest