inventor of the Internet?
Tony Whyman
Philip J. Nesser wrote:
>Just out of curiosity what network level protocol do you plan on running
>over ATM? The wonder and magic of IP is that it can operate over so many
>link layers. Even when we have our own OC48 ATM switch in every home we
>still need a protocol to run over it or all we get is a lot of very fast
>bits that don't do anything. (maybe we can do nuclear particle physics
>experiments when we get OC10000 switches :-)
There is certainly no reason why you cannot run IP over an ATM network and,
in the near future that is certainly what I expect to see happen. However, my
observation is that the Web will demand ATM to the desktop and, when that
happens why do you want the overhead of IP if the underlying network is end
to end? IP provides you with primarily addressing and secondarily,
fragmentation control. It is necessary if you have hetereogeneous networks
interconnected by routers, but is just straight overhead if your network is
homogeneous.
My vision of the future is that from my desktop I will need a few ATM PVCs to
connect me to a Mail Server, a Web Page Cache, etc., and I will set up SVCs
for telephony, video conferences and any private large volume data transfers.
The only argument for IP is about short lived connections where the cost of
setting up an SVC may be questionable and having a network of routers
switching packets might still be justified. But I think that that is missing
the point of the paradigm shift that the Web has introduced away from a
packet switching Internet to a Messaging Internet.
It is messages that will flow through the future Internet. Some will be
transient, like this EMail message - starting from one point and then hopping
around the network until it reaches its destination. Others will be more
persistent, like Web Pages and News, starting from one point and then being
copied out many times to be cached and then deleted by Web Caches, depending
on their popularity. Indeed, it is the Named Message rather than the
addressed packet that will be the general currency of the Internet, and the
Messaging Centre (Web Server, Mail Hub, call it what you will) that will
replace the Router.
--
Tony Whyman McCallum Whyman Associates Ltd Tel +44 1962 735580
FAX +44 1962 735581
Internet: why...@mwassocs.demon.co.uk
Compuserve: 100041,315
Dave Crocker
At 7:23 PM -0700 7/23/96, Tony Whyman wrote:
>There is certainly no reason why you cannot run IP over an ATM network and,
>in the near future that is certainly what I expect to see happen. However, my
>observation is that the Web will demand ATM to the desktop and, when that
>happens why do you want the overhead of IP if the underlying network is end
>to end? IP provides you with primarily addressing and secondarily,
Everything old is new again: "The overhead of IP"
Tony,
In case no one has noticed, most transaction-related exchanges are
just too darned short to be able to tolerate the connection setup time of
ATM.
In case no one has noticed, IP is proving astongishingly good at
providing real-time application use, in spite of having literally NO
facilities designed for it. One can only wonder what will happen when flow
support is explicit.
In case no one has noticed, ATM is proving to be nothing more than
a method of flexibly partioning fixed bandwidth. (flexibly? fixed? What
I mean is that re-allocated is pretty easy, but that it IS an allocation
model rather than being fully dynamic to the level of cell-time, as
originally promised.)
The alarm clock has gone off. The roses are wilting.
d/
--------------------
Dave Crocker +1 408 246 8253
Brandenburg Consulting fax: +1 408 249 6205
675 Spruce Dr. dcro...@brandenburg.com
Sunnyvale CA 94086 USA http://www.brandenburg.com
Internet Mail Consortium http://www.imc.org, in...@imc.org
Grenville Armitage
Tim writes, in part:
>>How 'bout ATM/IP?
Been there, done that, ran IP over it ;-)
gja
Tim Bass
Interesting.
Today ATM, yesterday superframe, yesteryear Greek runners delivering
messages....
The "Internet" just happens to be in the 'radar screen' of mass media
today and to use any term like 'founder of the Internet' just demonstrates
how we humans love to create icons and hyperbola and overestimate our
own importance in a single snapshot of time.
IMO, the inventor of the Internet was Samuel Morse, who demonstrated
that information could be transmitted via electrical energy over copper
wires and secured $30,000.00 in US Congressional Funding to build
the first Washington-Baltimore telegraph project. Of course, a good
argument could be made that any number of inventors developed
the "Internet". ( What would Glick or Mandlebrot argue, BTW ?)
Alas! I will spare the audience a paragraph of metaphors and syllogisms
that are polemic to the ATM, WWW, "in the media's radar screen" today
sensationalism and avoid a rancid tome of discussion over 'which of
the thousands of engineering innovations over modern history
contributed to the current state of Internetworking.
Maybe some Zen-like expression:
"A grain of salt is not the father of the ocean, nor is the wind, nor
the rain..."
(and I hoped the world would 'be a better place in my recent silence' :-)
.... wrong again, as usual.
Hopefully, we have better things to discuss on Big-Internet that
chicken-or-the-egg circular logic or a barrage of port 25 'male bonding'
.... (then again, maybe not ;-)
Let the cyber-games begin ! How 'bout ATM/IP? Will humans again believe
that today is the end of the line in technical knowhow ? Just like
the British believed that the electric telegraph would never replace
a system of towers and signal flags ( the first 'optical network' !!)
Continued Rambling ....... The real inventor of the Internet is
a Frenchman who demonstrated encoding techniques using a series
of optical towers and flags in the year 1791 using cross-arms
and pulleys. Remembering that in the year 1800 "Optical
Telegraph Networks" operated at about 20 chars/second. Then in
1845 (Oct) Morse Magnetic Telegraph Company profits were $413.00
and Postmaster General, Cave Johnson declares in annual report
" ... telegraph business will never be profitable ..."
----------------
Cave Johnson had an opinion, just like all the other we read
on the network today :-)
Regards,
Tim
--
We're just two lost souls
swimming in a fish bowl,
year after year,
Running over the same old ground.
What have we found?
The same old fears.
Wish you were here.
-Roger Waters
Jim Jackson
Why do I sigh when I here this ubiquitous ATM stuff being spouted?
4 years ago when we started out on ATM here, all sorts of people
assured us that they too saw the future as this and that they
envisaged Workstations with ATM as standard within 5 years. Watch out for
the flood of workstations with 100BaseT as standard - no sign of ATM as
standard anywhere :-)
[As an aside look at the relative endstation costs, both of the hardware
and the software support !!!!!!]
People were writing off 100M ethernet 4 years ago as a niche technology.
10BaseX ethernet switching was expensive - it is now approaching
commodity status with the attendant huge drops in costs.
1000BaseT is being persued - and other technologies are being worked on.
One Technology - No chance, stop dreaming. We will ALWAYS have a mixed
network - hence ALWAYS need the likes of IP etc.
Jim Jackson
Project Leader
Leeds ATM LAN Pilot
Robert Moskowitz
At 09:23 AM 7/24/96 -0700, Bob Hinden wrote:
>Sorry to disrupt this very serious discussion, but from reading the popular
>press, I would have thought that
>
> Bill Gates
>
>was the inventor of the Internet! At least the ISOC magazine will not give
>him the title of the "sexiest man....." :-)
>
Nah, Bill is the inventor of the Intranet. To get the proper perspective on
the distinction, re-read Stef's old Paradigm column from Connexions back in '87.
OH, and Bill doesn't have to be sexy. He is too rich for that.
Robert Moskowitz
Chrysler Corporation
(810) 758-8212