OT: diagnosing with ping (was Re: Dejawin Lost Connection)

Bill Vermillion fp at wjv.com
Fri Jul 23 07:39:50 PDT 2004


if:
then: nm(30,*)="Jay R. Ashworth"; dt(20,*)="Fri, Jul 23 09:59 "; 
if:
then: show ("1","1") "On" < dt < nm < "said:" 

On Fri, Jul 23 09:59  Jay R. Ashworth said: 

> On Fri, Jul 23, 2004 at 09:19:33AM -0400, Bill Vermillion wrote:
> > Change that line to 'all most all TCP/IP in MS came from the
> > original BSD code' which as I recall was partially done by BBN.

> This is covered in some depth in the excellent book _Where
> Wizards Stay Up Late_, an I don't remember, either, which way
> it went.

I've read that in the past.

> I know that Unix became the default OS for the Internet because
> of the 4BSD TCP implementation, which was done at Berzerkeley,
> but I just can't remember whether the design was done there, or
> whether BBN's IMP's did TCP. I don't think they did: I think
> that the first TCP router was Dan Mills' PDP-11 based Fuzzball.

In some email in the past I recall I got from a person who had
worked at BBN in the early days when this was done.

> I'll grab the book tonight and look; I actually know right where that
> book is.

And if I get really ambitious I'll look up my old BSD code - that
on an impulse I bought from McKuscik.    I just can't read the 
BSD 1 code as it is some bizarre format and compression for
machines that exist in memory and musuems for most people.

Vint Cerf is regarded as one of the co-inventors of TCP/IP.

This is from an article by Bill Joy.
========================================

   Bill Joy
   Sun Microsystems

Shift from Protocols to Agents

   I first encountered the Internet while a student at the University of
   California, Berkeley, in the late seventies. At the time, Ethernet ran
   at 3 Mbps, and most people thought that TCP/IP was fatally flawed and
   would not work well in high-performance networking situations. There
   were published papers purporting to prove its limited performance.

   I was the principal designer and programmer of the Berkeley version of
   Unix (BSD). After our very small group had worked on this project for
   a few years, we received support from DARPA to continue and expand the
   work. Specifically, we gained access to some code from BBN that
   implemented the TCP/IP protocol. Unlike the major technical
   universities (MIT, Carnegie Mellon University, and Stanford), Berkeley
   at that time had not been connected in any major way to the Arpanet.

                               [linered.gif]

   Protocol Implementations
   We got our first Internet connection up and working by getting a BBN C
   machine (which connected us to the Net), some C code from BBN
   implementing TCP/IP, and some 3-Mbps Ethernet cards provided by Dave
   Boggs of Xerox (who personally manufactured them after we paid for the
   components). I then designed the "sockets" abstraction for the ISO
   protocol stack and, by rewriting the BBN implementation of TCP/IP,
   managed to prove that TCP/IP wasn't slow, like the literature said,
   but could run at the full speed of our 3-Mbps Ethernet LAN cards.

   My BSD implementation of the Internet protocols was widely
   distributed, along with Berkeley Unix, in source form and became the
   standard code for implementing TCP/IP on a wide range of systems. At
   the time, many governments and other organizations thought that ISO
   protocols would displace TCP/IP, but these protocols didn't have a
   widely available source code implementation that was freely shared,
   and this meant that achieving compatibility among all the
   implementations was difficult. I think it is clear that having the
   source for a quality implementation was one of the major determining
   factors in the triumph of the TCP/IP stack.
            ____________________________________________________


The CYHIST archives cover a lot of the past computer history.
It's sort of an oral history proejct. That list had slowed
WAY DOWN in the last couple of years - but there were long
articles from the some of those who originated the pre-Arpanet
implemenations. In the past month there has been more activity
because a lot of the old archives are going to be deleted.

You can check those out at  http://maelstrom.stjohns.edu.
They are planning to take the oldest archives off line shortly so
if you want to snag a snapshot of history 'get them before they are
gone' as they say in so many TV ads.

Bill
-- 
Bill Vermillion - bv @ wjv . com


More information about the Filepro-list mailing list