[ih] Internet History - Commercialization (was Re: When did "32" bits for IP register as "not enough"?)

[Jack Haverty]

In the 90s I was at Oracle, and our products were users of TCP, SPX,
OSI, and anything else you had at the time.  Getting something to work
in all of those environments was a real hassle.

Even just within TCP, it wasn't simple...

At one point, I recall that there were over 30 distinct implementations
of TCP for the PC.  After the MS release, it wasn't long before there
was just a single implementation.   I think a similar pattern happened
in other OSes; multiple implementations eventually melded into a single
implementation which came with the OS.

TCP was designed to provide great flexibility to the TCP implementer,
who could (and had to) choose a packetization strategy, buffer schemes,
retransmission timing, and other such parts of a particular implementation.

That led to a range of implementations, with differences beyond just
cost and usability.   One might be highly memory-efficient.  Another
might be computationally "light".  For the same transfer of a particular
dataset across a network, there could be vast differences in the number
and sizes of packets produced by different implementations.   We could
see this as we tested with all the different TCP implementations, even
on the same computer.

When "the" TCP became part of the OS, all of those choices
disappeared.   That simplification was obviously an asset.  But IMHO it
was also a liability.

/Jack


On 2/18/19 11:37 AM, Vint Cerf wrote:
> I was on the FTP board when MS released TCP/IP for DOS. It killed the
> market for FTP.
> V
>
> On Mon, Feb 18, 2019, 14:16 Clem Cole <clemc at ccc.com
> <mailto:clemc at ccc.com> wrote:
>
>
>
>     On Mon, Feb 18, 2019 at 12:27 PM Dave Crocker <dhc at dcrocker.net
>     <mailto:dhc at dcrocker.net>> wrote:
>
>         On 2/18/2019 7:35 AM, Tony Finch wrote:
>         > I was wondering what effect KA9Q had on low-end adoption. I
>         turned up
>         > later, but I remember stories from early (1992 ish) dial-up
>         commercial
>         > Internet users who relied on KA9Q.
>
>
>         I believe KA9Q created a lingua franca for PC use of the
>         Internet,
>         within the technical community.  That counted as a major
>         improvement, IMO.
>
>     IMO: Phil (who was a friend and former lab partner at CMU) did an
>     outstanding great job; although I would say FTP SW folks in
>     Andover may have been more important from a commercial
>     standpoint.   Best I can tell, Phil's implementation was popular
>     in the ham community where he originally released it to use over
>     radio TTY HW.
>
>     The MIT's guys (I believe for Project Athena) and then created FTP
>     actually made a product that was tuned to PC Ethernet HW (and
>     DOS).  I had access to both implementations at the time. For
>     instance, we used the FTP stuff for a project at Mass General
>     Hospital, even though it cost a few hundred dollars and Phil's was
>     'open source'.   But FTP SW's solution was more polished and
>     integrated better into their environment.   Phil's stuff was a
>     'hackers tool kit' and although I personally had it running at
>     home, I can say I was reluctant to use it someplace where I was
>     not there to 'maintain it.'
>
>     One of the differences is that FTP guys did the important thing of
>     creating a socket implementation for windows and thus were able to
>     port a lot of the UNIX code using the 386 'DOS extender' from
>     Pharlap and early 386/C compiler.   For a short time, they seemed
>     to be winning the IP for PC battle until MSFT got the IP religion
>     and included an IP/TCP implementation in Win95.
>
>     Clem
>
>
>
>
>       
>
>          
>
>     ᐧ
>     _______
>     internet-history mailing list
>     internet-history at postel.org <mailto:internet-history at postel.org>
>     http://mailman.postel.org/mailman/listinfo/internet-history
>     Contact list-owner at postel.org <mailto:list-owner at postel.org> for
>     assistance.
>
>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20190218/6429c3e8/attachment.htm>