[ih] Internet-history Digest, Vol 10, Issue 17

[John Shoch]

We built on the original Ethernet and PUP, realizing fairly quickly that we
needed flat, unique, 48-bit addresses for both the 2nd generation Ethernet
(with DEC and Intel, and then IEEE) and the XNS design.  [There was the day
someone moved an Alto from one early Ethernet to another, and we
accidentally had two machines with identical network-dependent addresses on
the same network, talking to the same gateway....."Houston, we have a
problem."  A small preview of mobile systems.]

What an amazing insight -- but, sadly, not mine.
I have often said this is the best networking paper I wish I had written,
by Dalal and Printis, published 39 years ago:
"48-bit absolute internet and Ethernet host numbers"
http://www.bitsavers.org/pdf/xerox/parc/techReports/OPD-T8101_48-Bit_Absolute_Internet_and_Ethernet_Host_Numbers.pdf
https://dl.acm.org/doi/10.1145/1013879.802680

The prophetic last line of the paper:  "We encourage designers of other
local computer networks and distributed systems to use absolute
host numbers from our 48-bit address space. "
Worked for the Ethernet.........


On Thu, Jul 9, 2020 at 8:04 PM vinton cerf <vgcerf at gmail.com> wrote:

> Hi John - those were good times - I appreciated the hints - even if we
> didn't necessarily incorporate all of them into TCP, I think we may have
> tripped into problems and solutions that PARC encountered and perhaps
> solved in similar ways. Ironically, the 48 bit address space used at PARC
> might have been enough to last through a good part of this first half of
> the 21st C although it seemed excessive to me at the time. It's taking a
> long time to get IPv6 propagated and supported.
>
> v
>
> On Thu, Jul 9, 2020 at 9:48 PM John Shoch via Internet-history <
> internet-history at elists.isoc.org> wrote:
>
>> Vint C. and Dave C.,
>> Thanks for your nice mention of our networking work at PARC.
>>
>> Jack H., et al.,
>> In the early internetworking days there was a reasonable amount of
>> interaction between PARC and Stanford:
>> --At various times, both Bob M. and I sat in on Vint's classes or meetings
>> at Stanford.
>> --I spent the summer of 1976 (the bicentennial) working in Washington DC,
>> and got to visit Vint C. and Bob K. at ARPA, in Va.
>> --PARC had a nominal ARPA contract, which allowed us to have an IMP.  The
>> ARPA connections led to us participating in the TCP meetings -- Vint was
>> very gracious and welcoming.
>> --Some of it is documented in various TCP and Internet Experiment Notes
>> (IENs) -- meetings we attended, use of the PRNET, etc., etc.
>> --It is correct that we had some constraints on what we could say about
>> the
>> specific PUP design and implementation.  But we were able to take part in
>> discussions about broader internetworking issues.
>> --For example, in our contribution of IEN 20 on internetwork
>> fragmentation,
>> we wrote about alternate strategies, and then about picking one: "...our
>> current attitude is based on instinct, some experience....."  We sure
>> hoped
>> everyone would get the hint.....
>>  https://www.rfc-editor.org/ien/ien20.pdf
>> --Ron Crane and Yogen Dalal, who had worked with Vint at Stanford, later
>> came to Xerox (actually SDD, not PARC).
>>
>> Vint, those were good times.
>>
>> John Shoch
>>
>> PS:  It's almost dinner time here, and the discussion of Louis Kao's pot
>> stickers is really bringing back memories.
>> --After Hsi Nan, in Town and Country, Louis (Louie) ran a restaurant in
>> Menlo Park, one on Univ. Ave. in Palo Alto, a Foster's Freeze on El
>> Camino,
>> and later a restaurant in San Mateo.  I followed him to all of these, if
>> only for the pot stickers (and kung pao chicken).
>> --Some of you may recall Louis and Sandra's son, who was often in the
>> restaurant.  I'm told he has opened restaurants in the E. Bay, but I have
>> not yet been able to visit.....pork dumplings are on the menu.
>> http://www.noodletheory.com/
>> --
>> Internet-history mailing list
>> Internet-history at elists.isoc.org
>> https://elists.isoc.org/mailman/listinfo/internet-history
>>
>