Reed's views, was [e2e] Cannara's views

J. Noel Chiappa jnc at
Sun Apr 15 13:10:27 PDT 2001

<Some history corrections...>

    > From: Cannara <cannara at>

    > It's interesting that a networking system that, for instance,
    > couldn't anticipate needing more than 32 bits for addressing all
    > those machines the future held, when Xerox already had implemented
    > 80 in XNS

Ah, this is not correct. The 32-bit IPv[34] address was already fixed by
'77, and XNS was not done for several years after that.

(I don't know the exact date for XNS, but I was recently doing a search
for old networking stuff, and every early XNS spec I did find was marked
Demember 1981, including one on something like a character set that was
marked "Interim", suggesting it had to have been early on.)

    > From: "David P. Reed" <dpreed at>

    > You're right that we knew about Ethernet's 48-bit addressing.

Actually, no we didn't!

10M Ethernet came along somewhat later, and I distinctly recall trying to
figure out how on earth to fit 48 bits into a 24-bit (the "rest" field
back in the pre-A/B/C days) sack! The result was ARP, which Dave Plummer
and I put together in '82.


More information about the end2end-interest mailing list