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

Christian Huitema huitema at
Wed Apr 25 10:00:01 PDT 2001

In response to John Day's argument:

> In fact, even Ethernet reflects the state of our understanding at the
> time.  Then we didn't understand that addresses only had to have the
> scope of the layer in which they were used.  I really doubt that
> there will ever be an Ethernet segment or even a bridged ethernet
> with 2**48 devices or anything remotely close to it.  16 bits would
> have been more than enough.  What is interesting is that IEEE 802
> still hasn't figured it out.  Look at firewire addresses.

This is simply not true. There were LAN prototypes at the time that used
16 bit addresses, or even 8 bit addresses; people did understand layered
architectures -- at least they certainly did at Xerox Parc. Having
"world wide unique" addresses was a very well articulated trade-off
between transmission efficiency and management cost. It was motivated
primarily by device mobility, the possibility to attach a device to any
random network without requiring any administrative assignment of an
address; another important benefit was the possibility to join or bridge
two networks without having to renumber all stations. 

With hindsight, they might have realized that 48 bits was actually kind
of short, but that is another entirely different argument.

-- Christian Huitema

More information about the end2end-interest mailing list