[e2e] Mystery

Cannara cannara at attglobal.net
Sun Apr 29 16:58:42 PDT 2001

And, this was how every 3Com XNS 3+Share station started up each day, from at
least 1986 on.  :]  


Steve Deering wrote:
> At 11:26 PM -0400 4/28/01, David.Eckhardt at cs.cmu.edu wrote:
> >If memory serves me, one of the myriad forms of IPv6 address is
> >some constant prefixed to a 48-bit MAC address, which can somehow
> >be parleyed into a globally routable address.
> David,
> It's the MAC address that is constant, not the prefix (i.e., high-order
> part of the address).  The high-order part of a node's IPv6 address(es)
> depends on what subnet(s) the node is attached to, and is discovered
> automatically by the nodes.
> By the way, this approach to address autoconfiguration is an old idea,
> which I first saw in Xerox's XNS (which mutated into Novell's IPX),
> and was also defined for ISO's CLNP.
> The implication in the Sony announcement that a device might get an
> IPv6 address "for life" is the latest misconception that we (the IPv6
> design community) are having to counter -- we recently just barely
> stopped an IPv6-related press release from the European Commission
> that was to include language about permanent IP(v6) addresses becoming
> a human right(!).  Maybe someday when we figure out how to do flat
> routing to all devices in the Internet...  (Until then, we'll have
> to limit the right of humans to having only a DNS name for life.
> Or something.)
> Steve

More information about the end2end-interest mailing list