RFC 7217

[wollman at bimajority.org]

In article <2354C099-C801-43D2-8EE9-0DA0AAAD8181 at bangj.com>,
pusateri at bangj.com write:

>I read the draft and wasn't sure what problem they were trying to solve.

The problem of having zillions of IPv6 addresses in use on a network
with far fewer than a zillion actual end stations, which no network in
existence is engineered to support.  And also the requirement of
network operators to be able to trace problem IP addresses back to end
stations, which doesn't work if the end stations are constantly
generating new IP addresses and abandoning the old ones.  This has
caused many network operators to disable SLAAC and use DHCPv6 instead,
and I'm about to have to go down this path as well.

>Why not just use temporary addresses (RFC 4941)?

Because they're evil.

>On the server side, this draft could be implemented in rtadvd. Not sure
>who maintains that.

There is no client/server here.  The network infrastructure doesn't
*care* how the client generates its IID.  What it does care is that
the client generate one, unique, long-term-stable, IID, rather than
using nine different ones that change all the time.  rtadvd has
nothing to do with this.

(Note: FreeBSD doesn't currently ship with RFC 4941 enabled.  Most
other operating systems do.  RFC 7217 needs to be implemented, as a
replacement for RFC 4941, before it is even considered to turn this
switch on in FreeBSD..  RFC 7217 *actually solves the problem* that
RFC 4941 was intended to solve, without requiring network operators to
size their switch hardware for ten times the number of addresses as
they have end stations.)

-GAWollman