Re: [rfc2462bis] whether we need the M/O flags

[Iljitsch van Beijnum <iljitsch at muada.com>]

On 15-apr-04, at 17:13, Tim Hartrick wrote:

> I don't know of any implementations which depend on these bits for
> DHCPv6 invocation or termination.  That doesn't mean that none exist.
Also, the whole "other config" issue is still very much in a state of 
flux. Since those mechanisms haven't been defined yet, we can't be sure 
whether the M and especially O bits would be benificial to those 
mechanisms. So removing them *now* would be premature if nothing else.

(For instance, if we settle on well-known addresses for DNS + DHCPv6, 
then having the O bit will presumably be a good thing, as an unset O 
bit indicates that a host can just go to the WKA DNS immediately. 
Without the O bit it would either have query a DHCP server and time out 
if there are none available, or use the WKA DNS while querying the DHCP 
server and possibly the DNS query will time out if the well known 
addresses are unavailable.)

> There are multiple router implementations which allow these bits to be
> advertised.  What is the upside in making these implementations 
> obsolete
> at this late date?
I assert there is no upside. Just look at the whole ip6.int / ip6.arpa 
thing. Changing stuff that's out there in the real world just to make 
the specs look prettier is a very bad idea.

> Are we so short of other work that we need to waste time on breaking 
> things
> which are already working?  I don't get it.
Agree 100%.


--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------