Re: partial comments on section 5 of ibif

At 02:49 PM 10/9/2003 -0700, Randy Presuhn wrote:
>Hi -
>
> > From: "Vandana Rao" <[email protected]>
> > To: "Randy Presuhn" <[email protected]>; <[email protected]>
> > Cc: "'Bert Wijnen'" <[email protected]>
> > Sent: Thursday, October 09, 2003 2:14 PM
> > Subject: Re: [Ipoverib] partial comments on section 5 of ibif
>...
> > I'm not sure how this should be addressed then. I do understand your point,
> > but there is no way
> > that we can mandate this on IB devices without obsoleting existing h/w.
>...
>
>If managed systems do not know when they've processed multicast traffic,
>then those systems should not instantiate the multicast-related counters for
>those interfaces in the interfaces MIB, and, consequently, their vendors
>should not claim they conform to RFC 2863.   This is how this kind of
>situation has been handled since the earliest days of SNMP. The ibif
>document should specify the *correct* behavior for those counters in
>the context of what can happen in the InfiniBand protocol, not any particular
>chip's implementation.  This seems simple enough to me.

The "correct" behavior for those counters in the context of IB would be to 
be set to 0.
Looking at the ibif draft some more, I think that would be the right choice 
as the draft specifies that the
mcast and unicast packets are counted in the unicast counters as is done in IB.
-Vandana