Re: [rbridge] Draft TRILL agenda for Paris posted

Just a response on the MIB comment in this thread,

On Wed, Mar 21, 2012 at 10:15 AM, Thomas Narten wrote:

> IMO, the MIB document needs more work and is not as ready as you
> suggest. See David Bond's message at
> http://mailman.postel.org/pipermail/rbridge/2012-March/004870.html.
>
> It says:
>
>     I've submitted a new revision of the
>     draft-ietf-trill-rbridge-oam-02 which contains numerous changes as
>     seen in the change list below. One of the bigger changes was after
>     receiving input I removed a lot of the error codes that were in
>     the previous version of the draft. The thought here was unless
>     there is a clear need for an error type we shouldn't define it. If
>     a use case comes up in the future the error code can be defined at
>     that point.  Most of those error codes can be better represented
>     in the MIB.
>
> We need to decide whether to add more counters to the MIB, and if so,
> do so.

I believe the operative part here is, "Unless there is a clear need for 
an error type we shouldn't define it."

Counters are particularly easy to get carried away with. Implementers 
will tell you that every additional counter they must keep track of -- 
whether in hardware or software -- is a burden. Hence, every counter 
needs to be carefully justified. This has been the rationale in every 
MIB I have worked on.

For now we are well past WG approval and last call of the MIB document, 
and  in the final stages of publication. However as deployment of the 
protocol increases, there will be a process of discovery that may lead 
to additional management requirements, which would feed into a future 
version of the MIB. Part of this could be additional counters which are 
shown to provide enough bang for the buck.

Regards,
Anil
_______________________________________________
rbridge mailing list
rbridge@postel.org
http://mailman.postel.org/mailman/listinfo/rbridge