RE: Dot3OamEntry in EFM-COMMON-MIB

Hi Jayaprakash - 

Some responses in-line

> (1) Its a good idea that dot3Oam events need to be sent as SNMP traps.
> A SNMP trap can send all the attributes specified in the standard
> 
> But, I feel, running total fields would still need to be defined 
> in some table. What do you say?

Let me make sure I understand.  The defined error events provide thresholds for 
a) frame errors
b) symbol errors
c) errored seconds
We already have C30 counters and MIB variables for frame errors, so we do have a "running total" for frame errors.  But you're right in that I don't think we have symbol errors or errored seconds reflected in C30 or a MIB.  

My guess is that it may be possible to get a counter for errored seconds in the MIB.  A symbol error has been discussed in the past, and my recollection is that the majority of people were very against it as it has such a high update rate that it would be hard to maintain accurately across the management interfaces provided.  So I don't know that getting a symbol error counter into C30 would float, and if its not there, we're not going to get a counter in the MIB.  


> (2) A seperate issue regarding dot3OamEventConfigTable:
> 
> How do we set the peer values for entities defined in this table,
>  as I understand peer entities are not a part of the ifTable, will 
> there be a dot3OamPeerEventConfigTable? In that case SNMP traps 
> for dot3OamPeer events also needs to be defined.
>
> your thoughts?
>

Dot3 OAM had the self-imposed mandate for no "set" capability.  If you want to "set" parameters on a remote device, it could be done either locally at that device (e.g. CLI), or via telnet/SNMP/whatever.  But .3ah OAM can't help configure a peer entity.  

We've tried to pick defaults so that configuration, in most cases, isn't required.  

- Matt


> regards
> Jayaprakash