[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: snmpconf FW: November Milestones



Comments inline

> ----------
> From: 	Randy Bush[SMTP:randy@psg.com]
> Reply To: 	snmpconf@snmp.com
> Sent: 	Sunday, December 03, 2000 7:30 AM
> To: 	Dan Romascanu
> Cc: 	'snmpconf@snmp.com'
> Subject: 	Re: snmpconf FW: November Milestones
> 
> > The SNMPCONF WG might be interesting in this message, from one of our
> Area
> > Directors.
> 
> actually, i was saying it as an operator, and in response to a question of
> whether a mib should be designed for config as well as read.  my question
> of
> whether any large operators do puts was really meant at face value.  i
> really want to know if any large providers enable and use write.
> 
I think MIBs should be designed for both monitoring AND configuration.
If some vendors want to implemement just monitoring, then we can
allow for that in a compliant way by specifying two sets of MODULE
COMPLIANCE statements, one for read only and one for the full
support including config.

I also understand that up till now, a lot of operators have not used SNMP
(v1 or v2c) because of security concerns. And in fact, some MIB designers
have taken the read-only path as well, again because there was
no security in SNMPv1/v2c.

Therefore, I assume/hope that SNMPv3 will be exploited so that we can
start to take a serious look to using SNMP for configuration
management.

So when we ask real operators as to if they do use SNMP to configure or
not, then I think we should at the same time ask WHY they do or do not
do that.

What is the alternative? A bunch of inconsistent and non-interoperable
CLI interfaces????

Bert

> randy
>