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

snmpconf Re: I-D in configsnmp



on 09/19/2000 8:58 PM, Ehab Al-Shaer at ehab@cs.depaul.edu wrote:

> 1. in P.9, "if polling would not scale, consider using notification for
> reporting the old and new checksum objects". I am wondering based on what this
> notifications are generated? is it based on time or event-triggered? I think
> you meant event-triggered based on changes which means a notification will be
> sent everytime a change is performed on an object.
> 
> Thank you
> 
> -Ehab
> 
> 
Ehab, I have taken the liberty of forwarding this to the SNMPCONF list since
it relates to the BCP.

Ehab, this particular sentence perhaps needs to be reworded for clarity.
Event based in this context means that when the configuration is changed, a
notification will be sent. So we are in agreement if the object is pointed
to a single MIB Object. In the case of configuration, that would not be very
good. There are several techniques for dealing with this. One is a
notification any time the configuration in a branch of the tree changes such
as routing parameters for a peer group. The notification can contain the
entire row that changed in many case.

As with all notifications (and MIB Objects) they should be expressed at the
right level of abstraction, a point we hope to successfully make in the BCP.
What we mean by this is if you created a notification every time the output
queue depth changed on an interface, there would be quite a problem with too
many notifications.

Bottom line, we are saying the same thing. Mike and I need to fix the words
though.
/jon