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

Re: snmpconf BCP document



Bert, comments inline.
> Jeff,
> I have many hadnscribble notes on the document which I 
> will hand to the authors/editors.

Thanks we will do what we can to fix any errors or make things clear.

> 
> What bothers me most is that it seems premature to do BCP
> on policy-based management. Maybe we should try a
> BCP for the existing SNMP practices and a separate
> guidelines  informational doc with sample MIB for policy
> based management. At some later date, when we indeed
> have more experience and some current practices,
> maybe then we can clean that up and make it a BCP

The way that we hook-up mechanism and instance specific MIB modules is
something that is relevant and we have learned some of that from the
DiffServ and SNMPCONF effort. While there are probably still bugs in the
BCP example, the idea of showing how this is done is very helpful.


There are products on the market that do configuration with SNMP and some
even do configuration of policy, just not with exactly the PM MIB
Module. This knowledge is in the document along with other general
information. The feedback from the people that have commented so far is
that the document is helpful. Of course will learn a lot more over time
and then we can revise the BCP. Other areas have done this as well.

> 
> But I must say that it also bothers me that various errors
> and deficiencies in existing MIBs are spelled out and
> I find it hard to see what the recommended best practices
> are to prevent that in the future.

We may be looking at things differently since we have gone to
some length to include examples of what is helpful as well. We have
also attempted to talk about trade offs that engineers should think
about, since that is relevant to such a document. We hope the
consideration of some of these areas will reduce the likihood of errors
in the future, not only for mib modules but for application developers
as well.


> 
> In the meantime I have also gone through the sample MIB
> and have also there lots of comments scribbled in the
> document. And I think we would need to add text that
> describes how you can define some policies and a
> sample set of filters and then show how that then gets
> applied.
> 

This is an excellent idea, we will add it to the to do list. It had been
suggested in the past, we have not yet gotten to it.

/jo
> Bert
> 
> > ----------
> > From: 	Jeff Case[SMTP:case@snmp.com]
> > Reply To: 	snmpconf@snmp.com
> > Sent: 	Sunday, March 18, 2001 9:29 AM
> > To: 	snmpconf@snmp.com
> > Subject: 	RE: snmpconf BCP document
> > 
> > 
> > bert>I did go through this document and it seems to me
> > bert>that it is not really just a BCP, but rather a mix of:
> > bert>
> > bert>- BCP
> > bert>- Requirements
> > bert>- Guidance
> > bert>
> > bert>Specifically BCP stands for best CURRENT practice.
> > bert>So I wonder how the policy-based NM can already be
> > bert>current practice.
> > 
> > hi bert, et. al.,
> > 
> > i agree ... you are right that the document is a mix
> > 
> > which is bothering you more,
> > 	that it is a mixture, as many bcp documents are, or
> > 	your second point regarding the dimension of current versus new?
> > 
> > is there a particular action or solution that you have in mind?
> > 
> > see you soon
> > 
> > best regards,
> > jdc
> > 
> 

Thanks,
/jon
--

Jon Saperia		     saperia@jdscons.com
			     Phone: 617-744-1079
			     Fax:   617-249-0874
			     http://www.jdscons.com/