Re: [NGO] comments on CANMOD BoF

Seems like there are two schools of thought and approach represented  
in the threads over the past couple days.
Understand a single issue, work on it, more or less to the exclusion  
of other items and move on to whatever is next, repeat.
Understand a wider scope of the problem, and include key components of  
the system/architecture.

The former can produce faster incremental results, while the later has  
the advantage of producing a system that will more likely meet  
'customer' needs when deployed and require fewer incremental tweaks.

My take on the recent history is that the rate of progress with the  
single issue approach has been too slow. At least some of the time has  
been taken up with the administrative overhead associated with IETF  
working group spin ups and charter mods.  It seems that for this new  
effort, all the issues should be put on the table so a 'complete'  
management environment can be provided for all our customers.  All  
should be included in the charter and associated schedule.  The get  
worked on as resources are available just like any other engineering  
effort.

So,  I would certainly put into the list items like:
An administrative framework/security.
Integration of configuration with monitoring and event-based data in  
SNMP - this is where issues of SMI coexistence/translation etc would  
be dealt with.
Any modifications necessary to the base protocols to support the first  
two items.

My guess is that the group could probably reach consensus on the  
issues top level issues for a charter relatively rapidly.  Then with a  
path in place, discussion could focus on one or more elements at a  
time, but with the understanding that there are pieces to be put  
together.

/jon
On Mar 17, 2008, at 11:38 PM, Andy Bierman wrote:

> Balazs Lengyel wrote:
>> While agreeing that access control is important, I would also mention
>> that we just made the very first step for the DML. We still have many
>> months of hard work before us, so if we loose focus we are doomed.
>>
>
> Fine -- forget access control.  Go back to telnet for all I care ;-)
>
> But when drafts like partial locking have statements such as
> "must have some access rights", that is a pretty clear indication
> that the WG needs to consider issues like what is an access right?,
> and what does it mean to 'have', 'permit' or 'deny' that access right?
> Pretty basic stuff.
>
> (Remember the emphatic reason the WG could not work on a standard
> ACM 8 months ago?  "We don't understand the requirements!".
> I'm glad that the WG understands them now :-)
>
>
>> Balazs
>
> Andy
>
>
> _______________________________________________
> NGO mailing list
> NGO@ietf.org
> https://www.ietf.org/mailman/listinfo/ngo
>