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

Re: SNMP improvements



HI,

And one more item to add to the SNMP list below is
  * standard track MIB module so there are complete "solutions"
    for device manufacturers
  (What this is all about is that if you want to be a manufacturer
   of a small networking device, say an wireless access point,
   then there is not a standard set of MIB objects and notifications
   that exist to allow such a device to be managed (configured,
   monitored, controlled, and to emit notifications). Where a complete
   set of objects exists (such as a cable modem), it is only because
   an industry consortium has created it. The result is creation of
   objects (and notifications) that seem more appropriate as 
   common, and being on the "standards track".)
When there is a common set of objects and notifications, then it
is cost effective to create management apps.

At 11:11 AM 9/19/2003 -0700, Wes Hardaker wrote:
>>>>>> On Fri, 19 Sep 2003 11:03:56 -0400, "Harrington, David" <dbh@enterasys.com> said:
>
>David> The NM Workshop identified a number of long-term goals and
>David> which technologies were preferred by the operators to get us
>David> there. I recommend we take RFC3535 as a starting point and
>David> chart a course forward, with long term goals and a strategy of
>David> developing the necessary technologies to reach those goals.
>
>David> I'm willing to help drive this effort and/or help edit and review
>David> documents.
>
>Ok, sounds like a good informational RFC.  Note: it was needed last
>year, so its already late ;-)
>
>So if we're at:
>
>    SNMP <----------> We're here <-------------------------------> Netconf
>
>We need some concrete suggestions to get us forward (safely).  I
>listed a bunch in my presentation in SF and I can reiterate them here.
>This is a cut-n-paste of my previous thoughts on the subject:
>
>    In order to avoid pain, we MUST:
>
>        Move to XML in as short a time frame as possible
>
>        Make SNMP last till XML can fill its (size 4) shoes
>
>
>    Switching to XML quickly
>
>        We have the starting momentum, but there is much to do:
>
>                Security
>
>                Data Modeling Language
>                
>                Data Models
>                
>                Monitoring
>
>        We must start on these NOW 
>
>    Keeping SNMP alive that last few miles
>
>        SNMP has many known problems.
>
>        A few things are needed to fill the gap:
>
>                SMIv2.1
>
>                Bulking Fixes
>
>                Security infrastructure utilization
>
>                Better Encryption Support 
>
>
>Most importantly, however, what this does not mention is that we need
>to get the word spread about *how* we still need to use existing
>technologies in the interim.  I think it should be formally stated
>that in the future one of the goals netconf will be to make use of
>pre-existing SMIv2 documents such that it's still a suitable
>documentation mechanism for today while we wait for something better.
>IE, its the gaps that need documenting for how people should jump over
>them.
>
>-- 
>"In the bathtub of history the truth is harder to hold than the soap,
> and much more difficult to find."  -- Terry Pratchett