Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Fwd: Re: [802SEC] FW: Fwd: IEEE 802.11 SNMP MIB?

>Date: Wed, 04 Jul 2001 09:06:42 +0100
>To: "Bob O'Hara" <>
>From: Tony Jeffree <>
>Subject: Re: [802SEC] FW: Fwd: IEEE 802.11 SNMP MIB?
>Bob -
>There are some words in the current text of the 802 overview & 
>architecture on management; it mentions both OSI management & SNMP as 
>possible approaches. There are also some words in 802.1F on how to define 
>MIBs and on commonly applicable object definitions; however, 802.1F is 
>largely aimed at the OSI stype of MIB (and is probably overdue for a revision).
>There are examples of both types of MIBs (OSI & SNMP)in various 802 
>standards. 802.1 has lately taken the view that (at least for new 
>standards it develops) it is appropriate to define an SNMP MIB; 802.3 
>continues to develop OSI MIBs, but there are examples (for instance, Link 
>Aggregation) where 802.3 includes an SNMP MIB as well. So the short answer 
>is that it is up to you as to which style you use - my guess is that SNMP 
>is the predominant mechanism used in your marketplace (as it is for LANs 
>in general).
>Whichever path you choose, I would strongly recommend that you make a 
>clear distinction between:
>- The abstract definition of the managed objects that you need, in terms 
>that are independent of the management protocol that you plan to use; and
>- The definition of the syntax required in order to convey operations on 
>those objects in a given management protocol.
>It is very easy to leap into the second of these pieces of definition, 
>without clearly defining the first - the result (in the case of SNMP) is a 
>load of ASN.1 syntax that doesn't really answer the question "OK, so what 
>does this MIB actually *do*?". It is actually very helpful to do the 
>abstract definition first; most people in the room can understand what it 
>means, whereas once you start getting into the SNMP (or OSI) syntax, all 
>bets are off. A longer term advantage of this separation is that if the 
>management protocol of choice changes (for example, if SNMP got replaced 
>by a standard for web-based management) then it is relatively easy simply 
>to translate the syntax into the new protocol, without messing with the 
>underlying functionality.
>Take a look at the 802.1 and 802.3 standards for examples - for example 
>Clause 9 of 802.1X defines the managed objects for 802.1X in the abstract; 
>Clause 10 defines an SNMP MIB that allows those objects to be operated on 
>using SNMP.
>At 16:24 03/07/2001 -0700, you wrote:
>>Please read the email thread, below.  It asks if there is a policy as to how
>>MIBs are defined.  I don't know the answer to this.  What is the position of
>>802?  Is there one?
>>  -Bob
>>-----Original Message-----
>>From: Tim Godfrey []
>>Sent: Tuesday, July 03, 2001 4:16 PM
>>To: T. M. Kurihara
>>Cc:;; Bob O'Hara
>>Subject: Re: Fwd: IEEE 802.11 SNMP MIB?
>>Tom and Bob,
>>The 1997 version of 802.11 specified the MIB according to OSI rules, but
>>Annex D was removed prior to paper publication. Subsequent versions of
>>802.11 (802.11-1999 and supplements) are according to SNMP.  So the short
>>answer to your question is yes.
>>I have CC'd your question to Bob O'Hara, who is a member of the 802
>>executive committee and the editor of the 802.11 standard. He is an expert
>>on the 802.11 MIB and higher level 802 architecture, and will be able to
>>provide a more detailed explanation and history.
>>At 17:33 7/3/2001, T. M. Kurihara wrote:
>> >Tim, Bob DeRoche has asked for an authoritative response to his
>> >observation and question.  Could you respond or make arrangements to have
>> >a response for Bob?  Thank you very much for your assistance.  \\TomK//
>> >
>> >>From: Robert DeRoche <>
>> >>To: "Thomas Kurihara (E-mail)" <>
>> >>Cc: "NTCIP Profiles WG (E-mail)" <>
>> >>Subject: IEEE 802.11 SNMP MIB?
>> >>Date: Tue, 3 Jul 2001 18:09:23 -0400
>> >>X-Mailer: Internet Mail Service (5.5.2653.19)
>> >>
>> >>Tom,
>> >>
>> >>I was researching how property, value domain, and generic property domain
>> >>are expressed in other industries and ran across the following file:
>> >>
>> >>pdf. This is a supplement to IEEE 802.11.
>> >>
>> >>On document page 57, it describes changes to the "SNMPv2"  MIB that
>> >>in IEEE 802.11!!!  Are the IEEE 802 WGs now defining MIBs in the SNMP
>> >>This is a very important issue in regard to a number of NTCIP Standards
>> >>I would appreciate an authoritative answer.
>> >>
>> >>Regards,
>> >>
>> >>Bob De Roche
>> >
>>     Tim Godfrey at Intersil
>>     Ph 913-706-3777   Fax 913-664-2545