Re: FddiTime, and update to Path config table

[Anil Rijsinghani <[e-mail]>]

> Date: Fri, 24 Jul 92 12:43:01 EDT
> From: [e-mail] (Nelson Ronkin)
> To: [e-mail]
> Subject: Re:  FddiTime, and update to Path config table
>
> This is what our NMS does, it performs the conversion from 80 nsec units
> to milliseconds in the display, since that is more useful to the user.
> It is the job of the NMS to display information the way the user wants to see
> it.
>
> The MIB definition should, I believe, be as consistent as possible with
> the ANSI MIB, so I agree with Pablo that we should keep the 80 nsec units.
>
> Different users may wish to see this in different formats, but the MIB
> definition can only present one of these. It should be the job of the NMS 
> to provide this flexibility.

    As a matter of curiosity: does your nms have specific knowledge of fddi/
    do you have an application rather than a vanilla nms with compiler.

    Please don't rigidly stick to "consistency" as a reason on this type
    of issue.  The ANSI folks made fddi work.  The job this forum is trying
    to get done is make it manageable, from the point of view of a network
    manager rather than that of an implementor.  The former could
    care less about how times are internally represented.  Consistency
    is important to the extent that the protocol specified is adequately
    instrumented.  I would argue that a range of half what is required is
    more of a consistency problem.

    I was involved in a mib where time was represented in 1/256 seconds in the
    relevant IEEE standard!  We changed over to a more human-readable format.
    Mibs ought be reasonably usable without higher level applications or
    special features.

    Anyway, the vote is 3-1 against making this change so far and there
    are bigger and better issues yet to be resolved on this mib, so
    let's move on! :-)

    Anil