T.R | Title | User | Personal Name | Date | Lines |
---|
1321.1 | better to stick with the standard | TOOK::MATTHEWS | | Wed Aug 07 1991 16:46 | 13 |
| John, the names that you find incongruous are the MIB II standard
attribute names used in the standards specification. To the
question of can we use something else; the answer is yes. However,
to do so means that we have to deviate from the standard and that
leaves us vulnerable to being non-standard.
It is unlikely that we will change these names. The MIB II standard
and the CISCO documentation will use the standard names. Changing
them means we have to explain the mapping from our non-standard
names to what CISCO and the rest of the industry uses. I am afraid
that we need to get used to these names and use them as is.
wally
|
1321.2 | some do it | ENUF::GASSMAN | | Thu Aug 08 1991 13:15 | 5 |
| I've seen MIB editors that allow the person doing the mib to also input
'human readable' text to go with the attribute names. This makes for a
more friendly system.
bill
|
1321.3 | MSU has it. | MADMXX::C_OUIMETTE | Wampeters and Foma | Thu Aug 08 1991 16:45 | 5 |
| FWIW, DECMCC-MSU has the human-readable text addition for snmp.
Makes life easier.
chuck
|
1321.4 | Edit the MIB... | CHRISB::BRIENEN | DECmcc Bridge|Station|SNMP Management. | Thu Aug 08 1991 18:25 | 14 |
| Maybe I shouldn't mention this, but...
There's nothing to prevent a user from editing the MIB definition,
(changing attribute names to what the user wants displayed) before running
the MIB Translation Utility.
The problem becomes one of keeping documentation straight and recognizing
problems in parsing if multiple names conflict.
Assuming the Translation Utility is enhanced (in a future version)
to generate online help from the MIB definition, at least the first
problem would be eliminated...
Chris
|
1321.5 | whoa | MKNME::DANIELE | | Fri Aug 09 1991 14:31 | 12 |
| But I think there is a real difference between mcc and
MSU/other products. They use object ids as the real internal
identifier of an 'object' (a MIB variable). The name is simply an
aid to humans. DECmcc use the presentation name of an attribute as its
real internal identifier. The object id in this case is simply a
a string that is associated with the attribute.
Thus changing the name in DECmcc seems, at least to me, to have
much larger implications. Like potentially breaking things?
Mike
|
1321.6 | Talking about different things? | CHRISB::BRIENEN | DECmcc Bridge|Station|SNMP Management. | Tue Aug 13 1991 12:15 | 24 |
| > DECmcc use the presentation name of an attribute as its
> real internal identifier.
I think of the MCC attribute code as the "real internal identifier"
(though I suspect many would disagree with me).
MCC codes are what gets passed across the mcc_call interface...
> Thus changing the name in DECmcc seems, at least to me, to have
> much larger implications. Like potentially breaking things?
Really? If I change a *vendor specific* attribute name from
"proXfaceX25PhyInBytes" to "Bytes Received", what in DECmcc gets
broken? Nothing.
Maybe we're talking about different things:
I'm talking about making changes to the vendor specific attribute
names in the vendor MIB that get run through the MIB Translation
Utility.
Maybe you're talking about making changes to MIB I or MIB II
attribute names? I agree that the latter could cause major headaches...
Chris
|
1321.7 | | TOOK::STRUTT | Management - the one word oxymoron | Wed Aug 14 1991 00:17 | 24 |
| I think I would disagree with .1
I don't think the intent of the MIB RFCs is to define the user-readable
names (or strings) associated with each variable, but rather to have
a program-readable (and hence unique) variable name. Of course, I
might be wrong - if someone could indicate a page/section number in an
RFC, I'd appreciate it.
I believe that many have chosen to make the MIB name into the
presentation name that the user sees. That seems like a reasonable
DEFAULT. But it is certainly different from the way most other
MCC-manageable classes are defined.
It seems quite reasonable to use alternate text strings for
presentation names. However, should Digital do this (for standard MIB
definitions)? Or should we allow our customers to change them? (We
already do, but we don't document how). Then we have to consider how
to deal with Enterprise MIBS - we certainly would not want to prompt
the user of the MIB translator utility to provide the text for each
variable.
Any suggestions from you folks nearer to customers?
Colin
|
1321.8 | Consistent naming is required. | SUBWAY::REILLY | Mike Reilly - New York Bank District | Wed Aug 14 1991 15:50 | 30 |
| I am currently working with a customer whose has a large
multi-vendor environment. We have a large number of SNMP
'monitored' devices. When problems occur with a SNMP monitored
devices we consult with the customer service organizations of these
vendors. The 'language' we speak when dealing with these vendors is
'MIB variable names'. The tools I am using better have MIB variable
names which match what the support person is requesting. The
ability to change the presentation name for a variable is a nice
feature right up until the moment a support person asks me for the
value of the XyzNetDlkOutPkts and the XyzDlkOutPkts counters.
In my 'self' defined interface I have two variable called
Data link out packets counter #1, and
Data link out packets counter #2.
Which is which??
BTW: XyzNetDlkOutPkts was intended to mean the number of packets
send from the Network layer to the Datalink layer, and
XyzDlkOutPkts the number of packets actually send from the Datalink
layer.
What would be nice would be the ability to hit the 'help' key while
pointing to a variable name, and have the help which the vendor
supplied in the DESCRIPTION field of the concise MIB definition
appear in a pop-up window.
- Mike
|
1321.9 | nice wishlist item | TOOK::CALLANDER | Jill Callander DTN 226-5316 | Mon Aug 26 1991 13:35 | 7 |
| I agree the point and click would be nice, and it has been on the
"nice" list since we started the project. One of these days the
niceities will win time over the functionality, but for now
we are still trying to get the base system to a higher degree
of overall functionality.
|