intro

[Merike Kaeo <[e-mail]>]

I've been working with a number of FDDI implementations here at the
National Institutes of Health and am testing the FDDI MIB
implementations as I come across them.  It's been suggested that I share
my experiences with this group.  So, here's some comments I have......

- Until variables are in some *standard* space for all implementations
things are a pain.  The normal user will probably not want to compile
varying RFCs (be it the pre-1285 draft, 1285 and/or whatever the next
one will be for the SMT7.2 implementation) to get SNMP access for a
multi-vendor FDDI.

However, since the above is unavoidable til things settle down it would
be useful to users if vendors spell out which version they are using.  


- Getting the mib values means nothing to a layman.  After not looking
at it for a while I even forgot why T_Neg was 62500 instead of 5ms. 
Same thing, but only if you know why.  Same with codes for SMTCFSstate. 
I actually ran into some code which returned an in-house value for the
latter.  [No value is better than wrong value I
always say.....I kept second-guessing my sanity as to whether the ring
was or wasn't in wrap]

But this has nothing to do with the MIB........this just means I need a
manager that's smart enough to interpret the values.  They're coming....
[I'm talking scenarios of routers and servers in a ring w/out
concentrators, since most concentrators have good management
capabilities].


-  Most use I want to get from SNMP monitoring of FDDI:
     1. Show me visually where the ring is wrapped when it occurs
           [trap message of CFM state-change should take care of this]
     2. Warn me that ring is having problems (i.e error rate high)
     3. Let me set timer values via SNMP to optimize ring performance
          and enable/disable ports


So much for my opinions.   I'd actually be interested in getting put on
this mailing list if possible.  Thanks.

Merike Kaeo
National Inst. of Health
[e-mail]