Not all devices are created equal.
I know this seems like a piece of Captn’ Obvious wisdom, but it bears thinking about a little in context of network management.
One of the things which I see all the time is someone asking to do XYZ on the device. Whether that’s pull serial numbers from power supplies, or read the sticker on the back of a switch. There are some things that are just outside the realm of possibility, or would just be to difficult to put into place.
If you are seriously looking at implementing an NMS, you need to get friendly with SNMP. Simple Network Management Protocol is probably the most common management protocol on the planet.
To be honest, SNMP is a second language and I would highly recommend anyone who wants to get SERIOUS about network management pick up a book or two and start learning it. SimpleWeb has some tutorials, podcasts, and slide decks that they make available which may be a good place to start.
In a nutshell, SNMP MIBs fall into two major categories
Public – These are the standard MIBs that are defined by the IETF. These are your friends, the bridge MIB, dot3 MIB, Entity MIB, etc.. MOST vendors should support these.
Private – Also referred to as Enterprise, these are the MIBs which Vendors write to support their own device specific functionality.
Occasionally, someone brings in a non-snmp capable device and asks for it to be monitored. And then they complain because you can’t make the same pretty graphs.
If it’s not instrumented in the device, we can’t do anything with it.
Let me say that again…
If it’s not instrumented in the device, we can’t do anything with it.
Here’s an example: Say someone comes to you and says ” Hey! Can you please tell me what the serial number is on the power supply in XYZ vendors chassis switch?”
I check the MIBs and it seems that XYZ vendor hasn’t instrumented serial numbers as one of the piece of information which they make available. So the answer is ” No, I can’t”
Then they complain that this NMS stuff, or the specific NMS product sucks. Remember
If it’s not instrumented in the device, we can’t do anything with it.
kontrolissues 
This is actually one of the slightly less obvious reasons for buying better quality equipment. e.g. I’ve found that Cisco’s instrumentation is far ahead of say Dell’s when it comes to polling a switch for information on things like temperature. It’s really annoying when you can get data at the CLI, but not via SNMP.
Similarly, APC is pretty good about exposing useful UPS-related info via SNMP, whereas some cheaper UPSes don’t give you much.
100% agree
Actually, I’ve found that even within the same vendor things can very quite a bit. web managed vs fully managed can have varying degrees of SNMP instrumentation. In some cases, even if they were built on the same ASICs.
On the UPS, I bought an “snmp enabled” PDU about six months ago and although the CLI was ok ( allowed me to turn specific plugs off an on remotely) the SNMP was horrible. No writable MIBs at all, barely any readable.
Should have spent the extra $$$ and got an APC or Powerware.
Actually perhaps that should be re-phrased as “higher end” equipment, rather than “better quality” – the cheap Dell switch was fine at its core task of switching, but it just didn’t give out a lot of info for monitoring it.