Where are you coming from?
One of the first mistakes that new network operators make is that they don’t have a good idea of where they already are.
Take the case above. Say you want to get to Disneyland. Google can tell you were Disneyland is, but if you don’t have a starting point, there’s really no way to understand how you are going to get to where you want to be.
So the first thing you need to do when you’re trying to operate a new network is figure out exactly where you are.
This concept is known as baselining
At its essence, baselining is really nothing more than taking stock of where you are. Most experienced engineers instinctively know when they reach a new environment that they want to spend some time just getting to know the place before they make any changes. Network Management as a discipline takes this to a much more structured level.
There are a few types of baselines, performance and configuration been the major.
This is the simple act of contain and recording things in the network. One of the common mistakes that I see new network management practitioners make is the ” I want to monitor everything! ” move.
Now there are some grounds for this. Remember this guy?
“When you can measure what you are speaking about, and express it in numbers, you know something about it. But when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind.”
Lord Kelvin, 1891
But on the other hand, in the spirit of the ” Just because you can, doesn’t mean you should ” discussion, I refer you to This famous Albert Einstein quote which gets used all the time in NETMAN circles
“Not everything that can be counted counts….
and not everything that counts can be counted.”
For performance monitoring. Some people like to count exactly how many packets move over a given port, and to some degree, having this information can be useful to perform historical trending analysis, as well as providing SOME indication of where you MIGHT have bottlenecks in the future.
Any full feature NMS should be able to help you perform basic historical polling using the SNMP protocol. In the below video, I’m going to show you how to turn on polling for a trunk interface for a couple of switches, and then create a custom view to put all the trunk ports in the same view.
Here’s a quick example of how to setup a simple performance task using HP’s IMC
* If you need to provide pretty graphs to the bean counters to get budget for the much ended upgrades, this COULD be one of the most important activities for you. Finance guys LOVE graphs and statistics!
One approach, which I find really interesting, was put forward by Terry Slattery on his blog. I’ll let you read it, but in a nutshell, look for errors and don’t worry too much as long as the network is running fine.
Note: I may be putting words in Terry’s mouth here. read through the blogs though. Great stuff from a SUPER smart guy.
The most important thing is to PLAN what you are going to baseline. Think about what’s important to you, to your boss, and what data could be important to justify much needed upgrades down the road.
What about you? Any other interesting use cases for performance data you’d like to share?