The Most Valuable NMS Tool
A tool that you use regularly to see which interfaces are running at high utilization or have high errors is a start.
I've been attending and presenting at Enterprise Connect for several years now and have collected several interesting tidbits along the way. One of the interesting discussions happened a couple of years ago when I was participating in a panel about network management tools for Voice and Video. Someone asked me what I thought was the best network management tool. I thought I'd share some thoughts along these lines in this month’s blog.
What's the Best Network Management Tool?
My somewhat flip answer to the question "What is the best network management tool?" was "The tool that you have and use regularly is infinitely more valuable than all the tools that you have but rarely use." It parallels a similar thought among photographers: the camera that you have with you to capture a scene is infinitely more valuable than the high-end camera that's sitting at home in the camera bag.
In Network Management, one of the key factors in a smoothly running network is detecting when the network is not running smoothly. You only know when it isn't running smoothly when you collect enough information to detect abnormal behavior. A tool that you use regularly to see which interfaces are running at high utilization or have high errors is a start.
A tool that automatically detects and reports on error-disabled switch ports allows you to develop a view of how smooth the network is running. If you follow networking best practices, you will have implemented various configuration parameters that protect the network from serious accidents like spanning tree loops. These configuration parameters might be LoopGuard or RootGuard on inter-switch links. You may also use BPDUGuard on edge ports to prevent an accidental loop creation when someone "cleans up loose cabling" by plugging a dangling RJ45 connector into a nearby wall jack, creating a spanning tree loop. Your tool should identify any of these things and after a while, you become accustomed to when and where such problems occur and their symptoms.
I favor network management tools that are easy to deploy. In my experience, big infrastructure tools take a long time to deploy and offer much less value than tools that are easy to deploy. If you are stuck with a corporate decision to buy and install a high-maintenance tool, select some basic functionality and get it implemented quickly, then move to more advanced functions.
Regardless of the type of tool, you should start with a focus on two general functions:
1. Event management;
2. Collecting basic interface performance stats, including in/out packets, in/out bytes, interface errors, discards, and overruns from all network interfaces.
Create reports and alerts for each function. I like to have an overall event management report generated daily to list a count of all event types that have been received. The daily event summary report indicates the types of problems that the network is detecting and reporting. The single critical event is just as important as the device that is constantly reporting a low-severity problem.
The interface stats report should include several Top-N reports on interface utilization, interface errors, and interface discards. This report allows you to identify interfaces that are congested or that are experiencing high numbers of errors.
After a few weeks of running the above reports, you will have a good idea of the problem spots in the network and the general health of your network. You can then start to implement additional functions, such as configuration backups, checking for correct router redundancy implementations, and identifying redundant interfaces that are not properly configured and working.
Next page: Management of Voice and Video