How well can you spot irregularities in your network?
Your baseline should provide you with a clear picture of everything you have, including hardware and software. It maps your network as it exists and provides you with the configuration of every single network node.
Network monitoring is like playing a game of ‘spot the difference’. You’re always on the lookout for anything unusual or suspicious that could mean you might have a problem on your hands, and to do that means comparing what you expect to see against what’s actually happening out there.
You need an accurate network baseline to work from.
It stands to reason that you can’t identify problems and anomalies unless you know what ‘normal’ looks like for your particular network. To be able to effectively monitor and manage your network’s activity and behaviors, you need to have a clear picture of how it functions under everyday working conditions.
This means you need to create a benchmark – or baseline – against which you can measure performance.
Your baseline should provide you with a clear picture of everything you have, including hardware and software. It maps your network as it exists and provides you with the configuration of every single network node.
Creating a network baseline.
- First, make sure that you can see your whole network, so that your network monitoring solution can provide you with a complete inventory of physical and virtual devices and nodes (including virtual switches and virtualized application accelerators). This should be an automated process – the last thing you’ll want to do is manually collate your inventory.
- Once you have your inventory, you can then analyze traffic to obtain an overall measure of the health of your network, and a deeper understanding of how it is being used. Tools such Netflow and sFlow will provide much of the information that packet sniffers are able to, but without the overhead that they require. You’ll want data granularity to be as fine as possible, so that you can accurately determine application usage.
- Finally, you should allow this data capture to take place over an entire network cycle (for example, a week for a B2B network, or a season for a retail network). This is so that you can capture the normal variations within your business over a period of time. If you have regular seasonal peaks and troughs that make a significant difference to the volume and type of traffic, it’s a good idea to run baseline reports then, too, as part of the process of capturing what ‘business as usual’ looks like for your network.
Set your metrics.
You can now use your baseline data to determine network performance metrics and as a comparison or benchmark for future monitoring. Since there are no universal standards for the kind of performance metrics you’ll be looking at, you will need to review the specifications of all the devices in your network to determine what best performance for that specific device is. You can then develop your own appropriate set of metrics to work to.
Keep your baseline up to date.
Once your network baseline is established, you’ll want to make sure you keep it updated to take into account any organisational and network changes over time. This means setting a routine of regular network audits to look for alterations, additions and dramatic fluctuations in usage. On average, this audit might be monthly; although if you have a fairly stable or under-utilized network, you could look at conducting it over quarterly intervals instead.
