Top 4 KPIs for Network Monitoring
KPIs or Key Performance Indicators are the indicators that, in an ideal scenario, we should all follow. After all, when well selected, they are able to give a clear picture of the progress or impact of activities in a particular area. If it's network monitoring, KPIs can help your organization understand and measure their use by assessing their success or failure.
For a network monitoring KPI to work, it needs to follow a few rules:
- measure the impact of network performance on a company;
- offer a clear measure that can be quickly understood;
- identify actions or parts of the business that need attention;
- clearly highlight a problem that can be addressed as soon as possible.
Below we list 4 of the most important network monitoring KPIs and explain what each of them looks at and what advantages your business can expect from following them. Check out!
The main KPI used in network monitoring is availability. It indicates how long your network has been up and running daily and should be a concern of your IT managers. High availability during business hours is essential, but may vary by scenario.
While an accounting office will need high availability from 8 am to 6 pm, it is likely that a hospital will need it 24 hours a day. Then excluding previously scheduled maintenance, this institution needs to maintain an uptime equivalent to the planned uptime (which is easily discovered by subtracting maintenance hours from service hours).
How can a business measure this availability? The first step should be to focus on the entire IT infrastructure of a business rather than on individual components. After all, availability is calculated over the entire system.
Next, we will need to make that planned uptime formula a bit more complex. It looks something like this:
Availability = Planned Uptime - Planned Downtimes / Uptime x 100%
Improving network uptime is a common goal, which can be achieved by smarter distribution of your components, changing infrastructure, and constant optimizations throughout maintenance. To do this, discovering current availability with network monitoring is a must.
2. Latency and Performance
Latency and performance form a single KPI, which refers to the amount of time it takes to send something over a network. Too long latency can cause problems in organizations, so you should aim to bring the number you see here as close to zero as possible.
Latency and performance are said to be closely connected because they depend on each other. With too much latency, your employees won't be able to do simple things, such as previewing a presentation they received via email or uploading a video. Therefore, your team should strive to quantify and improve this KPI.
You can test latency automatically by choosing a tool that does it or estimating it manually. The most popular way to do this is by sending a packet and returning it to you and considering the size of the channel. If the observed latency numbers are:
0 - 100 ms = your uploads are instantaneous;
100 - 300 ms = there is a slight, noticeable delay;
300 - 1000 ms = Your network is slow.
Performance calculation, on the other hand, simulates common user activities and the time required to do so. It can identify bottlenecks, show the minimum configuration required for a system to function, and show how satisfied your users will be when using a particular network.
Another popular KPI in network monitoring is utilization. Utilization increases as more people demand resources from a network and can affect circuit performance. Because of this, it is very important to monitor this KPI to get a sense of the degradation to which the network is subjected.
There are many tools a manager can use to track network usage, either in the form of open source technologies or in the form of proprietary technologies. A best practice when monitoring this aspect of your networks is to monitor each activity and use that data to understand how to best distribute resources.
Among the free software that can be used to monitor networks are Zenoss Core, Icinga 2, and OpenNMS. In common, all of this software has open source that can be modified by your organization as needed and thus can meet all your monitoring demands.
Jitter is an English word that can be used as a synonym for nervousness. In a network, this is the exact metaphor we will use to make you understand the concept. When a delay occurs, there is a difference in the packet arrival interval, creating an oscillation of variable intervals.
This oscillation is the jitter. It is as if there is something that would prevent packets from being sent for a moment. In networks where there is more packet traffic per second, such as those used with VoIP systems, jitter can be a problem that impairs the quality of service.
If your network jitter is high throughout monitoring, it is a good idea to analyze which packets arrive late and check their ratio to those that do not. This will make it possible to correct the jitter buffer (which serves to remedy delays in packet delivery).
In order to use the network monitoring KPIs cited here efficiently, you must understand that they will never be static, and that the monitoring cycle may vary from business to business, and may comprise a few weeks or entire years. For this reason, it is best to choose the metrics that best fit your needs and introduce new KPIs whenever it seems necessary.
Most IT monitoring KPIs require some sort of qualitative analysis to interpret. Still, they are important in network monitoring because they quantify performance, enable you to meet goals, and help your business make better use of its resources.