Blog
As you may already know, your Wi-Fi environment is like a living creature. It changes all the time as different devices move in and around the environment. Performance one hour ago can be completely different than it is right now. It is not enough to take a snapshot from a single location and then mark the performance down in a log book. You need to track the creature's performance over time so that you can learn its behavior at certain times of the day and under certain load conditions. For example, measurements of the network under no load tell you it's maximum possible performance. Whereas measuring the system under heavy load baselines the minimum performance. We need to know what kind of creature we are dealing with and what issues are hampering it's performance. We may have a cheetah of a Wi-Fi network, but if our settings, configurations and designs are caging the creature, then you will never witness its majestic speed. You may be be tricked into believing you have a tortoise.
To know if you are getting best out of you network, you need to understand your existing environment.
- Determine what the expected performance is to be
- Measure the current performance
- Make adjustments and measure again
Also, network performance should be measured from end-user's point of view, as performance is affected by full end-to-end path (from the end-user's device to the application server).
It is very interesting to see how quickly some network managers are willing to replace their existing Wi-Fi hardware. I have often observed how people think that a hardware replacement will provide better performance and solve their Wi-Fi issues. However, you should not be so eager to spend your money. It may be an unnecessary investment.
Over the years, we have taken so many poorly performing networks and improved them with changes in network settings, AP locations or antenna placement. We've also seen where important features of APs or controllers are turned off or not optimal in how they are configured. So, make sure you've gotten the most out of your existing network before blindly accepting that a hardware replacement will "fix everything". Upon close inspection, we have seen that most networks are operating with the manufacturer's default settings. With a bit of attention, you may not have to blame your old hardware. Without understanding how the equipment works, you risk getting little-to-no performance improvement, even after an expensive hardware upgrade.
Wi-Fi Optimization Plan
Since the animal represents a complex system, there is typically no single remedy to solve your WLAN performance problems. But by isolating the issues you can start improving the performance one simple step at the time. For example, radio metrics, such as data rates, retransmissions and airtime utilization will help you to identify radio issues. Typical radio issues are high utilization, non-optimal channel allocation, TxPwr levels that are too low and non-Wi-Fi interference. But sometimes your Wi-Fi beast is choked off by some other network elements lurking in your Wi-Fi environment. It could be your switches, routers or firewalls performing content inspections... and so on. If the radio indicators are good but Wi-Fi performance is lower than expected, then isolate your back-end connection paths by measuring against different back-end elements.
Implementing WLAN Changes
Implement one change at a time and always measure the performance after each change. Correlate the performance against the baseline results. Then implement the next change and measure again. Wi-Fi optimization is a learning process. You will discover something new about your creature after every change and finally understand what kind of settings are best for the environment. So before you blame your hardware, consider taming your network through WLAN performance optimization. You may have a cheetah locked in a cage waiting to break out.