Posted on
More from News
Wi-Fi throughput testing has long been considered the best way to measure system performance and determine if a product is worth its salt or not, but is it really up to the task?
Well, not really, no.
The reason is that this kind of testing doesn’t produce enough data points to be truly useful. For example, if you were to use throughput testing to evaluate vendor equipment, the test wouldn’t reflect reality. This is because you can only generate a single data point under a specific scenario in a highly dynamic environment, which is insufficient.
More data is needed
When it comes to using throughput testing for validating WLAN design, the most accurate and useful way to do this is to test the Wi-Fi network system when it is fully loaded, with multiple clients connected at once, sending data through various access points (APs).
However, this is never done, because it can be difficult and time consuming, which is why throughput testing seems more of an attractive option. But if you have good signal strength, you’ll get decent results for a single throughput client – which isn’t an accurate reflection of what will happen when things get busy.
Throughput testing is also commonly used for troubleshooting Wi-Fi client problems, but it lacks the ability to quantify the real performance that users actually experience, meaning hundreds of people would be needed to identify one user’s problem.
User experience is more accurate than throughput testing
There is a better way to find great Wi-Fi products, troubleshoot the system and validate WLAN design. The answer lies in the user experience, as it is inextricably linked to how the network services and applications interact with the client devices being used by the corporate network.
By collecting, analysing and correlating information from WLAN controllers and client network transactions you will end up with a fuller and more comprehensive understanding of how the Wi-Fi network is really performing when fully loaded.
The cloud is key
While this can seem like a costly operation, reserved only for data scientists, the power of the cloud can mitigate these issues.
Some enterprises have started siphoning client transactions from networks and pushing the data through the cloud, so it can be processed and correlated more efficiently. The goal here is to quantify actual network experience, no matter the device or volume of data – which is something throughput testing simply cannot do.
Network data can be used to transform user client devices into Wi-Fi sensors that can provide a much clearer picture of performance, because you take into account the WLAN controller, network service and real application data.
This is a way to quantify the Wi-Fi performance that clients are actually experiencing, rather than relying on the synthesized data produced by throughput testing - data that cannot be applied to real life scenarios.
