Characterizing Wireless Network Performance

Accurate performance testing for wireless networks requires understanding how to test for worst case scenarios

As expensive and inconvenient as wires may be, when it comes to transmitting data, they are generally quite reliable. Once working, they tend to stay working, offering the same steady performance level until someone cuts them, unplugs them or digs them up.

However, modulated electromagnetic waves propagating in free-space (aka radio waves or Wi-Fi signals), are anything but steady. They interact with the environment through exotic physical mechanisms such as reflection, refraction, fast fading, slow fading, attenuation and ducting. Even with the best wireless protocols, the best chipsets, the best RF design, the best software and the smartest antennas, wireless performance is going to vary — and it’s going to vary a lot.

There will always be some location where, if the stars align, a user can achieve that magical, maximum physical layer throughput number of 300 mbps, in the case of two-stream 802.11n, for instance. But go far enough away from an access point (AP) and performance is guaranteed to eventually drop to 0.000 mbps. And everywhere else, where people are actually located, there will be performance levels of everything in between.

Performance can also vary significantly, even in a fixed location, due to motion in the environment, interference and the random background noise that came with this universe. Wireless performance is inherently statistical in nature, and accurate performance testing must account for this random component.

To read the rest, request this black paper now!