In the first part of this series (click here to read Part 1), we focused on WLAN design as it related to security and performance. For today’s discussion, we will turn our focus to the clients themselves. After all, without clients to use the WLAN we design, what’s the point?

Let’s start by taking a look at the environment around you right now. Take a moment to stop and count how many wireless devices you have within 3 feet of you right now. Think about how many of these devices you have on you at all times. If I were a betting man, I would bet that your number was somewhere between three and 5 wireless devices near you this very second. Don’t fret, the FCC and other regulatory bodies take steps to ensure that the amount of electromagnetic radiation emitted by these devices is not harmful to your body. For most people, it is very common to have a wireless watch, smartphone, and some type of tablet on them (or within a couple of feet) at all times. Even if these devices are all made by the same manufacturer, they all behave differently within a WLAN.

These behaviors are important to understand so the underlying WLAN is designed as optimally as possible. Yes these behaviors can be affected by orientation, battery level, and construction materials used, but they aren’t things that the WLAN designer will spend a lot of time evaluating. For purposes of WLAN design, we want to focus on the deeper technical specifications of the WLAN chipset used by the device. Within any WLAN chipset there are varying numbers of radio chains, antenna connections, and available output power. These pieces help determine multiple values that are important to the WLAN designer such as: receive sensitivity, maximum data rates, and maximum throughput. So how do these values affect the design?

When a WLAN is designed, one of the first things that should be taken into account is what the network will be used for. VOIP, Video, and large data transfer networks (such as graphics editing or scientific research) all have different requirements from a design perspective. In VOIP networks, we need super low latency/jitter and minimum signal coverage throughout the environment of -67dbm with overlapping coverage from 2 APs. At the same time, heavy data transfers might only require high data coverage from a single AP and latency/jitter might not matter as much. As you see, network use is a very important variable to understand prior to doing any AP placements on floor plans etc.

You might be wondering what questions you should ask about the client devices. Some important ones are:

• What chipset is used in the device?
  • This will provide valuable information such as supported PHY rates that will allow you to determine how fast the client should be able to transmit data.
• Are the devices mobile or stationary?
  • If devices are mobile, you will need to evaluate what type of security is required in the network and what tasks the mobile device will be performing, to determine if the expected level of service can be ensured. For example, using 802.1X security with a VOIP handset might provide choppy performance, one way audio, or dropped calls if the handset isn’t Wi-Fi CERTIFIED Voice-Enterprise by the Wi-Fi Alliance.
• How many devices are expected to be used in the network?
  • This is very important to help determine AP placements as larger numbers of devices may require more RF spectrum to be used in a given area. High density, or very/ultra high density, areas will provide higher levels of service by operating at 10-15mW output power (vs. 25-50mW) with smaller cell sizes so more channels can be used.

Speaking from experience, these questions are very important to answer both prior to designing and if the network is already installed, troubleshooting. I recall a time in my career working for a WLAN manufacturer (most of you know them as the SCA company) when I was troubleshooting medical devices at a hospital in the northern US. After evaluating the network and finding adequate coverage present, I began evaluating the client devices exhibiting problems. What I learned was that the devices used a subpar WLAN chipset that had not been manufactured in over 5 years. This particular chipset required a tremendous mount of CPU power that was drawn from the medical device itself. The problem was that the CPU on the device was sized just large enough to provide for the needs of the device, not including the WLAN operation. The WLAN connectivity had been an afterthought. It’s also important to note here, that even though device issues were not the fault of a poorly designed WLAN, it was still what users perceived as the cause of their problems.

To further illustrate the importance of the client types in the network take a look at the chart below. You will see that there are significant performance increases based on the number of spatial streams present in a device.

Expected TCP/IP Throughput (courtesy of DivDyn presentation)

If you were to take a 1GB file and transfer it across the WLAN, a 1 stream client would need just under 171 seconds, while a 3 stream client would only need 51 seconds. That’s a HUGE difference in transfer times! With that in mind, you can now see how important network use really is in design. If we were designing a network for file transfers, we might want to include more channels in the environment and things like load balancing to speed up file transfers from clients.

Lastly, as you probably have surmised by now, no two networks will ever be identical. Therefore, great care should be taken by the design engineer to develop and implement the best possible WLAN. Even the best designed network will need adjustments over time as changes to the environment happen constantly. The work of a good WLAN design engineer never ceases.

For a deeper dive into how these values affect the WLAN, check out this presentation by Darrell Derosia from the 2017 WLAN Professionals Conference.

-Scott