802.11: speed versus width on 80 MHz channels

When deploying an 802.11ac system, should you go for speed or RF channel width? This article provides the business case for why 802.11ac needs to be deployed on 80 MHz channels rather than 40 MHz.

When planning an 802.11ac deployment you may hear recommendations to use a 40 MHz channel. While increasing the channel bandwidth is one of the primary ways of increasing data rates, the question remains as to whether it makes financial or technical sense to deploy 802.11ac if you are restricted to the same channel size as legacy 802.11n equipment.

The primary reasons for deploying 802.11ac using a 40 MHz channel include spectrum availability, the need for Dynamic Frequency Selection (DFS) and destructive interference.

Wider channels mean you need more spectrum. Traditionally, in an enterprise setting where multiple access points (APs) are deployed, you would deploy them on different channels. If you use a 1-in-3 frequency re-use plan you would need at least 240 MHz of spectrum or 3 x 80 MHz.

DFS is a feature that lets the Wi-Fi network detect the presence of interference and dynamically move to a clearer channel. Deployment of a 1-in-3 frequency re-use plan with 80 MHz channels requires APs and clients to support DFS. From a client perspective, this means the client must be capable of responding to an 802.11 management action frame request to move to another channel.

When deploying a traditional 1-in-3 frequency re-use scheme, you also have to worry about co-channel and adjacent channel interference. Interference deteriorates the performance of the network. To maximise throughput, it is important to avoid contention on the same channel and interference with adjacent channels. The spectral mask for 802.11ac 20 MHz and 40 MHz channels is the same as for 802.11n. The mask for an 80 MHz channel is an extension of the 40 MHz mask. A 40 MHz channel causes more adjacent channel interference (ACI) than a 20 MHz channel, while an 80 MHz channel is expected to cause more than a 40 MHz channel.

In 802.11n deployments it is common to configure adjacent APs on non-adjacent channels, which means adjacent APs operate on channels that are at least 40 MHz apart. The power from transmissions in adjacent channels can spill into neighbouring channels and cause ACI. When deploying 80 MHz channels it isn’t possible to deploy adjacent APs on non-adjacent channels because there aren’t enough available today.

It is therefore reasonable to conclude that if you are deploying in an enterprise environment using a 1-in-3 frequency re-use scheme with 80 MHz channels they will suffer ACI and performance will be impacted. The more traffic on the adjacent channels the greater the ACI and the greater the impact on performance. This has led to some suggestions that deploying 802.11ac on 80 MHz channels could result in worse performance than deploying 802.11n on 40 MHz channels.

Performance in a 40 MHz channel

Technologies can be compared by looking at the maximum theoretical data rates. There are two technologies in 802.11ac Wave-1 products that are not in 802.11n. These are explicit beamforming and 256-QAM.

Although technically explicit beamforming was defined in 802.11n, it was not implemented because the specification provided the flexibility to implement beamforming in multiple ways and vendors did not support multiple mechanisms due to the implications for product costs. The 802.11ac specifications address this by defining a single beamforming mechanism.

Explicit beamforming requires the beamformee or the client to provide explicit feedback regarding the channel conditions to the beamformer or the AP. The beamformer then uses this information to generate beams toward the beamformee.

The advantage of beamforming is that it increases the range at which the higher data rates can be attained. The advantage of explicit beamforming defined in 802.11ac over the implicit transmit beamforming implemented in many of today’s 802.11n products is twofold. First, because it is based on channel conditions, the beamforming should be more accurate and result in a stronger received signal. Second, because vendors are aligned on a common mechanism it can be used for transmitted and received signals.

The introduction of 256-QAM means that in extremely good RF conditions a signal can be modulated to carry eight rather than six bits per modulation symbol. This results in an increase in the data rate of 1.33x (8/6). This higher data rate can only be achieved in the best RF conditions. However, coupling 256-QAM with beamforming extends the range at which the higher data rate can be attained.

The business case

If you deploy 802.11ac in a 40 MHz channel, explicit beamforming has a very high probability of increasing the range at which you can get higher data rates. In addition to the transmit beamforming capabilities of 802.11n you should also have the benefit of beamforming on the uplink. Explicit beamforming does introduce higher overhead, due to the need to report channel conditions to the beamformer, but overall you can expect a performance improvement.

It is impossible to say exactly what the throughput benefit will be as it varies with RF conditions and traffic loading but is it enough to replace the existing 802.11n deployment? The answer is clearly no. If you are currently in an 802.11g network is it enough to make you upgrade to 802.11ac over 802.11n? The answer to that depends on the equipment costs. The cost of 802.11n equipment is starting to drop significantly, while 802.11ac equipment carries the ‘new technology’ price tag.

The bottom line is if you are considering deploying 802.11ac in 40 MHz channels there will be difficulty justifying it. To get the true gain out of 802.11ac you need to deploy it in 80 MHz channels.