802.11ac Migration – Part 2: What Nobody’s Telling You About 80MHz and 160MHz Channel Bonding

|
802.11ac Migration - Part 2: What Nobody's Telling You About 80MHz and 160MHz Channel Bonding

 

In part 2 of this 3-part blog series, we strip away the hype and expose hidden details that WLAN vendors gloss over, or perhaps leave out.  Our aim is to tell you what to expect from 802.11ac in the real world and how best to migrate to it as your device population shifts to 802.11ac.

Part 1: What Nobody’s Telling You About 256-QAM

Part 3: What Nobody’s Telling You About Wave 2

So what use is 80MHz and 160MHz channel bonding anyway? Channel bonding is the single biggest performance multiplier, and it is the foundation for WLAN vendors’ claims of 1.3 Gbps speeds for Wave 1, and from 2.3 Gbps for Wave 2 up to 6.7 Gbps when using all 8 spatial streams, should anyone have the nerve to implement them. But let’s get real. Can you actually use 80 MHz and 160 MHz channels in your environment? Let’s explore.

Meru Channel Bonding Availability

We can discount 160 MHz right away. For most enterprise deployments this is useless. As you can see from the chart above, in North America there is only one 160 MHz channel available, and in Europe only two. This renders 160 MHz only useful for point to point WAN links. But what about 80 MHz?

You know from the 2.4 GHz days, that planning dense coverage is a bear with limited channels. Close proximity of cells using the same channel causes performance degrading co-channel interference. Thankfully that problem is less of a concern today, now that most devices support 5 GHz, and 2.4 GHz is simply considered a “lifeline” service for legacy devices.

However, 80 MHz channel bonding returns to that same problem. With only 4 or 5 channels available (depending on your region), it is difficult to plan a dense deployment of APs, because you must reuse channels in nearby cells. Without meticulous power management, you will create co-channel interference, which degrades performance and negates much of the expected gain from the bonding. As we mentioned in part 1 of this series, the top data rates of MCS 8 and MCS 9 are much more susceptible to low signal to noise ratio. The same is also true for bonded channels. So, to get the benefit of 80 MHz bonded channels and 256-QAM, clients need to be within 10-15 feet from the radio.

Recieve Sensitivity for MCS and Channel Bonding

How and Where to Use 80 MHz

Can you use 80 MHz channels everywhere? No! However, 80 MHz has tremendous value in two key applications: mesh backhaul and bridging (point to point and point to multipoint). In both cases you have AP talking to AP, and depending on distances, you can fully utilize the advanced features of 802.11ac. If you are using 802.11n for either of these use cases today, this is the first place to start. You can easily double performance of those links overnight. In client access however, the benefits of 80 MHz are more elusive.

For client access, one of the most practical uses for 80 MHz channels is in isolated high-density Service Areas such as a lecture hall, lobby area, cafeteria or conference center, where most clients have clear line of sight to the AP. Here, it is feasible to buffer the 80 MHz radio from the next major hotspot by surrounding it (above and below) with radios on other access points using non-overlapping 40 MHz or 20 MHz channels. Bear in mind that a 20 MHz channel overlapping the 80 MHz channel will basically render the 80 MHz channel useless, while a client is transmitting on that channel. But it is not strictly necessary to avoid overlapping channels. For example, one AP could be using an 80 MHz channel, and an adjacent AP could be using one of the 20 MHz channels within the 80 MHz range. The 802.11ac radio can recognize when the 20 MHz channel is in use, and holds off its 80 MHz transmission until it is clear. It’s clever, but mostly useless for Enterprises. Implementing 802.11ac’s Channel Sharing in a high density network is unwise – why compromise your 80 MHz channel for transmissions on a 20 MHz channel?

Consequently, careful channel selection and maximum channel separation is crucial to reduce the noise floor and prevent 80 MHz transmissions from being blocked by overlapping channel use on nearby APs. To the chagrin of some AP vendors, you must also disable automatic channel selection and interference avoidance schemes, because 802.11n APs may not be able to recognize an 802.11ac 80 MHz channel correctly resulting in a switch to a channel that overlaps 80 MHz, in order to avoid radar or microwave emissions.

 

80 MHz Channel Plan

Another good way to leverage 80 MHz channels is on the building perimeter, using directional antennas facing inward. Using 80 MHz cells at the building perimeter leaves you with a smaller RF footprint that needs to be buffered by other cells. So, this allows you to use 40 MHz cells more densely. However, bear in mind the line-of-sight requirements noted earlier for optimum performance of the 80 MHz cells.

Excluding hotspot scenarios, 40 MHz is a generally the better option. With eight 40 MHz channels available in North America and nine available in Europe, 40 MHz is more practical. But this is not a leap over 802.11n, since 802.11n also supports 40 MHz channel bonding. Nevertheless, you should be looking to migrate your channel plan to 40 MHz, if you have not done so already. For high density deployments where you want high-performance everywhere, such as in offices, residence halls, hospitals and where there is no one central place where everyone aggregates, 40 MHz provides a greater ability to re-use channels which allows for more cells closer together. That means you can scale capacity, without necessarily improving performance.

How 80 MHz Can Work Against You

While most 802.11n and 802.11ac clients support 40 MHz channels, not all 802.11ac clients support 80MHz channels, and no 802.11n clients can support it. So when an 802.11n client connects to an 802.11ac radio using 80 MHz channels, what happens? The radio drops down to a 40 MHz channel or 20MHz for that connection depending on the capabilities of the client. Obviously this means that during all communication with non-80 MHz clients, a 60 MHz or 40 MHz slice of channel capacity is going to waste. Unless you implement channel sharing (see cautions above), it cannot be used in another cell, as it is reserved for 80 MHz channel use. Therefore the merits of using 80 MHz are wholly dependent on your client mix, which is a moving target. In fact, if the mix is less than 30% 802.11ac / 70% 802.11n, then 80 MHz actually works against you, and you would have gained more capacity by utilizing two 40 MHz channels instead of one 80 MHz channel.

In part 3 of this series we look at the additional features in 802.11ac Wave 2, namely MU-MIMO and additional streams, and discuss how and where you can practically use them.

The entire guide contains bonus nuggets of information —

802.11ac Migration: Real World Best Practices
Get the Guide

 

5 thoughts on "802.11ac Migration – Part 2: What Nobody’s Telling You About 80MHz and 160MHz Channel Bonding"

  1. ip info says:

    Hi
    I have a bridged connection between 2 AP with a clear line of sight (distance between them is: 275m or about 900 feet). I used 80 MHz channel on 5 GHz, with low noise allowed it to 256QAM.

    Anyway, so far so happy, but you said in your great article: “Bear in mind that a 20 MHz channel overlapping the 80 MHz channel will basically render the 80 MHz channel useless”

    Does that mean -let’s say- if someone put his router (on 20 MHz) on the roof between the 2 AP will disturb my connection? or just that mean IF it is same AP.

    1. ip info says:

      Yes, if a nearby AP is using a 20 MHz channel that overlaps your 80 MHz channel, it will disrupt your whole 80 MHz channel, while there is client activity on the nearby AP.

  2. Von Clendenen says:

    We have an PTP around 12 miles apart. How to know when distance and channel width work together or not? I’m curious to know for 40MHz wide channel is there an distance limit or not? Would it be better to use 10MHz or 20MHz wide channel for going this distance or further distances? At what distance would it be advisable to not use 40MHz? At what distance does 10MHz become not an good width to use?

  3. Kyle Tiefenthaler says:

    Can you please give tell me what the labels for the X and Y axis on the line graph featured above, with the Y-axis going from 0-9 and the X-axis going from -48 to -82, along with a general title to the figure? Could you also give a brief detail explanation of the graph to help put this into context for me? I have some knowledge wavelengths/frequencies from my engineering physics courses, but I’m very limited in what In applying this information into networking and computers. If you could provide answers to my questions for the figures in each part of this forum I would be grateful because I find these topics extremely interesting and really would like to learn more about this topic/computers in general.

  4. Eric Camulli says:

    0 to 9 represents the MCS (y-axis) that you are able to achieve depending on your signal strength (x-axis). The poorer your signal strength (farther away from zero), the lower your MCS. This is important because the lower your MCS, the lower your data rate in Mbps that your device can connect at. Here is a link to the full MCS index chart http://mcsindex.com/.

    The higher MCS values are very sensitive to noise and interference. To achieve them (and to get the fastest data rates) you need to be very close to your access point with excellent signal strength.

Leave a Reply

Your email address will not be published. Required fields are marked *