Monday, July 9, 2012

Beamforming and How to survey for it?



Beamforming and How to survey for it?

Beamforming is popular these days with certain WiFi venders, this technology is maybe not a big mystery for the geeks in the WiFi industry, although I see and hear a lot of misconceptions around beamforming.

What is it?
Beamforming allows an access point to concentrate RF energy in the direction of the WLAN client/receiver in order to improve the Signal to Noise Ratio at the WLAN client/receiver and thereby improving WLAN performance (Increased SNR = increased data rates & decreased retransmissions).

So basically you could say beamforming affects the radiation pattern of the access point, so with this in mind anything that affects the RF pattern should be taken in consideration when doing a pre-installation onsite site survey, correct?

Before we elaborate on this further let's dig into beamforming first to grab a good understanding what it actually is (anyway how I see it ;-)).

Beamforming comes in 3 versions, static beamforming, transmit beamforming and dynamic beamforming (sometimes they are called different but for this blog we use these names to differentiate).


Static Beamforming:
Under static beamforming we can understand that an access point will use a fixed directional antenna to focus the RF energy (beam) in a certain direction, the coverage area in this direction will extend in comparison with a traditional omnidirectional antenna. Almost every vendor can do static beamforming when their access point supports external (directional) antennas. The directional antenna will be aimed (fixed) towards a certain area to improve SNR for that area, keep in mind that all mobile WLAN clients have small omnidirectional antennas on board that also need to send traffic back towards the directional antenna of the access point (more on that later). So static beamforming will be mostly used to improve SNR for a certain location or to extend a coverage area.
Some vendors like Xirrus, use an array of static directional antennas to improve SNR in 360 degrees, by assigning a different channel to each beam they can be packed together within a single array.


Transmit Beamforming (TxBF)
Under transmit beamforming we can understand that an access point will transmit 2 or more phase-shifted signals towards a WLAN client/receiver in order to achieve an constructive signal at the WLAN client/receiver. The phase-shifted signals as they travel through the air create constructive signal (in phase) points at various locations in time and space, the goal here is to achieve this at close as possible to the WLAN client/receiver and thereby increase SNR.
To do this correct, possible feedback is required coming from the WLAN client/receiver to help pinpoint the location of the WLAN client/receiver.
Explicit and implicit transmit beamforming are 2 forms of optional components of the 802.11n standard and require feedback from 802.11n clients, as till now 802.11n chipsets do not support either form of transmit beamforming.

Although Cisco has his own proprietary form of TxBF called ClientLink, version 1.0 works only with legacy 802.11ag WLAN clients without requiring feedback. Optimizing SNR values on the 802.11ag clients in a mixed environment (802.11agn) will increase airtime for all 802.11 WLAN clients and is beneficial for overall WLAN performance.
ClientLink 2.0 works for all 802.11agn clients (release 7.2 on the 3600 series access points).


Dynamic Beamforming 
Under dynamic beamforming we can understand that an access point will focus RF energy (beam) towards a certain WLAN client/receiver in a way it's kinda like static beamforming but with a big difference that it is not static and uses an dynamic antenna array that can change it's radiation pattern on a frame by frame basis. It selects the optimum antenna pattern for each communicating device in real time, while actively avoiding interference and minimizing noise to nearby networks and devices.
Test results showed that dynamic beamforming had the highest SNR increase in an overall WLAN deployment of the 3 beamforming technologies as described above.
The only WiFi vendor currently offering dynamic beamforming technology is Ruckus Wireless.


Site survey considerations
By understanding the 3 different types of beamforming, what do we need to take in consideration while doing a onsite pre-installation site survey?

This is how we see it.
Note that these are only additional considerations for the onsite site survey concerning beamforming technologies, during the actual site survey other considerations like capacity, throughput and so on need to be addressed also.




Static beamforming: 
Because static beamforming will cover only one certain coverage area following points need to be taken in consideration.
   Always measure with the antenna type that will be proposed in the bill of materials.
   Adapt your TX power settings before you measure (see additional note at the end).

Transmit beamforming:
In this case only cisco has a proprietary form of TxBF, we recommend to follow their recommendations (I really appreciate all the documentation Cisco put's out there) 
   "Although ClientLink increases the SNR and data rates of clients at the edges of the cell, it does not extend the maximum range of a cell. This is because certain packets that must be heard by all clients (such as beacons) are sent as broadcasts, which ClientLink cannot optimize. These broadcast packets effectively become the limiting factor on the cell size. Therefore cell spacing should not be increased when using ClientLink. In fact, to be conservative, the ClientLink feature can be turned off during a site survey." this makes sense to me I would say ClientLink feature must be turned off during a site survey.
   Take note that ClientLink 2.0 is enabled by default, and needs to disabled by a cli command.



Dynamic Beamforming:
In this case only Ruckus supports this technology, I have a tiny issue with their recommendation to do onsite site surveys. I've got this information from their technical guy at a Big Dog seminar.

   Ruckus recommends that onsite measurements need to be done with the WLAN client/receiver in the beam and with the access point at full tx power, note that the beacons are not broadcasted in the directional pattern but in a omnidirectional pattern to communicate with other WLAN clients/receivers. Also sending out traffic in a directional pattern at full power (20dBm EIRP in Europe) raised an big question mark in my mind that is related towards WiFi math and physics, certainly if you want to support smart mobile devices (see additional note at the end). 
  I would recommend doing the site survey with the access point tx power settings at 14dBm IR.


Additional Note: this note refers to the "antenna reciprocity theorem" (damm that’s a hard one if you're native tongue is not English ;-)) which I refer to when it comes to tx power settings and antenna options.

Basically what this theorem says is the following, If an electromagnetic force of some particular magnitude is applied to the terminals of antenna "A" and the received current is measured at some other antenna "B" then an equal current (in both amplitude and phase) will be obtained at the terminals of antenna "A" if the same electromagnetic force is applied to the terminals of antenna "B".

In human terms, no matter with type of antenna is used on both sides if the input power on both ends is the same then, if you can hear me, I can hear you.
Explained in a real world scenario: if you're laptop has a wireless NIC with a lousy omnidirectional antenna on board and is sending at a tx power input of 17dBm (50mw), and I have the same laptop with the same setup sitting 1km away from each other. In this case we would probably not be able to set up a wireless communication because we don't hear each other, if I connect a high gain directional antenna to my wireless NIC and aim it towards you're direction whenever I can hear you, we will be able to set up a wireless link even if you're lousy omnidirectional antenna don't reach me, as long as we have the same tx power input on both laptops (tell this to all the IT managers that for security reasons are eager to block WiFi signals so that they don't bleed into the parking lot).

So lets put this in an access point and a mobile wireless device relationship. 

Consider what happens if an access point has a tx power input of 17dBm with a directional antenna and the mobile wireless device has an tx power input of 14dBm with a low gain omnidirectional antenna, due to the fact that the access point is more powerful then the mobile wireless device, then "reciprocity" is broken and around the edge of the cell the mobile wireless device will attempt to associate with the access point that it will never reach.

Consider what happens if an access point has a tx power input of 14dBm with a directional antenna and the mobile wireless device has an tx power input of 17dBm with a low gain omnidirectional antenna, due to the fact that the access point is less powerful then the mobile wireless device, also "reciprocity" is broken, but there is a big but here ;-), reciprocity is broken in a way that augments, rather than detracts.


So taking all of this in consideration, and if you're WLAN needs to support smart mobile devices (BYOD, which have a limited power input) it's better to have your access points on limited tx power settings, so take this in account when doing the onsite site survey.  

Please reply, if there are any remarks or somebody has additional information regarding this topic.

About Us

Skyline networks is specialized in a professional range of wireless LAN services worldwide. We offer customized services for every aspect of your WLAN.

The Skyline Blog Authors:

Joeri De Winter
Justin Cetko

Joeri De Winter

Follow Joeri

Justin Cetko

Follow Justin

Twitter feed