The Not So Definitive Guide to Beamforming
While Wi-Fi suppliers will all say they "support" beamforming, since it is an optional feature built into the 802.11n standard, not all beamforming was created equal. In fact there are a number of fundamental differences with respect to its implementation.
Anyone who's ever used a directional antenna has formed "beams", the goal of which is to provide a stronger signal to Wi-Fi devices (this a good thing). But directional antennas are static and can't follow a mobile client. Transmission and reception of signals is focused in a single direction. But Wi-Fi clients tend to be everywhere and moving around which is why Wi-Fi is desired in the first place.
Beamforming is a specialized method of radio-frequency (RF) transmission most often used within access points (APs). Within a multipath environment, beamforming helps to control and manipulate Wi-Fi transmissions so signal reflections can be combined to be best heard by each client.
APs that support beamforming focus the RF energy they radiate directly at a receiving client device. This can be achieved by amplifying and phasing the signals (using the Wi-Fi chip) or by physically focusing the energy using specialized high-gain directional antennas (smart antenna based). The purpose of either implementation is to improve signal reception at the client and, as a result, increase throughput.
But beamforming can do more than improve average throughput rates. It can be implemented in a way that delivers predictable throughput at longer ranges — performance that is sustained even when the client device is in motion. By focusing transmissions toward receiving clients, some beamforming techniques avoid wasted energy, eliminating unnecessary RF interference in other directions.
Today, there are two popular methods of beamforming (see figure): chip-based and antenna-based. A third method will ultimately combine the two.
Electronic Beamforming using Wi-Fi Chips
Chip-based transmit beamforming is specified in the IEEE 802.11n standard. This means pretty much any Wi-Fi systems vendor can implement beamforming, but that will require client support. Within the chip-level implementation there are three alternative methods: legacy, implicit, and explicit. These methods vary by how much, if any, feedback information from the client is necessary for them to work.
- Legacy beamforming uses signal-processing techniques and multiple transmit paths to optimize the signal sent by an 802.11n AP to older 802.11 a/g/b clients in the downlink direction. It operates in legacy 802.11a/b/g networks without requiring client feedback and can statistically yield 1-3 dB of signal gain with two radios.
- Implicit beamforming leverages some information from the client device upon initial association to determine how best to form Wi-Fi beams toward receiving clients. Still, without explicit client feedback, APs have no way of knowing if the beams are optimum or working properly. Today there are no solutions supporting implicit beamforming commercially shipping.
- Explicit beamforming requires the same signal-
capabilities in the client as in the AP so that APs can continually gather the information about the client environment needed to make dynamic best-path decisions. As chip-level beamforming is optional in the 802.11n standard and there is currently no client support for this option. For the chip vendors, implementing explicit beamforming capabilities provides one of several possible differentiation paths.
Chip-based approaches almost always implement beamforming using omni-directional antennas that transmit signals all directions. Because this type of beamforming is using the Wi-Fi chip to "time" RF transmissions, the chip (because it's busy beamforming) is unable to simultaneously perform spatial multiplexing functions vital to achieving higher data rates that are supported in 802.11n. Therefore most chip-based beamforming implementations will only provide minimal performance and range gains to legacy 802.11a/b/g clients.
More importantly, current chip-based beamforming provides no valuable client feedback, so the system doesn’t know if beamforming is actually working. Additionally, there’s no way for these systems to deal with interference when it undoubtedly crops up.
In contrast, antenna-based beamforming is implemented in special multi-element antenna arrays and firmware that sits above 802.11’s PHY and MAC layers. Also called smart antenna technology and required only on the AP side only, it works with all 802.11 clients (a, b, g, n) and operates in a dynamic mode to adapt to changing environmental conditions and client locations. Currently, implementations are based on individual vendors’ intellectual property.
A major misconception of antenna-based dynamic beamforming is that benefits are only on downlink transmissions from the AP transmitting to the client. However, on the uplink, intelligent antenna arrays use PD-MRC (Polarization Diversity Maximal Ratio Combining) to achieve better uplink for weak clients. These smart antenna system also have the ability to perform predictive receive (listening in a direction). This allows them to predict where the next Wi-Fi packets for any given client will come from and which antennas are best suited to receive those transmissions.
These emerging dynamic beamforming and beamsteering implementations use intelligent antenna arrays (a collection of antennas that are controlled by software) to focus Wi-Fi signals only where they are needed while simultaneously avoiding interference. These smart antenna systems operate independently from the underlying wireless chipsets. This enables concurrent support for beamforming, spatial multiplexing, channel bonding and other essential 802.11n techniques essential to realizing 802.11n performance promises.
Smart antennas continually learn about the Wi-Fi environment, gathering information about clients and adjusting transmissions to keep them optimally focused at all times. They do this using control software on the AP that automatically adjusts the antenna array configuration on a per-packet basis to select the best-performing and highest-quality signal path and optimum data rate for each receiving device.
This approach can result in 9dB signal gain and interference rejection of -17dB or more for a total increase in signal-to-noise ratio or effective gain of up to 27dB. Rejecting interference can actually be more beneficial to overall system performance than improving signal gain. To put this gain into perspective, an increase in client-received signal of 9 dB can yield a 50% increase in channel capacity.
Smart antenna–based beamforming leverages 802.11’s built-in acknowledgement mechanisms (eg. MAC ACKs) to continually determine the quality and performance of a physical RF link. Expert software extracts important information from all 802.11 packets received, such as the sender’s performance, the optimum data rate, signal strength, error rates and approximate location. It then ranks the optimum antenna patterns for each communicating device, keeping track of the best performing signal path for any given client at any point in time.
The firmware tracks thousands of possible connection paths to each client and dynamically selects, on a per-packet basis, the best path.
- Beamforming produces asymmetric signal strength (good downlink, poor uplink), which is very bad in the enterprise.
Not true. Dynamic (antenna-based) beamforming maximizes throughput at ANY range. In common enterprise networks 80+% of all data comes from the Internet or internal servers to the client device (i.e. downlink). On the uplink, antenna-based dynamic beamforming uses PD-MRC (Polarization Diversity Maximal Ratio Combining) in addition to having unbeaten receiver sensitivity, thus Ruckus’ system is perfectly tuned and balanced for the enterprise. These smart antenna system also have the ability to perform predictive receive (listening in a direction) and because they constantly learn about the environment, they are able to predict which potential antennas will be best to receive on.
- Beamforming creates the hidden node problem.
Sure, it is possible in some situations for beamforming to be so effective that while transmitting to one client station, another client station on that AP can’t hear that transmission. Hidden nodes are just as likely with omni-directional antennas but do not show to be problematic in practice. Hidden nodes are also very effectively handled by the 11a/b/g/n RTS/CTS (request-to-send and clear-to-send) mechanism.
- Beamforming stop works after X number of clients connect.
Antenna-based dynamic beamforming is extremely lightweight and maintains "state" per client similar to encryption - which maintains keys for every client. Just like encryption beamforming works equally well independent of the number of clients.