MobileParadigm » beamforming

Comparison of beamforming technologies

Paul DeBeasi — Mon, 25 Jan 2010 10:00:00 +0000

This is the fifth post in a series that delves into the subject of beamforming. In this post, we compare all three beamforming methods (static, transmit, dynamic).

Although all three beamforming technologies provide SNR gain, they are quite different. In addition, all types of beamforming can exacerbate the following problems:

Sticky client problem: The sticky client problem occurs when stations (STAs) remain associated with a distant AP (i.e., when they are “stuck” to an AP), even when a closer AP can offer the STA a better data rate or voice call quality.
Hidden node problem: The hidden node problem typically occurs when one or more STAs on the same channel cannot “hear” one or more of the other STAs, resulting in channel interference. In some beamforming instances, this can also be true of STAs not hearing the AP.

A static beamforming array, such as the Xirrus product, is useful in dense deployments such as a large conference room. In addition, a directional antenna can be used to form a static beam to direct energy down a hallway or toward the interior of a building.

The extent to which transmit beamforming (TxBF) will reliably improve SNR is dependent on many factors such as the effectiveness of sounding the channel between the AP and the STA. The WLAN industry is in the early stages of TxBF deployment and I expect that the industry will gain useful real-world experience with Cisco’s proprietary TxBF solution (ClientLink). I also anticipate that silicon vendors will eventually integrate standards-based TxBF mechanisms once the IEEE ratifies implicit and/or explicit beamforming.

Dynamic beamforming is useful in many deployments because it can provide SNR gain for all STA types in the uplink and downlink directions. Since dynamic and transmit beamforming operate in ways that are complimentary, it is conceivable that future enterprise products will combine both techniques to provide even greater SNR gain. Dynamic beamforming will not likely become widely deployed unless WLAN silicon vendors support this feature, thus driving down development cost and complexity.

Refer to the chart for a more detailed comparison of each beamforming method.

Dynamic beamforming

Paul DeBeasi — Thu, 21 Jan 2010 10:00:21 +0000

Dynamic beamforming focuses RF energy in a particular direction and with a particular shape in order to increase the Signal to Noise Ratio (SNR). In a sense, it is like static beamforming in that it focuses RF energy, but it is different because the antenna array’s radiation pattern can change from frame to frame (see Figure).

Figure 1: Dynamic Beamforming (Source: Ruckus Wireless)

Dynamic beamforming is also different from TxBF in several ways because it can:

Dynamically change the propagation pattern on a frame-by-frame basis in order to optimize the pattern for every STA over time
Provide SNR gain for both legacy and 802.11n STAs without requiring any changes in the STA
Be used in conjunction with other 802.11n performance-enhancing techniques such as spatial multiplexing
Improve both downlink and uplink performance
Provide interference rejection

Dynamic beamforming has the potential to offer the highest SNR gain of the three beamforming technologies. Note that broadcast traffic, such as Beacon frames, will usually be transmitted using an omnidirectional pattern in order to communicate with nearby STAs in all directions. The only vendor currently offering an enterprise class dynamic beamforming technology is Ruckus Wireless.

Next time we will compare all three beamforming methods.

Transmit beamforming

Paul DeBeasi — Mon, 18 Jan 2010 10:00:08 +0000

In this post, we continue my discussion of beamforming by focusing on transmit beamforming. As previously mentioned, beamforming is a method of concentrating radio frequency (RF) energy in order to improve the signal to noise ratio (SNR) at the receiver, thereby improving network performance and predictability.

Transmit beamforming (TxBF) is a method of transmitting two or more phase-shifted signals so that they will be in-phase at particular points in space where the transmitter believes the receiver to be, thereby increasing SNR. Two forms of TxBF are optional components of the IEEE 802.11n draft amendment to the 802.11 standard. Explicit and implicit TxBF require feedback from 802.11n stations (STA) and thus will not operate with legacy stations at all. Enterprise WLAN vendors do not support implicit and explicit TxBF at this time because the 802.11n chipsets do not currently support either form of TxBF.

Cisco has introduced a proprietary form of TxBF called ClientLink. ClientLink can work with 802.11 g/a STAs because it requires no modifications in the STA. ClientLink is designed to improve the SNR for legacy STAs in the downlink (AP-to-STA) direction only. A boost in SNR can improve the STA’s “rate over range” performance because the modulation rate for 802.11 STAs will increase as the SNR increases. Improving rate over range performance is particularly important for legacy STAs because they can consume considerably more airtime than 802.11n STAs, and therefore can reduce the achievable throughput of 802.11n STAs. Alternatively, Air Time Fairness (ATF) mechanisms can also regulate legacy STA airtime consumption.

TxBF changes the phase of the original signals in relation to each other and transmits the phase-shifted signals using two or more antennas to the STA (see figure). As the signals propagate through the air, they additively combine at various points in space. The figure shows an example of two out-of-phase signals propagating from an AP to a legacy STA. The green dots represent the points in space where the two out-of-phase signals combine to form a signal with an SNR that is up to 3 dB higher than (i.e., twice as high as) the original signal. In a multipath-rich environment, even higher levels of gain are theoretically possible. Cisco is the first enterprise WLAN vendor to implement TxBF, and it claims it can achieve 4 to 6.5 dB of SNR gain in a multipath-rich environment. A 6.5 dB gain is an increase of approximately 4.5 times the original signal.

Figure 1: Transmit beamforming (source: Cisco Systems)

The challenge with TxBF is figuring out how to modify the transmit signal phases for the greatest possible gain. In the ClientLink implementation, the AP uses frames received from the STA in the uplink direction to determine how to modify the phase in the downlink direction. TxBF assumes that the uplink channel characteristics and the downlink channel characteristics are “reciprocal” (i.e., the same) in both directions. In reality, the uplink and downlink channels may not be reciprocal, especially when the STA is moving. So, in practice, TxBF performance gains will vary from moment to moment and STA to STA.

Unlike static and dynamic beamforming, TxBF does not change the antenna radiation pattern. Cisco’s TxBF implementation uses omnidirectional antennas that cause the signals to radiate in a doughnut-shaped pattern. Therefore, referring to TxBF as “beamforming” is somewhat misleading because it does not actually form a directed beam. Cisco’s TxBF achieves an SNR gain at “points in space” (i.e., the green dots in the figure). So the receiving STA must be located at the right point in space in order to achieve the maximum SNR gain. In contrast, static and dynamic beamforming achieve SNR gain throughout the radiated coverage area because both techniques focus the radiated energy.

TxBF and spatial multiplexing are mutually exclusive. This is because spatial multiplexing transmits different signals on each antenna, whereas TxBF transmits the same (phase-shifted) signals on each antenna. So, it is impossible to use TxBF and spatial multiplexing at the same time. That is why enterprises should use TxBF to provide SNR gain for legacy STAs only. In contrast, static and dynamic beamforming can operate in conjunction with spatial multiplexing.

Next time, we will look at dynamic beamforming.

Static beamforming

Paul DeBeasi — Thu, 14 Jan 2010 10:00:08 +0000

This post continues my discussion on beamforming which is a method of concentrating radio frequency (RF) energy in order to improve the signal to noise ratio (SNR) at the receiver, thereby improving network performance and predictability. In this post, we discuss static beamforming.

Static beamforming provides a fixed radiation pattern by using a directional antenna. Virtually every vendor provides APs with removable antennas so it is easy to swap an OMNI antenna for a directional antenna. Some vendors, such as Xirrus, use an array of static beamforming antennas to create densely packed multiple channels in much the same way that a cellular tower uses directional antennas to create cellular sectors. By assigning a different channel to each beam, many non-overlapping channels can be densely packed together within a single array (see Figure).

Figure 1: Static beamforming (source: Xirrus Networks)

Static beamforming can provide an signal-to-noise (SNR) benefit to both legacy wirless protocols like 802.11a/b/g and newer 802.11n stations. Since the antenna propagation pattern is static, the AP cannot adjust the radiation pattern on a frame-by-frame basis in order to track a station as it moves through the enterprise. Therefore, unlike transmit beamforming and dynamic beamforming, it does not take advantage of knowledge of the WLAN channel between the AP and the station in order to further optimize signal propagation. An array of directional antennas, such as the Xirrus Array, can improve SNR in 360 degrees.

Next time we’ll look at Transmit beamforming.

Introduction to beamforming

Paul DeBeasi — Mon, 11 Jan 2010 19:00:44 +0000

Beamforming is a method of concentrating radio frequency (RF) energy in order to improve the signal to noise ratio (SNR) at the receiver, thereby improving network performance and predictability. Enterprise WLAN vendors are now integrating beamforming technology into their access points (APs) so it is important to understand the types of beamforming and the benefits that beamforming can provide.

Beamforming is not new. At the most basic level, beamforming affects the radiation pattern of a wireless signal. The radiation pattern refers to the way in which the electromagnetic waves propagate outward from the antenna element. For example, the most commonly deployed antenna is the omnidirectional antenna. The radiation pattern for the omnidirectional antenna is in the shape of a doughnut (see Figure 1). This type of antenna is a good choice for hotspots or any environment where the intent is to propagate the WLAN signal in a broadly dispersed pattern.

Figure 1: OMNI radiation pattern (source: Cisco Systems)

WLAN venders currently offer three types of beamforming. Each type affects radiation patterns in different ways.

Static beamforming involves the use of internal or external antennas that have a fixed radiation pattern (such as a YAGI antenna) that emits a directional radiation pattern (see Figure 2). A directional pattern is a good choice in environments where the intent is to propagate the WLAN signal in a particular direction, such as down a hallway or toward the interior of a building.

Figure 2: Directional radiation pattern (source: Cisco Systems)

Transmit beamforming propagates two or more phase-shifted copies of a signal on a frame-by-frame basis, so that they will be in-phase at particular points in space where the transmitter believes the receiver to be, thereby increasing SNR.

Dynamic beamforming changes the antenna radiation pattern on a frame-by-frame basis using a central processing unit (CPU) controlled antenna array in order to increase the SNR at the receiver.

In subsequent posts we will look at each of the beamforming types. Next time, we will look at static beamforming.

Demystifying radio Management (free)

Paul DeBeasi — Fri, 03 Jul 2009 00:23:01 +0000

My new report entitled “Demystifying Radio Management“ is now published. Thank you to everyone who provided feedback on my drafts. Fortunately, Burton Group has decided to make this report available for free. Below is a summary and also a link to the free report. As always, comments are welcome.

Summary:Wireless local area network (WLAN) vendors have developed a variety of solutions to deal with the challenge of controlling WLAN radio signals. They include: WLAN architecture, beamforming, and air Time fairness (ATF). Single channel architecture (e.g., Meru Networks) has several important technical advantages over multiple channel architecture (e.g., Cisco Systems or Aruba Networks), but multiple channel architecture (MCA) is deployed in over 95% of the enterprise market. Transmit beamforming will become the most dominant form of beamforming because the IEEE 802.11 committee will standardize this approach. ATF will be widely deployed because it helps enterprises deliver predictable wireless performance.

Click to read Demystifying Radio Management

The new battlefront: RF management

Paul DeBeasi — Wed, 28 Jan 2009 16:15:03 +0000

Directional antenna pattern

Wireless LAN (WLAN) Radio Frequency (RF) management is poised to become the new competitive battlefront. This includes technologies such as beamforming, smart antennas, and any other techniques used to control the wireless LAN physical layer. One might think that 802.11 technology innovation is slowing down, and that future competitive battles will primarily rely on marketing fluff, but not so.

Where we’ve been

It may seem that most of the seemingly difficult WLAN problems have been solved. For example, WiFi Protected Access Two (WPA2) (802.11i-2004) provides strong, interoperable, wireless LAN security. WiFi Multimedia (WMM) (802.11e-2005) provides basic quality of service capabilities and power save mechanisms. And the emerging 802.11 high-throughput standard (802.11n) provides high performance improvements using various multi-input, multi-output (MIMO) configurations.

There are also other important, but lesser known standards such as fast roaming, (802.11r-2008), spectrum & power management (802.11h-2003), and radio resource management (802.11k-2008). So, we’re done with 802.11, right? Well, not exactly, there are still lots of standards under development and new technology hurdles to overcome.

Where we’re going

I think that the key area to keep an eye on is RF management. Companies such as Xirrus and Ruckus Wireless have used beamforming for years as a way to improve wireless performance, predictability, and reliability. Aruba Networks uses Adaptive Radio Management (ARM) and Meru Networks uses VirtualCell/VirtualPort to manage RF. Recently, Cisco Systems announced M-Drive and ClientLink technologies aimed at RF management.

Unfortunately, the technology is complex, difficult to understand, and filled with vendor jargon. It seems to me that this is a perfect setup for pseudo-techno marketing spin. The poor IT manager must learn the technology and then figure out how to compare the various vendor solutions. So over the next 4-5 weeks I am going to do a deep dive on this topic. My goal is to answer questions such as:

What are the important components of RF management?
Why is this topic important to the enterprise?
How does one compare the various vendor solutions?
Which vendor is in the best position to provide RF management innovation?
What does the future hold for RF management?

This research will form the basis of my next Burton Group report. In addition, I will likely write several detailed blog posts and articles on this topic.

How you can help

As always, I began my research by casting a fairly wide net in the hopes of quickly and thoroughly learning the topic. Naturally, I’ll speak to all the major vendors and read all their white papers. I will also rely upon engineering textbooks and technical articles. But probably the most useful source of information is the collective wisdom of practicing network engineers.

So if you are willing to share your RF management experience, or you know someone that might, please contact me (here).