Wi-Fi

802.11n: MIMO really needs smart antennas

I remain very optimistic about 802.11n being a disruptive technology that will alter everyone’s perception of data networks. If smart antenna systems and beam forming are included in the mix, it's a for sure thing.

As Wi-Fi standards go, 802.11n has a lot to live up to, especially after hearing how 802.11n's advertised throughput, security, and reliability will allow Wi-Fi to replace existing wired networks. This means 802.11n's RF technology needs to be rock-solid, just like Ethernet cables, while facing ever-changing environmental conditions.

Initially I felt it was entirely possible. 802.11n's new RF technology was certainly enough to take on all real-world demands, but I'm not so sure now. I'd like to explain, but before doing so I feel it's important to really understand what challenges 802.11n technology must overcome in order to become rock-solid. To begin with, Ethernet bits flow nicely through solid amorphous materials like copper, whereas Wi-Fi bits travel through a variety of media and environments, which can affect the following parameters:

  • Received signal strength is dependent on the distance between the transmitter and receiver. Physical obstructions along the link path that absorb or disperse the RF signal also affect signal strength. Ultimately, received signal strength must exceed the receiver's noise floor by a certain amount; otherwise, the signal cannot be processed.
  • In-band RF interference comes in two flavors. The first flavor is non-802.11 RF capable devices like cordless phones or microwaves, which happen to share the same frequency band as Wi-Fi networks. The second flavor pertains to co-channel and or adjacent channel interference from other Wi-Fi networks. Both types of interference if strong enough will create sufficient RF noise to make it difficult or impossible for the receiver to distinguish between the interference and real traffic.
  • Out of band RF interference is something most people don't think about. This interference emanates from devices that are not normally considered RF transmitters. Any electromagnetic (fluorescent light) or thermal (lightning) radiation has the potential to disrupt the RF link between two Wi-Fi devices.
  • Multipath interference or fading occurs when a RF signal encounters objects on its way to the receiving antenna. These objects could reflect or refract the original RF signal, creating variations that have different timing and phase characteristics. When the original RF signal and variations reach the destination antenna, that receiver usually has a difficult time trying to sort out what's what. I went into more detail about this subject in an article named "Multipath environments and how they affect Wi-Fi propagation."

Many people will argue that the previously mentioned types of interference exist in both wired and wireless networks. I agree, with the exception of multipath interference or fading, which is unique to RF propagation. The TRSFC crew may disagree and bring up the topic of electron or photon barrier activity in a captive medium, but that's another topic. The simple reality is Wi-Fi networks are much more susceptible to interference than wired networks.

The fallout from poor signal quality is the retransmission of digital traffic to meet TCP/IP requirements of error-free data transmission. With sufficient errors, the connected 802.11 devices will renegotiate the transmission rate incrementally until the error count is below a set level, which dominos into lower data throughput and decreased network efficiency. The following chart graphically shows the extent of signal reduction caused by interference. I'd like to thank Ruckus Wireless for use of the chart.

signal-strength-versus-noise-rev.jpg

Pre-802.11n solutions

Prior to 802.11n there were various methods to reduce the affects created by interference. Most helped to a limited extent, and I wrote an article "How to make the best of 802.11 multipath environments" that looks at the different solutions.

Now that we are on the same page as to what a RF signal has to contend with on its way to the receiving antenna, let's proceed to the next topic. 802.11n uses RF technology based on MIMO, antenna diversity, and spatial multiplexing to help deal with the above-mentioned challenges. I'd like to take a few moments to explain the inner-workings of MIMO as a prelude to pointing out why MIMO in and of itself is not the definitive answer.

MIMO: Antenna diversity

Antenna diversity isn't new to Wi-Fi technology. It's just becoming official as part of the 802.11n standard. Wikipedia does a great job of explaining antenna diversity:

"Antenna diversity is especially effective at mitigating multipath situations. This is because multiple antennas afford a receiver several observations of the same signal. Each antenna will experience a different interference environment. Thus, if one antenna is experiencing a deep fade, it is likely that another has a sufficient signal. Collectively such a system can provide a robust link. While this is primarily seen in receiving systems (diversity reception), the analog has also proven valuable for transmitting systems (transmit diversity) as well.

Antenna diversity can be simple as "receive selection combining." Where a multi-antenna device transmits using the same antenna from which it just successfully received digital traffic. Or as complicated as equipment using "maximum ratio combining," which allows multiple RF signals to be sent simultaneously between two proprietary devices. The following graphs from Ruckus Wireless show the difference in signal gain between the two different approaches.

receive-diversity.jpg

MIMO: Spatial multiplexing

If you remember earlier in the article I mentioned that RF signals will be altered as they traverse multipath environments. Well, spatial multiplexing is counting on that, as it's the only way a receiving 802.11n device will be able to distinguish between the different RF signals. The Ruckus Wireless chart below depicting spatial multiplexing helps explain the process. As you can see in the first graph, the signals are similar enough to make it difficult to distinguish the two, whereas the second graph depicts two uncorrelated signals.

spatial-multiplexing.jpg

If everything is working correctly, one 802.11n device using spatial multiplexing will transmit a unique data stream using N (number of antennas) antennas. The receiving 802.11n device with at least N antennas will then receive N unique data streams. Therefore, the link's total throughput capacity is equal to the individual data throughput multiplied by N antennas. If you're interested, I went into more detail about this in the article "802.11n, MIMO, and multipath environments."

MIMO: kind of hit or miss

Now it's easy to see how antenna diversity and spatial multiplexing theoretically improve throughput and the reliability of Wi-Fi networks. My concern is what happens when dealing with real-world environments that are constantly changing. For example, if there isn't enough alteration to a RF signal, the receiver using spatial multiplexing will not be able to distinguish it from the rest. Another example pertains to antenna diversity. What if it's a bad assumption to transmit using the same antenna that worked the best for receiving? Seems to me that too much is left to chance. 802.11n networks need to be more self-determining and less reliant on the RF environment if they are going to compete with wired networks.

Smart antennas and beamforming

It took awhile but with all that background information, we can now tackle smart antenna technology. The term smart antenna in reality is a misnomer as all of the intelligent signal conditioning takes place before the RF signal gets to the appropriate set of antennas. Beamforming is the technology that does all the hard work. The following definition is from a University of Washington website. It's the best explanation of beamforming I've come across. The site even has interactive models to help explain the technology.

"Beamforming is a general signal processing technique used to control the directionality of the reception or transmission of a signal on a transducer array.

Using beamforming you can direct the majority of signal energy you transmit from a group of transducers (like audio speakers or radio antennae) in a chosen angular direction. Or you can calibrate your group of transducers when receiving signals such that you predominantly receive from a chosen angular direction."

Beamforming isn't new, being a key component of both radar and sonar systems for many years. Recently, telco and Wi-Fi researchers have become interested in beamforming and the ability to steer signals to where they do the most good. Ruckus Wireless is one such company and has a great deal of research expertise in beamforming. Ruckus Wireless also has been instrumental in introducing products into the Wi-Fi market that have beamforming capabilities. BeamFlex is their interpretation of beamforming, and the following description comes from one of their technical articles:

"Central to BeamFlex is an agile antenna system with multiple antenna elements that can be combined in real time to offer an exponential increase in diversity order. With N number of high-gain, directional antenna elements, a BeamFlex antenna array provides 2N-1 unique radiating patterns to maxi­mize range and coverage in a home.

A Diversity Combiner composed of low cost, software-controlled circuitry allows the BeamFlex software to manage antenna combining in real time. The core of the BeamFlex software is an expert system that constantly learns the environment - the RF conditions, communicating devices, network performance and application flows.

A Path Control module selects optimum antenna combinations on a per packet basis to ensure a quality signal path to each receiving device.

The Transmission Control module sets the transmission policies including data rate and queuing strategy based on application and station knowledge. The BeamFlex software interfaces to the 802.11 MAC layer and is compatible with standard 802.11 chipsets. Residing in the host processor, it adds minimal incremental CPU load and memory utilization."

In my research on smart antenna systems and beam forming, the Ruckus Wireless approach has surfaced as a very elegant design. It has the potential to alleviate my concerns about the inability of MIMO and spatial multiplexing to be reliable enough. The individual advantages are as follows:

  • BleamFlex antenna arrays can rapidly present many different antenna configurations. Which translates into significantly different RF signal patterns that will afford spatial multiplexing technology the best opportunity of success.
  • BeamFlex antenna arrays use both horizontal and vertical polarized antenna elements, once again, to create RF signal patterns with increased diversity and ensure recognition by the 802.11n receiver using spatial multiplexing.
  • BeamFlex architecture uses application-level performance parameters when making decisions on how to optimize the signal quality rather than information from the PHY and MAC layer that doesn't take into account QoS or application networking requirements.

The following diagram depicts current equipment from Ruckus Wireless, which include all of the above-mentioned features.

beamflex1.jpg

I'm more interested in a symbiotic relationship between the BeamFlex antenna and 802.11n technology so as to have the best of both worlds. Ruckus is continuing work on this front as shown in the following diagram.

beamflex2.jpg

Final thoughts

I remain very optimistic about 802.11n being a disruptive technology that will alter everyone's perception of data networks. 802.11n's antenna diversity and spatial multiplexing are vast improvements over what's been available in previous standards. I'm just concerned that the required reliability will not be there until additional RF signal conditioning like that offered by Ruckus Wireless is used to combat environmental variables.

About

Information is my field...Writing is my passion...Coupling the two is my mission.

33 comments
engsat
engsat

Research topic needed in smart antenna area : Dear sir! I am preparing research topic in the area of smart antenna system for future wireless system. Kindly guide me some area that i explore in my PhD work Best regards

brian.collins
brian.collins

MIMO is just that. It IS a smart antenna technology. In this digital age smart antennas really are dumb antennas plus signal processing. MIMO works by exploiting lots of altenative routes for the signal between the transmitter and the receiver IN A COMPLEX REFLECTION-RICH ENVIRONMENT. In effect it transmits different parts of the data stream through these alternative spatial paths and puts the whole transmission together at the receiver. There are various choices for the antenna elements, but this is not what was understood as 'beam forming'. It's a very different technology! Brian Collins www.bscassociates.co.uk

jdiamond
jdiamond

A question really - if there are multiple positions receiving signals and beamforming is being used, are different beams created for thes different positions? Does the transmitter actually keep track of where the targets are and is it constantly adjusting this because targets could move around or even "objects" could be moved to cause interference (or microwave ovens switched on etc.).

alegentile
alegentile

Hi, I found some improvement with the "N", but not much. Infact, the Belkin "N1" is worst then the "N". Thanks.

JCitizen
JCitizen

Hey, Michael! Now that CBS is going to be running the show; are we going to lose quality reporting like this? I shudder to think!

Michael Kassner
Michael Kassner

Will the addition of beamforming technology be enough to push 802.11n over the tipping point in the battle against wired network devices?

JCitizen
JCitizen

I've always suspected antenna technology was still in it's infancy, if one were to gage it by the future improvements possible. Your web site is the bomb. I will have to visit it more often when I have time. (edited) to add message

Michael Kassner
Michael Kassner

MIMO's spatial multiplexing as defined by 802.11n Draft 2.0 only uses passive techniques to derive multiple data streams. That is why spatial multiplexing will not work in a non-multipath environment. MIMO's spatial multiplexing would be useless in a completely void environment such as empty space. Beamforming is an attempt at techniques that will enhance the individual waveform differences so spatial multiplexing works better.

Michael Kassner
Michael Kassner

All beamforming technology that I am aware of is proprietary, thus I have to rely on their information and test results. The developers appear to consider the link between the receiver and transmitter as an ever-evolving symbiotic relationship. Using data-rate stepping and RSSI information the beamforming technology will adjust what antennas or combinations of antennas are used to deepen the differences in RF signals. Which allow easier differentiation at the receiver.

engsat
engsat

Dear sir! I am preparing research topic in the area of smart antenna system for future wireless system. Kindly guide me some area that i explore in my PhD work Best regards

Michael Kassner
Michael Kassner

I have been researching this a bit more and honestly can not tell if it is for real or not. Mr. Mathias is one of my heros, but maybe pulling a fast one on all of us. I'm starting to suspect that it is as Daaahhh I just looked at the date of the article. Good one.

seanferd
seanferd

It's like a giant MASER with receiving capability. Wow. Talk about steering beams. I wonder what kind of magnetic field that thing produces. "Now we can transmit from the center of a black hole." -Daily Balthorism.

JCitizen
JCitizen

What was up with that? The article didn't go far enough to explain just what that hugh contraption was in the picture!?

JCitizen
JCitizen

Sorry Michael, but that name almost set me back. I had to reread the article to make sure I wasn't being scammed! I love ze Schermin scientists though! I haf allvays looked up to ze mad scientists of yor. Seriously though, what did you think of this article when it came out: http://blogs.techrepublic.com.com/hiner/?p=694&tag=nl.e019 I would think tech like Xirrus could be part of this area of discussion.

Timbo Zimbabwe
Timbo Zimbabwe

... but for home networks, this is good tech. I switched to 802.11n with MIMO almost a year ago. I get good data transfer speeds and have yet to have a connection issue between my PC and the router. We have 3 or 4 cordless phones and a microwave that gets used quite a bit. I never seem to lose a beat.

robsimkins
robsimkins

Anyone else feel that regardless of any killer modulation improvements like beamforming.. the WLAN vendor community would go mental if the spec is changed to include it? Yet more time added to the ratification process at the eleventh hour? and even more so if the change involved anything physical (like antenna elements) - think of the cost implications!? Remember that the IEEE/Wi-Fi Alliance are constantly under scrutiny and have been questioned in the past of their capability to define WLAN standards in a timely manner - they're under pressure to get this standard out within the next few months and that means beamforming isn't high on their agenda. Although perhaps it should be? As an integrator I see 802.11n for the majority of enterprises being up to 3 years away. While they upgrade their LAN to support 802.11n (Gig Ethernet and PoE+ at the edge, 10G Ethernet at the core?), they will also need to draft in 802.11n adaptors to their end devices (I know very few organisations that are buying 802.11n capable devices across the board right now (with a 2/3 year lifespan expected). On the other hand - 802.11n draft 2 has performed consistently in testing and pushes reliability and performance out to the edge of the cell, interference shouldn't be an issue as 802.11an will be de-facto for 300mbps - as 802.11gn will have to be deployed at 150mbps to allow 3 channels to be used (300mbps allows just one non-overlapping @ 40MHz..) In summary! * 802.11n will be ratified without beamforming and probably do a great job. * 802.11an will prevail over 802.11gn * Gigabit Ethernet LAN will be a catalyst for its success Plus, hopefully we see lots of price reductions for 802.11n silicon so that it puts 802.11abg to bed.

seanferd
seanferd

Very interesting article. As far as I know, there is no type of interference in a wired system that is an analog of what radiative systems have to deal with. I certainly wouldn't bundle an ethernet cable with a bunch of power conductors though. That's my TRSFC-oriented input. :)

JCitizen
JCitizen

haven't kept up with radio wave technology can get easily confused. When and if you get a chance to answer my question on one of the previous articles I will feel caught up on things a little better.

MWRMWR
MWRMWR

...so tell me the joke. I tried variations on 051408 05142008 140508 20080514, put in / and -, Googgled ... all to no avail. It can't be "May the fourthteenth be with you ". Is it an RFC or some ISO or other standard number ? Is it a frequency ? Guessing really Herz my brain so please tell.

seanferd
seanferd

When shipping atoms becomes like shipping bits.

Michael Kassner
Michael Kassner

I would appreciate hearing your reasons why you feel that this technology would not be a good fit for enterprise customers.

Michael Kassner
Michael Kassner

Hello Rob, How have you been? You have made some great points. The pressure to keep 2.4GHGz alive will force many developers to user proprietary means to avoid the frequency issue. Companies like Meru already have it figured out really. Their technology allows the use of one channel throughout the entire network. I have deployed their system once and it works quite well, actually. http://www.merunetworks.com/pdf/whitepapers/Virtual_Cells_WP4_0705.pdf I suspect that along with beam forming to be thorns that force the use of proprietary equipment. 802.11a is great, but as we all know the coverage area is significantly less and does have an affect on number of required APs and thus overall cost.

Michael Kassner
Michael Kassner

I thought you guys would get all over the multipath interference and electron or photon barrier activity in a captive medium comment. I am curious as to learn if that would actually be considered multipath interference as the photons/electrons bounce off of the exterior surfaces.

Michael Kassner
Michael Kassner

I hope I was of some help. If not just let me know, I will find someone who does know.

JCitizen
JCitizen

I'll let you know...Somehow I lost the original article but it will pop up again I'm sure. If it was the circuit question you got me straight for sure!

Michael Kassner
Michael Kassner

If I did not explain it correctly, please let me know. Beamforming is pretty amazing technology, almost to the point of adding AI to RF emitters.

seanferd
seanferd

Multipath interference and electron or photon barrier activity. I don't know what, for the purposes of the transmission of information, the negative aspects of barrier activity might be. Is there unwanted reflection causing interference in conductors that alters signal to noise ratio on the scale of random noise? The flow of electrons in a conductor is a bit like fluid mechanics, but I don't know how this might affect a signal other than speed. As to EM propagation in a wire, I really have no idea what effects might be important. I would expect that it is much more problematic with radio in that the medium is far from homogeneous. Radio doesn't benefit from the more useful aspects of a barrier. It doesn't have a waveguide or protection from external interference. Optically, I suppose you plan for the frequencies that can transmit information the fastest that propagate best in a chosen medium. Fiber diameter and the coating I would assume are chosen to get the best internal refraction [insert plug for research into optical metamaterials with negative indices of refraction here] and the medium needs to be free of defects (barriers!). I suppose it all boils down to scale, aside from the non-homogeneity of the medium. How does a radiant signal compare with a captive medium signal for wavelength, distance to reflection (for that matter, possible distance to reflection), and the timing of reception of multiple reflections of the signal? Hmm. Perhaps some sort of analogy between multipath interference and LASER interferometry...

JCitizen
JCitizen

that especially in reference to Neon's basement project, I would think reflection is the most under used facet in the interference category. Where the idea that you can use interference to actually enhance your performance.. For instance I use a metal hutch to reflect radio waves from my AP downward toward one of my work areas where I do a lot of laptop work. Just one example*

Editor's Picks