The King is Dead, Long Live the King: 802.11n dramatically improves Wi-Fi outdoors

IEEE 802.11n is the new international standard for wireless Local Area Networks, incorporating new smart antenna technologies (MIMO – Multiple In and Multiple Out) permitting a 5x performance and 2x coverage improvement for WLANs. While this new technology is becoming the de facto standard in consumer and enterprise networks, it has not yet made an appearance in outdoor, metropolitan scale networks derived from Wi-Fi technology. (Note: Meraki just released today the first 802.11n outdoor mesh product)

Many of these same MIMO techniques are being incorporated in both WiMax and in LTE for cellular networks. Sadly, neither is being produced in much volume and fixed WiMax networks do not incorporate MIMO technology.

There has been much dispute about whether the specifics of 802.11n designed for indoor networks would apply to outdoor networks and bring the economy of scale of 802.11 to outdoor applications. At Novarum, we tested the effects of 802.11n on outdoor performance. We found dramatic improvements in using indoor 802.11n technology outdoors, so much so that 802.11n has become, for us, the recommended baseline for new network deployments.

First, let’s review the key pieces of technology incorporated in 802.11n and how it might affect outdoor performance.

In the course of Novarum’s Wireless Broadband Review in 2007 and 2008, we examined over 25 deployed Wi-Fi networks (including all major vendors), 46 deployed 3G cellular networks and 4 fixed pre-WiMax networks. In the case of Wi-Fi networks, we noted the dramatic effects that client selection had on network performance, coverage, and ultimately, user satisfaction.

We both examined 802.11n clients against the installed multi-vendor base of 802.11g infrastructure and constructed our own testbed from early outdoor 802.11n components to evaluate the effect of 802.11n when deployed in the infrastructure itself.

It is important to recognize that in almost all outdoor Wi-Fi networks, the client access uplink is the weakest link in the communication chain. Legacy 802.11b/g clients experience VERY high packet retry rates of between 100% and 1000%, and there are often deep multi-path fades of between 10-30 dB within a few tens of feet. The Wi-Fi protocol is VERY good at masking these effects – instead they are most commonly seen indirectly – by lower throughput and higher delay variance. These effects are seen even for deployments of very high access node density of 50 nodes per square mile or more.

These deep fades and very high packet retry rates made mobility difficult, dramatically affecting throughput, and causing packet delay variance to be so high as to make streaming media difficult, thereby materially decreasing the overall capacity of these networks.

The improvements that 802.11n provides outdoor networks astonished us – particularly for a technology that has been disparaged as inappropriate for outdoor deployment. Deploying IEEE 802.11n technology has dramatic effects outdoors – both with legacy systems and even more compellingly with green field deployments.

Let’s summarize what we found in Novarum’s experiments:

  • 100% throughput improvement of 802.11n Wi-Fi clients with legacy 802.11g outdoor infrastructure;
  • 100% throughput improvement of legacy 802.11g Wi-Fi clients with new 802.11n outdoor infrastructure;
  • 200% throughput improvement of 802.11n client with 802.11n outdoor infrastructure;
  • Similar coverage of 802.11n clients and infrastructure in the 5.4 GHz band as for legacy 802.11g clients and infrastructure in the 2.4 GHz band – making the 5.4 GHz band useful for client access;
  • 25% decrease in access latency and a dramatic improvement in latency variance;
  • a low power 802.11n client has the same throughput and coverage as a high power 802.11g with 10x the power and antenna; and
  • coverage to smartphones at low power and with poor antennas dramatically improves.

These results have a dramatic impact on outdoor wireless networks, bringing the benefits of MIMO technology at consumer price-points.

These are the improvements we expect in outdoor networks that use 802.11n technology:

  • 400% to 800% increase in system capacity and throughput;
  • 200% to 300% improvement in spectral efficiency through increased link budgets, reduced packet errors, increased modulation rates and improved fading performance;
  • effective client access to the 200 MHz of the 5.4 GHz band;
  • 802.11n clients dramatically improve legacy 802.11g networks and new 802.11n networks dramatically improve legacy 802.11g clients;
  • Streaming media applications will perform as we expect and will be much easier to deploy;
  • Better backbone designs by reducing the interference of the backbone mesh through beam-forming antennas rather than omnidirectional broadcast;
  • Decreased deployment cost due to decreased node cost, possibly dramatically.

While not optimally designed for outdoors, 802.11n will substantially increase the performance of and customer satisfaction with outdoor wireless networks.

We can expect the first product announcements of outdoor Wi-Fi networks incorporating 802.11n shortly. We also expect all major vendors of outdoor Wi-Fi equipment to ship 802.11n products by the end of 2009.

Novarum recommends that all new outdoor Wi-Fi networks use 802.11n products in their infrastructure and 802.11n clients wherever possible.

Related articles:

Real world measurements show muni Wi-Fi networks outperform WiMAX and cellular

——————————————————————————————–

About the author

Ken Biba is a co-Founder and Chief Technical Officer of Novarum, an advisory firm specializing in wireless data networks. Ken has over 30 years experience in the network information systems industry combining a background of general management with a strong product and marketing focus in network systems and information security. Ken was an early engineer of the Internet in 1975. He has co-founded and managed four notable networking companies-Sytek, which was focused on cable TV-based local and metropolitan data networks, Agilis which delivered the first wireless handheld computers, Xircom, which pioneered local area networks for mobile computing, and Vivato, which was focused on scaling Wi-Fi infrastructure to cover campuses and metropolitan areas. Ken’s perspective as CEO, board member of public and private companies, and as a technologist brings unique insight to the business, market and technology of bringing useful wireless solutions to users. Ken has a Bachelor of Science in Physics (Magna Cum Laude, Tau Beta Pi) and a Master of Science in Computer Science from Case Western Reserve University. This article is posted on the Novarum blog.

- – - – - – - -

*Sign up now to be included in the Muniwireless 2009 Vendor Directory. Advertise on Muniwireless.com. Download our 2009 Media Kit and contact us.

*Join the MuniWireless / Muni WiMAX Linked In Group for networking and discussions.

*Sign up for the MuniWireless / WiMAX newsletter.

*Register now for the Muniwireless – Freedom to Connect Event, March 30-31, 2009 Washington DC

Comments

  1. Is this equipment-like a USB add-on? It wuld be kinda cool to have wireless portable internet service anywhere you go, provided you have a wi-fi setup on your computer.

  2. No, 802.11n extends the 802.11 specification for Wi-Fi equipment such as wireless access points. Having Wi-Fi access in most places is the whole point behind muni WiFi and you don’t need a USB dongle, as you do with WiMAX. Laptops and many smartphones already have Wi-Fi built in.

  3. Brett Glass says:

    As a wireless ISP, I would be extremely concerned about any attempt to use 802.11n outdoors. First of all, it’s a spectrum hog. A single 802.11n access point uses two thirds of the available Wi-Fi spectrum at 2.4 GHz. This means that two 802.11n operators could mot use the band simultaneously without interfering. (In fact, a single operator cannot have two 802.11n access points in sight of one another without SELF-interfering.) This makes frequency coordination impossible, and creates a “mutually assured destruction” situation in the spectrum commons. Users CANNOT steer clear of one another by agreeing to use different channels. This will not only negate any benefits that MIMO might bring; it will also interfere extremely destructively with 802.11g and 802.11b. (There’s a reason why 802.11n is called the “Wi-Fi Killer.”)

    Secondly, the creation of high gain antenna arrays for customer premise equipment would be unwieldy and prohibitively expensive. Good wireless broadband requires antenna gains of 15 dBi and higher; a “triple” 15 dBi antenna would be extremely costly. And the omnidirectional or wide beam sectoral antennas that would be required at the access point would ensure that the 802.11n equipment’s attempts to take over most of the frequency band would impact other users greatly. Also, at great distances (several miles), any antenna array at one end of the link “looks” to the radio at the other end like a point source, obviating many of the useful effects of having an array. Yes, there would still be some “beam steering,” but not as much as could be achieved more simply and effectively with a sectoral antenna.

    Finally, due to the FCC rules, the power of MIMO equipment is more severely limited than that of equipment that uses narrow swaths of spectrum. This matters a lot at long distances, where path loss (and not problems due to obstacles) is the dominant factor in signal quality. The FCC rules do allow a small amount of extra power for beam forming antennas. However, an array of omnidirectional antennas (as is normally used for 802.11n) would create a signal pattern with multiple peaks in multiple directions and so would likely not qualify for this exemption (or shouldn’t, at any rate, because of the interference this would cause). And you couldn’t co-locate arrays of more directional (say, 180 degree sectoral) antennas near one another, because you would get cross-interference between the two access points due to near field radiation and spectrum overlap.

    For all of these reasons, 802.11n is not a good choice for real world outdoor applications, such as rural broadband. Municipal networks, especially, should be good “spectrum citizens” and not monopolize the limited Part 15 spectrum, which is already being used by WISPs, public institutions, and private companies. 802.11n may be OK indoors — where the walls will attenuate it before it pollutes the spectrum for everyone else — but should not be used outdoors.

  4. Brett:

    Sadly, I completely disagree. And the data from field experiments validates my position – which was the point of my post.

    First, 802.11n can be deployed in either single 20 Mhz channels (just like current 802.11) or in two bonded 20 MHz channels. In the former mode 802.11n nodes cause no more issue with channel planning than current 802.11 products.

    Second, because of the greater link budget from beam steering, spatial diversity and MRC, 802.11n MIMO products can cover the same area with LOWER total power and/or improved bit error performance and hence high modulation rates and thus higher throughput.

    Third, I urge you to see the wave of new outdoor 802.11n products to see the innovative antenna solutions packaged for both performance and low cost.

    Fourth, 802.11n has really two complementary applications outdoors: for client access and for long distance backhaul.

    Our field experience is utterly conclusive that 802.11n will improve client performance on access with LESS power and less interference – by taking advantage of the multipath that exists on the client access side.

    On the backhaul side – for long distance links with no multipath – 802.11n frankly adds minimal value – these simple links are largely best served as they are today.

    However, for more complex mesh backhaul systems as we see in urban areas – with inherent multipath and a complex radio environments – 802.11n will DECREASE interference and INCREASE performance while staying within FCC limits on power. There is actually substantial innovative potential on the backhaul design to take advantage of 802.11n to minimize power and increase performance.

    802.11n is entirely legal and appropriate to be used outdoors. And like its use indoors, will dramatically increase performance but will challenge the conventional wisdom of last generation deployment models.

  5. One additional example.

    In our testing of over 25 deployed WiFi muni networks we tested a variety of clients. Our results are documented in the Novarum blog at http://www.novarum.com/novarum_blog/2008/11/.

    In particular, we tested a “high power” 200 mW Ubiquiti PCMCIA mobile client (with roof mounted diversity antennas) against a USB attached early 802.11n 30 mW client. When tested in the same locations, at the same time, against multiple access points by different vendors – both of these clients performed almost identically.

    With lower total TX power and hence lower contributed interference, the 802.11n client delivers the same throughput as a client with 8x the amount to transmitted power.

    802.11n wins conclusively in urban environments where we can take advantage of multipath.

    In unobstructed, non-multipath environments – 802.11n and WiMax MIMO for that matter, will have only incremental advantage.

  6. Ken, the basic problem that I personally see with outdoor wireless is not the difference between 5Mbps or 50Mbps, it’s cost per square mile verus revenue. Yes, some cities can afford whatever gets put up regardless of the cost (although those cities are now fewer and farther between), but we aren’t cracking the majority of the cities any longer. Even $1500 per square mile plus $200 for installation costs is way out of the range of most cities. In Meraki’s case, you then also have to use their authentication system if I’m not mistaken.

    We need to start by deciding what’s most important and that’s reducing costs and ROI. If it’s a city only system, cost is important. If it’s an open access WiFi system, then cost comes first followed by a planned upgrade if additional capacity is needed. Nobody cares if it’s mesh or not and the reality is that radio failure is very few and far between anyway. I say build it at the least expensive cost possible, see how it is used, see what revenue or savings it can generate, then start expanding it. I haven’t seen any system that has ever had a problem of bandwidth shortage. I do agree on your analysis if interference and multipathing. However, I think that the community as a whole needs a kickstart in the rear to get municipalities back on board.

    If you go back to them and say $3K-$9K per square mile or some investors feel that this price point starts making financial sense, we have a winner. If we keep trying to push the $100K or more per square mile option, I feel that this market is going to die and we are all wasting our time. We have a good idea, we just need to get the cities who now have very little money to spend, to get back on board.

  7. Rory,

    Meraki’s new 802.11n outdoor mesh product dramatically lowers the cost of deployment. Read my article: http://www.muniwireless.com/2009/02/24/meraki-releases-80211n-mesh-product/

    They are selling their 802.11n mesh nodes for $1500 list price. That price is about half to less than half of what the other vendors are offering.

  8. Brett Glass says:

    802.11n is generally run on 40 MHz channels. A few systems do offer an option for narrower channels, but this is not the default.

    As for “beam steering, spatial diversity and MRC:” As you yourself note, Ken, none of these work at long distances. But the interference from 802.11n systems can propagate for miles, harming efforts to deploy broadband to unserved areas. And so do mesh networks, which can make spectrum totally unusable by other systems.

    Municipal wireless networks are, to put it simply, a bad idea. They waste spectrum, have poor coverage (especially inside buildings), and at the same time harm privately deployed systems. Greater interference from 802.11n would make matters even worse. The government should not compete with private business — and especially should not DISABLE private businesses by hogging the spectrum. Private enterprise can do the job better. We are, and we’re proud of it.

  9. I have seen what 802.11n does a working Wi-Fi hotspot.

    If want to see for your put three off the shelf N AP’s within a few hundred feet of each other. (Real world stuff not some controlled lab test with 200 Mw card)
    Now fire up your Wi-Fi sniffer, see all the noise??
    Now shut down the N AP’s, Gee what happened to the noise?

    As we can not control “the other guy” we must control ourselves.

    I use the older Meraki units, they work great. I would never try to cover any amount of space without Mesh networking, it just makes sense to me to use mesh.
    In fact I am consulting on a RV park setup at this time, they are deploying 16 plus Meraki single radio units over 7 acres, the cost of the gear plus mounting gear is less than $ 6000.00 total for park owner.
    We have two 8 meg down 2 meg up cable modem connections to share between 30 or so concurrent connections at peak times. The only reason I am going with such a high AP density is the park has had three other wireless networks which have all failed to provide coverage inside the RV’s across the whole park. The owner wants 100 % coverage in his park, I figured I may bake a few brains but they will have internet as they get nuked.

    If I used the lowest cost N AP from Meraki it would have increased the job cost 1/3 or more and as most users have b/g cards and even the ones with N cards will default to g/b at 30 feet or so, we would spend more and have less.

    Now remember in most cases the network bottle neck is the Internet connection not the traffic between client and AP’s (well not in well designed network anyway). One of the first things I do on any “public” network is to choke the end user bandwidth down to 1024/256 so more users can have better internet, so why do I care about the speed to the end user to the backbone?? If they are getting more than the allowed Inet pipe it does them no good.

    The Wi-Fi world is at a turning point, within two years we will see a new set of 4G AWS / 700 MHz / Wi-Max client devices built in most laptops / desktops.
    I tested the older single radio Meraki mesh devices for over a year before I felt we could deploy them in the field, I will hold out until I can test one of the 4G systems.

  10. So many points – let me try to review a few:

    1. In a multipath environment – 802.11n APs will deliver superior coverage and performance to legacy .11g clients outdoors. They will service more types of client cards over a greater area. This is simple test results we and other folks have done.

    2. In many applications, performance is simply not the issue. In which case cost of coverage is dominant issue. Leaving aside the issue of particular vendors for the moment – when we can drop the cost of a 2 radio mesh to $500/radio (as Meraki has done) AND use 802.11n for improved fade resistance to improve the quality of coverage to legacy clients – this is substantially less than any other commercial mesh vendor and sets a new standard for the cost of coverage. Our tests of 802.11n AP clearly demonstrates this.

    And frankly, since penetrating fiberglass RV shells is a multipath environment, if effective “indoor” coverage is a design goal – 802.11n will in fact improve that coverage rather dramatically. The 30% improvement in cost might be a minimal investment over a fourth future wireless network.

    While not always compelling, the ability of .11n to increase coverage also means decreasing the number of hops in a mesh. And since mesh performance decreases exponentially with the number of hops – even 2 or 3 hops will make a cable connection look like dial-up and be effectively unusable.

    3. 4G/LTE/700 MHz/WiMax … are all interesting possibilities but all have their own challenges. When deployed on licensed bands we will see DSL-like performance at 3x DSL prices, and there are only few frequencies available for un or lightly licensed deployment of WiMax devices. The price points for these “private” WiMax systems argue for applications with wide area coverage with a modest user load. WiFi will likely always be the appropriate choice for enterprises, campuses, MBU and in many cases municipalities.

    4. Individual vendor products always reserve the right to do it wrong.

  11. Okay, my turn on this:

    1) No argument. However the gain in performance will only be achieved when there are no other 802.11b/g anything around that create interference and everyone has 802.11n on their laptops. That isn’t going to happen for many, many years. For good or bad, we are stuck with 802.11b/g and that’s what we need to make work right now.

    2) I agree that Meraki’s equipment is inexpensive as long as you use their authentication model. This may have changed but I haven’t followed them for a while. However, nobody cares if it’s mesh and we have a 2 radio model that costs less than half that with a signficantly larger coverage area that compensates for fade resistance with better RF engineering and overlap design.
    3) I doubt that the new technologies will be successful in the short term since very few cities are purchasing anything anyway.

    As Bubba said, 802.11n hasn’t shown it can play nice with others yet. It was supposed to be compatible but instead it’s a random noise generator that basically escalates the battle in terms of interference. If I’m in IT and my 802.11 network goes down in my office becase some 802.11n AP was installed outside my window, I would probably make sure that AP never sees another radio it can connect to until it came off the pole.

    Until the world all goes 802.11n, it’s going to be a red herring and with 802.11a/b/g equipment coming down to less than $100, it’s difficult to cost justify 802.11n just yet.

  12. There is simply little evidence for 802.11n not playing nice.

    We MEASURED 802.11n clients and APs IN THE WILD – in the presence of other .11g products and as compared to other .11g products.

    Put a traffic sniffer on any outdoor .11g products and what you will often see is HUGE link level retry rates. I have seen 1000% retry rates on client uplinks for .11b/g products AND 150% retry rates for downlinks.

    When using .11n, these retry rates drop dramatically. And further, .11n units seen to do a better job in deciding the right modulation rate that legacy .11b/g units.

    The effect of .11n then .. is to dramatically DECREASE interference by more efficient use of the airwaves.

    This is simply not speculation and opinion .. this is MEASUREMENT.

  13. Ken, I’m not arguing the point that you make. However, I would like to see the equipment/antenna combination used. For example, I know that most AP’s are using 26dBm – 30 30dBm radios with 6-9dBm omni directional antennas. I argue that is not the best general design combination for municipal deployments and have also proven that in the field. Therefore comparing a new technology against what I consider a weak design to begin with creates an incorrect financial calculation. I believe that a simpler solution and much more cost effect option is available and have deployed it successfully. 802.11g is capable of supporting multi-pathing successfully, granted, not as good as 802.11n, but the cost of deployment more than offsets that differential.

  14. I have measured over 26 deployed .11g WiFi muni networks … let me assure you … they almost universally offer poor service with huge fading problems.

    Almost ALL deployed WiFi networks do not even use diversity!

    What we will find is that the new generation of .11n muni capable APs will cost no more and likely less than the current generation of less than capable .11g APs.

    .11n will not only INCREASE the quality of the coverage but DECREASE the cost.

  15. Actually Ken, you actually supported my argument. I believe that all of those deployments were done exactly like I described in my earlier email. This creates the exact problem you are describing.

    As for diversity, there are several implementations of it and not everybody was using a true diversity design. That was the same with the use of the term “mesh”. Too many different implementations and feature differences. Depending on the type and processor overheard, diversity simply makes up for a lower quality antenna design. 8″ apart doesn’t really make a huge difference in the signal quality unless the antenna has flat spots in radiation pattern that were signficant. Again, this supports the position I took earlier. $2000-$3000 for the radio, $30 for the antenna. Kind of like putting bicycle tires on a Drag Race Car. You can make all 800 horsepower but if you can’t get it to the ground to convert it into forward momentum, it’s sort of useless.

    There is no doubt that 802.11n is better than 802.11b/g technically. However, your statement about decreasing the cost is completely dependent on the premise that a $1500 802.11n will allow a reduced number of AP’s compared to standard 802.11g. I still point out that you can deploy 3-5 802.11b/g AP’s for every 802.11n AP. Of course, I don’t have all the rules and costs your big cities in the East have which the cost of the AP superfluous as a percentage of the total package, but just at the unit cost, will an 802.11n AP cover the same area as 3 802.11b/g AP’s for the same omni-directional antenna gain?

  16. Actually Ken, you supported my argument. I believe that all of those deployments were done like I described in my earlier email. This creates the exact problem you are describing.

    As for diversity, there are several implementations of it and not everybody was using a true diversity design. That was the same with the use of the term “mesh”. Too many different implementations and feature differences. Depending on the type and processor overheard, diversity simply makes up for a lower quality antenna design. 8? apart doesn’t really make a huge difference in the signal quality unless the antenna has flat spots in radiation pattern that were signficant. Again, this supports the position I took earlier. $2000-$3000 for the radio, $30 for the antenna. Kind of like putting bicycle tires on a Drag Race Car. You can make all 800 horsepower but if you can’t get it to the ground to convert it into forward momentum, it’s sort of useless.

    There is no doubt that 802.11n is better than 802.11b/g technically. However, your statement about decreasing the cost is completely dependent on the premise that a $1500 802.11n will allow a reduced number of AP’s compared to standard 802.11g. I still point out that you can deploy 3-5 802.11b/g AP’s for every 802.11n AP. Of course, I don’t have all the rules and costs your big cities in the East have which the cost of the AP superfluous as a percentage of the total package, but just at the unit cost, will an 802.11n AP cover the same area as 3 802.11b/g AP’s for the same omni-directional antenna gain?