Wi-Fi

What is the big deal about 802.11n?


802.11n is being touted as the networking standard that will unwire the network world. Most people are not convinced, yet curious to see how one defends such a claim. The 802.11n standard has two top-level mandates (unwiring the world not being one of them): Achieve higher data rates and retain backward compatibility with legacy 802.11a/b/g devices.

The significance of requiring backward compatibility shouldn't be taken lightly since it complicates the development process exponentially. Yet the 802.11n group has come up with a soon-to-be-ratified standard that meets both objectives.

In a previous post: 802.11n, MIMO, and multipath environments I discussed MIMO smart antenna technology, which is by far the most visible improvement being leveraged by 802.11n. In this post, I will touch on how MIMO and other 802.11n technological advances dramatically improve data rates, system stability and reliability. Maybe even convince some that it is a viable alternative to a wired connection/network.

802.11n's improved technology

802.11n's version of OFDM: 802.11a/g already uses OFDM (Orthogonal Frequency-Division Multiplexing) to achieve data rates of 54 Mbps. 802.11n OFDM technology builds on 802.11a/g OFDM modulation by creating support for multiple channels (MIMO), allowing more bandwidth per channel, and higher code rates. This brings the maximum data rate of a single 802.11n OFDM channel to 65 Mbps. MIMO antenna systems: The 802.11n standard allows up to 4 MIMO transmit/receive antenna pairings. 802.11n OFDM has a maximum data rate of 65 Mbps, multiplying that by the 4 MIMO antenna channels raises the data rate to 260 Mbps, which is a significant improvement when compared to 54 Mbps. 40 MHz channels: To further improve data rates, 802.11n allows the use of 40 MHz channels-twice the existing 20MHz channels used by 802.11a/b/g-which effectively doubles the data rate to over 500 Mbps. 40 MHz channel size is also the most controversial tenet of the new standard, having the potential to disrupt existing 802.11a/b/g networks due to co-channel interference. Aggregation: Aggregation is an important feature developed to overcome shortcomings of having to be backward compatible with 802.11a/b/g networks. It improves mixed-mode performance and efficiency by bundling several frames together that are destined for 802.11n devices, while still being able to transmit single data frames to legacy devices. RIFS (Reduced Inter-Frame Spacing): RIFS is a required 802.11n feature that also improves performance by reducing the amount of dead time required between OFDM transmissions. It should be noted that this feature is restricted to greenfield deployments.

Consumer and enterprise interest

With pre-release equipment already for sale and claims of equipment being firmware upgradeable to the ratified 802.11n standard, is it time to switch? What does this mean to consumers and businesses that may want to start using 802.11n equipment?

Consumer interest: 802.11n promises many things, greater bandwidth, better range, efficiency and reliability. These promises are starting to be tested as pre-release equipment is being integrated into many households. The general consensus is noticeable improvement, especially when considering coverage area, throughput, and reliability. There is also a movement to eliminate mixed mode networks as soon as possible to take full advantage of 802.11n technology. Enterprise interest: Enterprise use of 802.11n will take longer to integrate because IT departments have to be more concerned about everything working all the time and being cost effective. The introduction of 802.11n equipment by the major wireless networking houses has given the standard an unequivocal stamp of approval, even without final ratification. Network designers and administrators are already looking at several unique uses for 802.11n equipment. With data rates exceeding some wired Ethernet networks, using 802.11n equipment for a backbone link that's difficult to cable may now become a consideration. Increased bandwidth and larger coverage area per device may make it economically feasible to unwire the office or install a wireless network in the new addition.

Conclusion

802.11n is indeed a significant improvement, just like 802.11g when compared to 802.11b. All the theoretical data looks good, but with any new technology there is a learning curve. 802.11n technology is complicated, which means problems are likely to be complicated as well. Consumers and businesses alike will have to decide where along the accumulative learning timeline they feel most comfortable about moving to 802.11n.

About

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

44 comments
tagullett
tagullett

OK. What am I missing? I fail to see the attraction for the 'n' technology when most users (home/small business) don't have connectivity above 8 Mbps. I have a feeling we will all be forced into it because the manufacturers will just stop making 'g' routers.

dgs010243
dgs010243

It is the natural technological trend, independent on the inherent bootlenecks. It promise a superior channel throughput. I have many links related to this topic on my wireless, wimax ... blogroll: http://dansomnea.tripod.com/diary

emetesh
emetesh

Any examples of n being used in long distance Line-of-Site Wireless WANs? Would that be 3 directional antennas for each link? Thx.

inachu
inachu

I am sick of Linksys wifi. I bought 3 of them B,G,N each one always dropped signal. I bought a few weeks ago a d-link extreme and I am very pleased with my speeds and no drops from other wifi routers fighting for signal space.

jason
jason

I didn't read through all the posts, so forgive me if this is redundant. My personal opinion is this is great for the techie consumer for their home/personal use. However, in the LAN environment, most companies have already begun the migration process to Gbit networks and 802.11n seems like a step backwards. Yes, I do agree that it could work in situations where you need connectivity but you can't run cables, but this wouldn't be a viable solution for companies that need quick access to network intensive programs. Jason

mcvr.chane
mcvr.chane

really intresting and thanks for the early info

Baltimore Jason
Baltimore Jason

This would be all fine and good if it wasn't for the fact that most of the 802.11 draft n equipment on the market is plain and simply garbage, rarely beating 802.11g in terms of throughput or reliability. I'll with hold my judgment until equipment made using the ratified standard has been out for a few months, but until that point I am most definitely sticking with my very reliable 802.11g equipment.

wolffjw
wolffjw

Okay, so I'm not too bright on 802.11n, or even earlier versions, but this article doesn't really help someone like me. Can someone please explain in a low level enough language for me to understand? How does this physically work? What are the physical connections? Does this use a steelhead, or some wireless iSCSI switch?

Neon Samurai
Neon Samurai

11n could be beneficial for the local network. The pipe out to the ISP may be 8 meg or less but within the local network you could have shared multimedia, shared network storage, video gaming or anything else bounching between two nodes.

Michael Kassner
Michael Kassner

If you have both nodes use the same number of RX/TX antenna pairs you will see an improvement equal to the number of pairings times the original SISO data rate. That improvement is based on using link connection best practices and not dependent on antenna directionality. I have tested a few 1 mile links with 802.11n and directional antennas with good success. The improvement might even be greater if MIMO takes advantage of multipath environmental situations along the route. This is more difficult to confirm as different antennas would have different horizontal coverages angles, which create different RF path opportunities. Real world data rates would have to be determined experimentally. One other advantage is the data rate gain from being able to use 40MHz channels.

w2ktechman
w2ktechman

Honestly, having wifi is not going backwards, if it is used well. Running cables everywhere adds extra overhead. For the main access and backbone, wifi is not recommended. For conference rooms, warehouses, or hard to reach with cable areas wifi is definately a great tool to utilize.

Michael Kassner
Michael Kassner

It is very interesting if you read articles from wireless developers. they have no real argument for the wired versus wireless debate other than the inability to lay cables or costs. A new tact being used is the proverbial "good enough" approach. 100-300Mbps may just be good enough for most office applications. That mentality does have some merit and will appeal to the CFO's. I try to approach the whole situation with the frame of mind that 802.11n is another important tool that I have to help solve network situations.

Neon Samurai
Neon Samurai

I watched the modem 54k standard wars and I'm not ready to jump back into that same mess with everyone's own version of 11n. The routers are pretty but my 11g does just fine for now.

Michael Kassner
Michael Kassner

Hello Jason, I was wondering if you could give some examples? I have several different 802.11n devices working on test beds and have seen significant improvements. Even so, I still would really appreciate hearing about any negative reviews that you may have.

w2ktechman
w2ktechman

Now I am not an expert on 802.11n so this may not be totally correct. Basically it uses 2-4 antennas to trasmit and receive data rather than 1. Because of this, interference such as multipath (multiple signals created from one, bouncing around and being received at different times) are used to increase the data rates instead of decreasing it. New version of OFDM allows for less overhead and more data == more throughput. Or simpler. More data transfers in the same time, farther reaching (distance). And, there are versions for the 2.4 Ghz and 5ghz bands. Better signal quality, and less interference.

emetesh
emetesh

Interesting. So would you speculate and say that IF: You had an operation link using g @ x miles. And you changed to n radios and added more identical pairs of antennas. You would still be able to cover the same distance and increase bandwidth? Because of the same frequency?

JCitizen
JCitizen

(sorry w2ktechman) I was trying to point this to Jason. and I will now just reiterate; that when the bean counters look at the time and costs it is a no brainer to make the small local expansion wirelessly; especially when it is possible to do so with security. I would never recommend it for a major node or remote office expansion either; unless that remote office is a long way from its MDF/IDF. Sometimes we ran into real resistance from land lords at business locations we didn't own and were fortunately able to talk them into letting us pull cat5 for one of our major nodes. If we hadn't we would have been forced to make the AP in the POT service closet at that locale and wirelessly bridge the rest of the expansion.

Baltimore Jason
Baltimore Jason

Michael, Sure, without getting into third hand accounts of trouble with draft n equipment, I personally have experimented with the linksys wrt300n and the first generation netgear n router(the model number escapes me at the moment), and I found both to be flaky at best. I'll grant that it's been several months since, but once I switched the two locations I'm referring to back to trusty wrt54gl's I haven't had any connection issues. Also, I'll concede, that when they worked I did see very impressive connection speeds, but they just didn't seem to be built with the same sort of consistency that older G routers are.

chrisl317
chrisl317

These have already begun with several patent lawsuits originating out of Australia. If I remember right even IEEE was wondering why the "N" group hadn't procurred a letter of assurrence from the patent holders first to develope the "N" standard.

Michael Kassner
Michael Kassner

I have linked an article that I have recently wrote about that very subject. http://blogs.techrepublic.com.com/wireless/?p=137 There is a great deal more optimism that this will get resolved today, than when I first submitted that article in September. I also have just submitted another article that discusses the 802.11a/g patent infringement issue between Buffalo and CSIRO as it affects Buffalo's newest router that now has DD-WRT as the pre-loaded firmware. I will link that when it becomes available.

w2ktechman
w2ktechman

on that page. It states most of the big players, but missed Cisco completely....

Michael Kassner
Michael Kassner

w2ktechman indeed hit the differences from 802.11n and the earlier standards of 802.11a/b/g. If you would like more information please let us know as I/we can get more specific if you would like. I also am not sure what you mean by physical connections, if you further explain that I would be more than happy to try and help.

JCitizen
JCitizen

for refreshing our memory of the history and definition of the term.

Michael Kassner
Michael Kassner

In one respect using 802.11n is the same as using multiple 802.11g or 802.11a devices. So that aspect would be correct. 802.11n has many other improvements, such as an improved OFDM and receiver firmware, which will also increase data rates. As an aside, I try not to use the term bandwidth as it has two definitions and can be misinterpreted, especially when talking about 802.11n. WiKi mentions: "Bandwidth is the difference between the upper and lower cutoff frequencies of, for example, a filter, a communication channel, or a signal spectrum, and is typically measured in hertz. Bandwidth in hertz is a central concept in many fields, including electronics, information theory, radio communications, signal processing, and spectroscopy. In computer networking literature, digital bandwidth refers to data rate measured in bit/s, for example channel capacity (digital bandwidth capacity) or throughput (digital bandwidth consumption). The reason for this usage is that the channel capacity in bit/s is proportional to the analogue bandwidth in hertz according to Hartley's law." Reason I mention this as 802.11n allows the usage of 40MHz wide channels, which is an increase in bandwidth over the normal 20MHz channels used with 802.11a/b/g. That increase in bandwidth will also give a substantial increase in data rate or digital bandwidth.

Neon Samurai
Neon Samurai

ddwrt seems to be the right balance between OpenWRT technical and Tomato easy (vendor firmware doesn't make the list). I setup RC4 on a new 350n and it runs perfectly. You just have to do the update in the right order: For version24 - flash mini firmware and do factory reset to write all default variable values. - flash the standard (std-blah.bin) and do factory reset to write all outstanding default values. - configure to your liking. The only thing I miss after running Tomato for months is mac address handling. On Tomato, you look at what devices are connected, click a mac and it dumps the mac strait into the static IP config table and asks you for a name and IP. With ddWRT you have to copy the mac from currently connected devices then manually switch to the static IP assignment and add it in. Oh, and I wish they'd fix the browswer inferface. It only works with IE currently. If you try firefox or konqueror you can make changes to the form fields but when you hit the apply button the later two browsers try to download the .asp file. I've only seen this error with apache without the PHP module installed; browser asks for page, apache tries to feed page to browser as a download instead of processing it. Booo.. version 23 worked with every browser I threw at it, version 24.. only IE.. I hope this limitation is only during the RC releases. So far v24 RC4 running on wrt350n v24 RC5 running on wrt54gs Both are happiness even with being limited to IE for the browser interface.

rob.smith
rob.smith

What type of antennas were you using? Much of the N hype is completely overlooking the impact of MIMO antenna selection. Using the basic antennas may get you the performance you need locally but trying to push the coverage area beyond the typicall 350feet will require upgrades to the basic system. And that introduces a whole new level of complexity on enterprise integration.

JCitizen
JCitizen

Good thing your waiting - it looks like the firmware will catch up before long; unless your already planning to use Linux based 3rd party software. I went on Amazon after your post, and was reading the reviews and it looks like most of those people bit off more than they could chew. The imbedded video review by Richard V. Burckhardt made it even more interesting. So far I've never seen a Linksys router that couldn't use better firmware as a new model. But I have limited experience. On each occasion the devices performed flawlessly after the 2nd firmware update.

Neon Samurai
Neon Samurai

I'm not about to turn on 11n networking but I gotta admit, I'm counting down the days until I can afford a 350n for home. I bumped my 54gs ddwrt firmware from version 23 to rc version 24 after getting home and I can feel the difference between the two hardware platforms. I do love the 54g series though; my router has treated me well even after the week when I repeatedly bricked it until I finally got a stable firmware on with tftp. Sidenote, we where having an issue with slow ssh and ftp connections on the businesses old router befsr81. ssh connections are nearly instant now though I believe it's due to ddwrt doing a far better job of providing local DNS caching.

JCitizen
JCitizen

Every since I read about the WRT54G and how well it worked with the Linux firmware that enthusiast were writing for it, I had a feeling the future was bright for Cisco/Linksys. Just having a flexible gateway solution was cool enough for me; but with the WRT54GS WOW!

Dennis
Dennis

I have installed 3 different Wireless N systems in the past year. The most recent one I ended up returning the system since it offered NO improvement whatsoever in distance. I had an existing Wireless G and needed to get the signal to a different building. The existing router got the signal to the building but not in it. So, logically I thought "N" would do it. I measured signal strength at several different points and it was NEVER at any point stronger than the G. It also died "at the door" to the second building. Consumers are wasting their money on "N" stuff at this point.

Michael Kassner
Michael Kassner

Buffalo makes several routers that work quite well with DD-WRT, i.e., not the one I mentioned that comes pre-loaded with DD-WRT. Just thought I would mention that as another option for you.

Neon Samurai
Neon Samurai

I got lucky and baught version 1 hardware before I knew what I was doing with wifi so I've been through openWRT/xWRT, Tomato and am now on DD-WRT. Actually, I spend the last few days in the market for a router to put on a client's network. I couldn't find a wrt54gl anywhere so I ended up with a wrt350n v1 now running dd-wrt rc4 very nicely. I may have to pop online and see who still has the 54gl in stock and ships to my area though. It's a downgrade from the 350n hardware wise but it has it's own benefits.

Michael Kassner
Michael Kassner

Thanks Jason, I totally agree with your assessment of Draft 1 equipment. I have found that Draft 2 devices seem to have a great deal more stability, range and reliability. You also have latched unto one of the best 11g routers that exists. The WRT54GL is a special edition specifically developed with more memory and a better chip set so that DD-WRT and other third party firmware can be installed. Are you running the standard firmware? Interestingly Buffalo has just come out with a router that has DD-WRT pre-installed and should be a very good device and hopefully Buffalo will be able to resolve their patent issues.

JCitizen
JCitizen

something would get done the next time they revisit the issue. Do you know of any professional organizations that lobby congress that could help put more pressure on them to overhaul the patent system? [which is what America needs anyway]

w2ktechman
w2ktechman

so much technology is unavailable because of companies that horde patents. This is just another of the many over the last few years that seems to be hindering progress. Personally, I think that if you hold a patent for something, you should be actively working on generating a product for it. Without this product generation going on, the patent should be invalidated.

Michael Kassner
Michael Kassner

I could not reply on your post, so I thought I would try it here. In my humble opinion if WiMax becomes a viable solution it will be huge. I have a 3G data card for my notebook that I use constantly in my line of work (network field engineer). Even with available Wi-Fi networks I still prefer the data card, as I do not have any problems with VPN connections and security is obviously better. I see WiMax as the next evolutionary step to the 3G data networks. It appears that it will be a rocky road though as Sprint and Clearwater are re-thinking 4G/WiMax and how to get there. 802.11n will more than likely survive in a different role, especially for private and in-building situations. Because WiMax will have the same problems of RF penetration that 3G/GSM face. That is why Sprint and other telcos are also interested in Femtocell technology, which addresses the RF penetration issue in buildings.

w2ktechman
w2ktechman

for quite some time. Except for the very mobile users, I think WiMax is best suited for non-business use. I think that 802.11 is where most home/business should be focused. However, there are many who will/would disagree with my assessment, and I am sure a lot of people (home users) will decide that WiMax is for them.

Michael Kassner
Michael Kassner

It is really interesting to follow 802.11n and 802.16. Cisco is either hedging its bet by developing equipment for both or they feel that there is a need for both. The new 802.11n equipment they have just released is pretty exciting.

w2ktechman
w2ktechman

not much on the horizon for 802.11 networks once WiMax is avail.

cdiebel
cdiebel

Cisco just spent $330 on Navini networks to become quickly involved in 802.16e...approx $100 million more than its market value at the time.....

Michael Kassner
Michael Kassner

I think the patent in question has to do with multipath transmission technology and Cisco has just recently released products for 802.11n Draft 2. Also maybe Cisco is already paying royalties as are some other companies.

wolffjw
wolffjw

Thank you very much. w2ktechman's answer was what i needed.