Q: in China ,we just can use UNII-3, just 4 channel(and 165 is not availabe),if we use 40Mhz,just two channel point can be used.it is bad news.any good suggestion?

A: Yes you can still manage pretty well by using only the UNII-3 (4 non-overlapping 20 MHz or 2 non-overlapping 40 MHz channels) along with the standard 2.4 GHz channels as long as you select dual-band equipment.  You will also have the 3 non-overlapping channels in the 2.4 GHz band for a total of (7) 20 MHz channels.  Now it will come down to the types of applications, the user expectations and the limitations of any pre-existing equipment or applications.  You may decide to leave your normal data users on the 2.4 Ghz band while moving critical data and real-time applications to the 5GHz band.  Oh course this would require the use of 5 GHz voice/video client devices which you may or may not have access to. You will need to determine what is most important to this network.  Will high speed data be most important or will 5 GHz voice services be the most important? 

Q: Can you (later of course) explain the difference between Data and voice RST requirements?   I understand the SNR reqs.

A: Yes, the raw RST levels are published in the 802.11n amendment, although you may see slight differences posted by the equipment manufacturers for their devices.  The 802.11n TG recommends that you add at least 11 – 12 dBm of margin (cushion) to these raw signal strength requirements.  The margin allows some leeway in the ambient energy levels to account for unexpected dead or weak spots which may be caused by signal reflections canceling each other or by obstructions in the building such as pipes, ducts, elevators, tile, brick or rebar.  In most cases the 11-12 dBm fade margin will be sufficient for data applications but since real-time applications such as voice are more affected by signal drops you should as a general rule add even more fade margin into your site surveys prerequisites.  Most VOIP manufacturers recommend an additional 5 dBm of margin over and above the amount added for data.  That is why the chart column for voice has a higher number than the one for data.

Q: Is this a clear field?  ie how did you get multipath in it?

A: Yes, this is a wide open, clear field and as you can see, 802.11n works just fine here.  One of the many misunderstandings concerning 802.11n, is that it requires obstructions or reflections to cause the multipath.  That’s not the case.  The multipath referred to in 802.11n and MIMO is the multiple paths of the spatial streams created by using several antennas each with their own radios.  This has very little to do with the correlated multipath caused by reflections that we normally think of in RF work.  It’s true that additional non-correlated multipath signals may help to improve the separation of the spatial streams, but because the signal strength and SNR is also reduced by this it is not always helpful.

Q: is it possible to get a picture of the field (later)

A:  Sure, here it is… 

You can see the RF cage in the middle.  We use that for isolation testing (mostly 2.4 GHz).  It is about 300 ft. to the back side of the RF cage from where we placed the APs.  If you look closely at the heatmaps from the site survey you can make out the attenuation and reflection effects caused by this Faraday cage.  The full distance is about 550 ft. to the beginning of the residential area (fenceline) and about 300 ft. wide (left to right).  It makes a great place to test WiFi radios and antennas.

Q: Hi, when do you think clients will have 3 or 4 spatial streams please?

A: Good question, but I don’t have a good answer.  You can find 3 stream clients today (Intel WS5300 and newer) but I haven’t heard of any four stream clients yet.  This could be because nobody as yet is asking for them, except you and me. J

Q: Which wireless hardware vendors would you recommend for small to medium sites and large sites respectively?

A: It seems that all of the hardware vendors have strengths and weaknesses and none of them is perfect.  For larger enterprises that already have an installed WLAN base, I’d say if you’re familiar and satisfied with the vendor stay with them, especially if you’ve invested in a WLAN controller-based system (lightweight architecture).  There’s no arguing that Aruba and Cisco are the tops of the heap for that scenario, and they have such large installed bases that it’s easy to find engineers familiar with those products.  But if you are interested in technical innovation with a broad support base of additional product lines, you should take a look at what Motorola is doing with the Symbol line since they purchased it.  They have a very comprehensive hardware offering which includes a scalable WLAN controller selection and nextgen line of 802.11n APs.  They also offer an excellent management interface, site design, planning, and survey solution, and of course, the AirDefense Wireless Intrusion Protection System suite.  For large, medium, or small WLANs, I would suggest that you also look at AeroHive.  I am not 100% sold on the lightweight architecture philosophy.  I agree that management needs to be centralized but I do not like the idea of having to transport every data frame back and forth between the controllers and APs for processing.  And I don’t trust the single point of failure for the entire WLAN segment you have if a non-redundant controller breaks down.  AeroHive has an elegant solution.  They provide a centralized management plane for configuration and profile management that is just as powerful as any of the WLAN controller architectures but allows all of the APs to retain their independence and perform frame processing locally.  In other words you can maintain AP and SSID profiles and push configurations out to any number of APs from a centralized management console, but once that’s done the APs have no further need to be connected to the manager.  The APs communicate directly to each other using “Cooperative Control”.  They even support mesh backhaul by default so that you do not need to connect each AP to an Ethernet connection, unless you prefer to.  I think AeroHive offers the best of both the autonomous and controller-based worlds.

Q: What site survey software and hardware (laptops included) would you recommend to wireless LAN Techs for use in maintaining and surveying wireless LANS?

A: There are many good site survey applications available on the market.  I personally prefer the Ekahau and AirMagnet programs.  Either one, Ekahau Site Survey Professional or Airmagnet Survey Professional will have the accuracy and features needed to provide professional level site surveys on a commercial basis.  (Contact Scott Williams at This e-mail address is being protected from spambots. You need JavaScript enabled to view it or one their offers through WiTS).  The hardware part is not so clear cut.  The two main requirements for a usable site survey terminal are weight and battery life.  Site surveys force you to be on your feet, walking around holding the survey PC for hours at a time.  Even the lightest laptops will feel like a ton of bricks after a short time.  A comfortable shoulder strap or PC tray should be near the top of your list of required accessories.  Big screens are nice and make it easier to mark your position on the graphical floorplans that you use during the walkabout, but they are not worth trading for battery life.  There’s nothing worse than surveying for hours and then losing all your work because your battery cuts out before you can swap it with a fresh one.  If you find you need to frequently survey outdoors you should choose a PC with a water resistant keyboard and screen.  The most popular site survey PC I’ve seen used by most survey professionals is the Lenovo tablet PC but since I’ve been known to drop my PC and run into walls while performing site surveys, my personal choice is for the Panasonic Toughbook tablet PC line.  The field-tough Toughbooks are a little more expensive (not that much more) but are moisture resistant, drop protected, and have screens that can be seen pretty well in bright sunlight. 

Q: Which wireless training provider(s) would you recommend for overseas techs (living in the Caribbean) interested in pursuing CWNP training?

A:  Of course I’m biased on that question to my own company.  WiTS travels all over the world and also offers the best online training experience available.  Our online instructor-led classes are as close as you can get to an in-person class, including the ability to perform all of the same labs, on the same equipment but through a remote access connection.  WiTS would love to discuss your training requirements further. 

Just drop a note to Scott Williams ( This e-mail address is being protected from spambots. You need JavaScript enabled to view it ) and he will help you customize a training solution.

Q: Does that mean that you design around your weakest client in terms of receive sensitivity and tx power?

A: Well , yes.  But, only to the extent of the weakest client that you need to provide an assured service level.  In other words, if you are deploying an 802.11n WLAN upgrade and need to support high speed data for laptops with integrated client adapters that can TX at a maximum of about 16 dBm, and you also need to provide good coverage for users with Vocera B1000 voice badges that have client power of only about 14 dBm; then you would design and survey for the Vocera clients.  But, if you were only allowing the Vocera connectivity as a courtesy and the mission critical applications ran on the laptops then you would design and survey primarily for the laptop’s 16 dBm output power while providing Vocera users with best effort access only.

Q: what about speeds using AES?

A:  In order to get the full speeds of 802.11n you should use only AES (CCMP).  Since the CCMP (AES) equipment contains additional hardware to offload the encryption/decryption duties from the WiFi chips there is no degradation by using AES encryption.  In tests here in the WiTS labs for our new certification classes, we have found that using CCMP encryption is exactly as fast as using no encryption at all (which is strongly NOT recommended).  Bottom line is that AES (CCMP) encryption is unbreakable and fast.  Use it whenever possible.

Q: how ccmp selection increases thr throughput ?

A: Actually CCMP doesn’t increase the throughput speed, it simply doesn’t reduce it like TKIP and WEP do.  While preparing for our new “802.11n Essentials” and “Certified Wireless Design Professional (CWDP)” classes (which are coming out this Fall), we extensively tested 802.11n connections from multiple vendors using 1. No Security, 2. WEP Security, 3. TKIP Security, and 4. CCMP (AES) Security.  We found that enabling either WEP or TKIP reduced 802.11n signaling speeds to a maximum equal to the legacy rates of 802.11a/g or 54 Mbps.  We also found that enabling CCMP did not restrict the maximum signaling rate of the 802.11n connection (300 Mbps in our tests) from what was available with No Security enabled.  In actual throughput tests we could find no measurable difference between the speeds of an open, non-secure 802.11n file transfer (roughly 90 Mbps) and one which used CCMP encryption.  However, the throughput speeds of 802.11n with WEP or TKIP was never able to go faster than what we normally see with 802.11a/g (usually no better than 22 – 24 Mbps).  The use of AES encryption does not impair 802.11n speeds but WEP and TKIP do.  We suggest using only CCMP (AES) with 802.11n.  You can read the details of these tests in the new CWDP Study Guide from Sybex when it comes out Q4 – 2010.

Q: Quick question, what channels are rather to be used due to the radar signal conflicts?

A: Well, the only channels that are required to be placed under Dynamic Frequency Selection (DFS) control as of the 802.11h amendment are the channels in the UNII-2 and UNII-2e (middle 5 GHz bands).  Those channels are: UNII-2 - 52, 56, 60, 64, and UNII-2e – 100, 104, 108, 112, 116, 120, 124, 128, 132, 136, and 140.  Those channels can still be used but you can’t select them individually and set them statically.  You have to let DFS select an available channel from that range after it tests the air interface to see if radar is in use on that channel.  The DFS rules make the use of these channels less useful for applications that have critical roaming requirements such as voice over IP over WiFi services.  But they are still very useful for high speed data applications and 40 MHz channels work very well in these bands.

Q: How to measure the cell overlap in percentage?

A:  Ah, good question.  That is not as easy as everyone seems to assume.  To try and work that out manually takes a lot over measuring and calculating.  The easiest way to assure good cell overlap is to use a professional quality site survey application such as Ekahau Site Survey Professional or Airmagnet Survey Pro.  Either of these programs will allow you to see the channel overlap as color coding in the heatmaps view or as a percentage in the formal reports.  From that you can adjust the APs to provide the overlap required for your application(s). 

Q: What do you think about Single Channel solutions? How will this affect planning and site surveys?

A: There is nothing wrong with using single channel architectures (SCA), especially if you are in a controllable environment that does not have a lot of competition from nearby neighbors who you can’t control.  SCAs are a great solution to situations where you need to have the fastest roaming conditions.  Since there is only one channel, scanning times for roaming clients are greatly reduced.  And in the case where virtual cell architectures are in use in addition to SCA, there is essentially no handoff latency due to roaming.  But, that is not always the only criteria for a QoS network.  Remember, for collision avoidance, there can be only one transmission at a time within each BSA (cell or audible channel).  If there are a large number of users contending for the channel and especially if those users are sending not just voice packets, which are usually very small in size, but also data traffic, which normally uses much larger packet sizes than voice, there may be discontented users in your enterprise.  Once more, the use of the 5 GHz channels can really help to ease the strain, even with single channel architectures.  Depending on your hardware you can determine where best to put the single channel voice users and if necessary setup additional single channels for other types of applications.  Of course, you can control which applications run over which channels by assigning unique SSIDs for each application.  Bottom-line: SCA can provide an excellent real-time platform but you may need to get creative when mixing real-time applications with other types of applications, such as data, locationing, or telemetry, to get the best results for everyone.  But, of course that also goes for multi-channel architectures.

Q: Would you still statically configure AP Channels with the newer "automatic remediation" solutions for rogue APs.

A: I prefer to statically map the channels because I believe I have a better understanding of how RF and the 802.11 contention mechanisms work then the software algorithms written into the current crop of WLAN controllers do.  This is especially true in mixed client/application environments.  If everybody is using VOIP and that is the only criteria then channel and power control are easy.  But when you have different, sometimes antagonistic, applications all competing for the same resources, you have to be able to take a higher and broader view of the WLAN than the automated radio management routines are capable of.  This is especially true for WLANs that are located in areas that can be affected by neighboring residential or business wireless networks, or locations that cannot control the existence of Rogue APs.