WLAN Book.com and @WLANBook

Argument Against Using iPad for WiFi Surveys

Argument Against Using iPad for WiFi Surveys

Argument Against Using iPad for WiFi SurveysI came across a CWNP forum post with comment below arguing against using iPads as a WiFi survey platform. Regular readers of this blog and followers of @WLANBook know that we think using iPads for WiFi Surveys is a good idea. We like the concept so much that we have created an “app for that“.

I agree that the iPad isn’t the perfect platform for WiFi Surveys but I disagree 100% that it is a bad idea. My counter arguments are after comment below.

I was thinking about site surveys using an iPad, too, but after some research and discussion with a couple of software vendors, I came to conclusion that this was a bad idea. There are many reasons why you can’t use an iPad (or a droid tablet, for that matter) for a quality site survey. The most important reasons are:

(a) iPads and most of other tablets have a scaled-down Wi-Fi chip. If you perform active surveys, no matter how good your AP is (e.g. 450 Mbps), your measured throughput rate will be too low.

(b) Site survey software typically extrapolates data and performs some other CPU-intensive operations. iPads don’t have sufficient computing power for that within a reasonable amount of time.

(c) iOS and Android does not allow developers to capture raw frames. This makes it impossible to gather vital information necessary for a quality site survey. To give you just one example, if you can’t capture raw frames, you cannot detect multi-SSID APs; the application will “think” that it is dealing with several different physical APs located very close to each other, all of them using the same channel. As a result, it will show a totally wild interference picture.

There are other less important reasons. E.g., it’s difficult to mark your location on the floor plan with a finger.

The bottom line is that while iPad is an attractive platform, one can’t use it for each and every task, unfortunately. There are things that require a traditional OS with a fast Intel CPU. I’d go for a light Intel i5 or i7 notebook or tablet, Windows 7, and a good site survey tool (Ekahau, TamoSoft, AirMagnet).

Using iPad for WiFi Surveys is Good Idea

I was thinking about site surveys using an iPad, too, but after some research and discussion with a couple of software vendors, I came to conclusion that this was a bad idea. There are many reasons why you can’t use an iPad (or a droid tablet, for that matter) for a quality site survey.

There isn’t a universal definition of a “quality site survey” but if a post WiFi survey report includes below then survey was was pretty good.

- Desired RSSI coverage map
- Desired S/N** ratio
- Throughput tests if doing active survey

Our iPad app can provide above and the results be closer to what the end user will experience because the actual device they will be using was used for surveying. **Noise value from radio card is not possible with current iOS API but can be estimated.

(a) iPads and most of other tablets have a scaled-down Wi-Fi chip. If you perform active surveys, no matter how good your AP is (e.g. 450 Mbps), your measured throughput rate will be too low.

True, hardware is different / smaller than some laptops but expecting to get 450 Mbps or even 300 Mbps from any device (including laptops) is not realistic. Most of the data rates on vendor boxes for 802.11n are marketing hype actual results are much lower. Check out post at WLANPros.com about iPad 1 and 802.11n.

(b) Site survey software typically extrapolates data and performs some other CPU-intensive operations. iPads don’t have sufficient computing power for that within a reasonable amount of time.

I don’t have any real world benchmarks (either) but based on field tests the iPad2 is very usable for all tasks including heatmap generation.

(c) iOS and Android does not allow developers to capture raw frames. This makes it impossible to gather vital information necessary for a quality site survey. To give you just one example, if you can’t capture raw frames, you cannot detect multi-SSID APs; the application will “think” that it is dealing with several different physical APs located very close to each other, all of them using the same channel. As a result, it will show a totally wild interference picture.

True iOS wireless APIs dont’ allow access to raw frames. If there was enough demand for this type of app I guarantee someone could create a custom driver for iOS that would get raw frames from wireless card. Not sure if I agree with multi-SSID AP issue mentioned above. The app should be able to correctly account for this scenario since most modern APs use virtual APs with different Mac addresses per SSID. Also most survey access points are configured with single SSID per frequency.

There are other less important reasons. E.g., it’s difficult to mark your location on the floor plan with a finger.

Use stylus, also zooming in/out on iPad is pretty easy.

The bottom line is that while iPad is an attractive platform, one can’t use it for each and every task, unfortunately. There are things that require a traditional OS with a fast Intel CPU. I’d go for a light Intel i5 or i7 notebook or tablet, Windows 7, and a good site survey tool (Ekahau, TamoSoft, AirMagnet).

iPad is solving issues that laptop based solutions can not.
– Battery never lasts long enough
– Too heavy
– Cost of laptop high relative to iPad
– Cost of software very high relative to typical iOS/iPad app.
– Survey with actual device/wireless card that will be used on network
– Perform throughput test with actual device/wireless card that will be used on network
– Need for external PC/USB card
– Need for external GPS sensor
– Rear facing camera to snap photo of physical site survey issues (closet, AP placement, etc)

It easy point out negatives only without including the positives of a tablet/iPad solution (or negatives of a laptop based solution).

80/20 Rule

iPad meets 80/20 rule.

iPad WiFI Survey easily has 20% of the core features needed and will get 80% of end results needed for a quality site survey.

Most poor quality site survey reports have nothing to do with software / hardware used but knowledge of operator.

If you liked this post, subscribe using below

WLAN Book RSS Feed

RSS Email WLAN Book by Email

.

Filed under: WLAN Site Survey, WLAN Tools

9 Responses to “Argument Against Using iPad for WiFi Surveys”

  1. Greetings Zaib,

    I can’t agree with some of the “pro-iPad” arguments, but before I explain why, a disclaimer: I work for one of the companies mentioned in the original post and we make a site survey tool for Windows. That said, I have no intention to bash a future iPad product that some users might (or might not) prefer; I’ll try to be unbiased.

    As far as “too heavy, too expensive, need a card, need a GPS sensor” group of arguments is concerned, I think they are a bit subjective. There are lots of great light Windows notebooks and tablets on the market today, their price often being close to the price of an iPad. They have built-in Intel, Atheros, or Broadcom adapters that you can use with site survey software, and a 100 gram Bluetooth GPS sensor in your pocket is not a big burden. I mean, if you love iPads, you can find a dozen arguments why they are cooler. If you like Windows notebooks, you can do the same.

    As far as purely technical arguments are concerned, I think you overlooked some crucial facts.

    1.”The app should be able to correctly account for this scenario since most modern APs use virtual APs with different Mac addresses per SSID.” Sure, they do. You have AP #1 with the MAC address of “00:02:00:00:00:55″ and you have AP #2 with the MAC address of “00:02:00:00:00:52″. Both work on channel 1 and have different SSIDs. Same physical AP? Different APs? You’ll never know if you can’t capture raw frames. There is no need to explain why this matters.

    2.”True, hardware is different / smaller than some laptops but expecting to get 450 Mbps or even 300 Mbps from any device (including laptops) is not realistic. Most of the data rates on vendor boxes for 802.11n are marketing hype actual results are much lower.” A TCP or UDP data throughput rate is directly related to the PHY rate. A PHY rate over 200 Mbps is very common in a good WLAN, if the client is not too far from the AP. This means that a data throughput rate might reach 100 or 120 Mbps. You can’t measure this with iPad, where the max. PHY rate is about 50 Mbps.

    Finally, there is a very important point that was never mentioned in the original post or in your comments. With raw frames, the application can obtain far more important pieces of information, for example channel width and secondary channel number. iPad tells you that there is an AP on channel 6. What the secondary channel, 2 or 10? Is there a secondary channel, or it’s 20 MHz only? Answers to these questions cannot be obtained with iPad, but they have a huge impact on the quality of information in your site survey.

    Oh, and I question the accuracy of dBm signal level readings that you get from the iOS API. They might be very inaccurate.

    Yes, if a custom driver for iOS becomes available, then some of the problems will be solved. But for now, I think the rule is 20/80 rather than 80/20. Sorry for the long post.

  2. I can’t agree with some of the “pro-iPad” arguments, but before I explain why, a disclaimer: I work for one of the companies mentioned in the original post and we make a site survey tool for Windows. That said, I have no intention to bash a future iPad product that some users might (or might not) prefer; I’ll try to be unbiased.

    >> NP..I welcome discussion.

    As far as “too heavy, too expensive, need a card, need a GPS sensor” group of arguments is concerned, I think they are a bit subjective. There are lots of great light Windows notebooks and tablets on the market today, their price often being close to the price of an iPad. They have built-in Intel, Atheros, or Broadcom adapters that you can use with site survey software, and a 100 gram Bluetooth GPS sensor in your pocket is not a big burden. I mean, if you love iPads, you can find a dozen arguments why they are cooler. If you like Windows notebooks, you can do the same.

    >> Name one business class windows notebook that meets above and price for me with extra battery and external card if needed to work with software. For work I can not purchase consumer class PCs that will not be available in 10 months leaving all engineers with different types of machines.

    As far as purely technical arguments are concerned, I think you overlooked some crucial facts.

    1.”The app should be able to correctly account for this scenario since most modern APs use virtual APs with different Mac addresses per SSID.” Sure, they do. You have AP #1 with the MAC address of “00:02:00:00:00:55″ and you have AP #2 with the MAC address of “00:02:00:00:00:52″. Both work on channel 1 and have different SSIDs. Same physical AP? Different APs? You’ll never know if you can’t capture raw frames. There is no need to explain why this matters.

    >> Not clear on config and why it would be configured as described? Is this production network that is misconfigured or survey AP-on-stick?

    2.”True, hardware is different / smaller than some laptops but expecting to get 450 Mbps or even 300 Mbps from any device (including laptops) is not realistic. Most of the data rates on vendor boxes for 802.11n are marketing hype actual results are much lower.” A TCP or UDP data throughput rate is directly related to the PHY rate. A PHY rate over 200 Mbps is very common in a good WLAN, if the client is not too far from the AP. This means that a data throughput rate might reach 100 or 120 Mbps. You can’t measure this with iPad, where the max. PHY rate is about 50 Mbps.

    >> What app are you using to test?

    Finally, there is a very important point that was never mentioned in the original post or in your comments. With raw frames, the application can obtain far more important pieces of information, for example channel width and secondary channel number. iPad tells you that there is an AP on channel 6. What the secondary channel, 2 or 10? Is there a secondary channel, or it’s 20 MHz only? Answers to these questions cannot be obtained with iPad, but they have a huge impact on the quality of information in your site survey.

    >> Not clear on above…is user surveying for diagnosing network or survey during design phase / validation phase?

    Oh, and I question the accuracy of dBm signal level readings that you get from the iOS API. They might be very inaccurate.

    >> Would love to learn more this.

    Yes, if a custom driver for iOS becomes available, then some of the problems will be solved. But for now, I think the rule is 20/80 rather than 80/20. Sorry for the long post.

    >> 0_o joke?

  3. I think there is another variable that has not been mentioned – antenna design on mobile devices and tablets. Much of the survey data is dependent on the signal measured by the chipset and the differences in antenna design between laptop integrated cards, USB cards, and mobile device antenna designs can greatly influence (skew) the results. This could be especially problematic since antenna designs for mobile devices are often constrained and sub-optimal due to space and placement restrictions.

    I would advise engineers to survey with a device that is most representative to their production client devices that will be used. If that is tablets then there may not be any issue surveying with a tablet. On the other hand, if most of the clients will be laptops, phones, mobile scanners, or embedded M2M devices than a tablet survey may not accurately reflect the what users will experience in production.

    What are your thoughts?

    Thanks,
    Andrew vonNagy

  4. Yep, I agree. I think with increased # of tablets using WiFi for business apps (users and customers) it should be a requirement to survey with tablet.

  5. Great points were made on both sides. Logical and well thought out – but with bias based on individual experiences.

    The best survey platform I’ve ever used has been a very expensive, yet professional and rugged, Motion J3400 Tablet. W/12 internal antennas, so no matter what way you hold it, there are always 3 2.4GHz and 3 5GHz antennas w/o interference.

    With its ‘Work Anywhere’ case and sling – I’ve done hours and hours of survey work without the arm or shoulder pain normally associated with a site survey with a powerful (read big) laptop/tablet.

    On the other hand – I really like the 80/20 rule and for most work now I use an Apple MacBookAir 11″ – using USB NICs and Windows in a VM or Bootcamp. I’ve found this setup quite light, easy to hold, and works like a charm.

    It has been VERY stable with AirMagnet Survey Pro (Other systems tend to crash AM Survey software periodically) – The save and continue button is your friend!

    A professional quality – read lots of details with great ways to slice and dice the data – on a native Macintosh platform would be fantastic. (NetSpot is a start down this path)

    One mis-conception from the above discussion is the RSSI is paramount to a survey. Sure – it’s one of the variables. But measured co-channel interference is a greater predictor of client health and throughput.

    Clients need an entire set of criteria, RSSI of the primary AP, RSSI of the secondary AP, Co-Channel Interference, Jitter, Latency, SNR, Raw Spectrum Interference, Adjacent Channel Interference, Frequency Load, among others. All these need to be brought to light and confirmed during the Verification Survey!

    RSSI is required… but not sufficient alone to tell how well any given client is going to react.

    Keith

  6. I don’t know … it seems now a days I am surveying for mobile devices.
    Yeck, I just started using a tablet for surfing at home.

    I have to survey for really hot AMs if I never expect a wifi cell phone to work. I’d like to survey with a device I am expecting to have to provide coverage for or at least something close. I am sure first couple trys wont be super good but after seeing years of another vendors product that still runs like crap even with a high powered laptop i am ready try something else. All in all I am not really impressed with the current wifi tools out there. Blue screens and crashes are the normal along with outrageous prices.

    I do worry about Apple not playing nice with the drivers. Maybe a droid version that is open with the drivers.

    Andy is right about the antennas not being good. Does anyone now what an iPhone’s dBi is?

  7. @Andrew regarding antenna design – yes, I fully agree with you. If you use a laptop, you have some flexibility in selecting an adapter for a survey. Depending on the hardware compatibility list of your site survey software, you can often choose between using an integrated adapter or one of the supported external cards. If you use iPad, you have no choice at all; you have to use the iPad adapter and nothing else.

    @Keith – yes, exactly, a set of criteria. My point: no raw frames – no details on co-channel/adjacent channel interference.

    @Zaib – “Name one business class windows notebook that meets above and price for me with extra battery and external card if needed to work with software”. Yes, I agree about external battery; it’s often required. But including the cost of the notebook is not fair, IMHO. It’s like “it costs $50,000 to drive my kids to school, because I need a car.” But you already have a car, and you use this car for many other purposes:-)

    “Not clear on config and why it would be configured as described? Is this production network that is misconfigured or survey AP-on-stick?” – a production network. You create multiple SSIDs on one AP and they are assigned unique MAC addresses. This is a typical scenario.

    “What app are you using to test?” – doesn’t really matter. I can use iPerf or TamoSoft Throughput Test or any other utility. Regardless of the tool being used, you can achieve a 100 Mbps TCP throughput rate and even a higher UDP throughput rate if your PHY rate is above 200 Mbps.

    “Not clear on above…is user surveying for diagnosing network or survey during design phase / validation phase?” – Let’s say you’re surveying a network that has already been deployed. Your site survey software detected two APs with high signal levels in a given area, on channel 1 and channel 6. Your site survey software needs to build an interference heatmap. If the AP on channel 6 uses 40 MHz channels and the secondary channel is 2, you have much interference. If the AP on channel 6 uses 20 Mhz channels or if the secondary channel is 10, you have virtually no interference. How are you going to address this in iPad?

    “Would love to learn more this.” – me too. I never compared the dBm values reported by iPad with the ones reported by Windows software, but I suspect that considering the fact that you have no control over the scanner in iPad (scanning interval, number of samples, set of channels), the values might be highly inaccurate. It would be interesting to perform such a test. I’ll try to find some time for that.

    “0_o joke?” – well, partly :-) You see, I’m a pragmatic guy. If I need a job done, I look for the right tool. I don’t look for a tool that works on a specific platform that I like. I like Windows, but if I need to set up a high performance and reliable web server, I’ll use FreeBSD or Linux. Same with site surveys: If someone needs a job done, he’ll probably use a Windows computer. However, there are many people who use Windows IIS as a web server quite successfully, and I’m sure that there will be many people who will use your iPad application.

  8. I’ll post iOS throughput test soon between iPad 2 on 5Ghz and a wired device You may be right that it is limited to 50 Mbps.

    We are going to continue working away to make app better and hope that Apple releases a public API in future. The fact that they have a public API on Mac OS X gives me some hope that they will have one on iOS too.

  9. I was trained in past to survey with weaker client . Has this changed now that 802.11n is abundant. I would think the iPad is a weaker client but normal for site survey as it uses the SISO streams and more people are using these anyways.

    I think doing so will only help you in the design process. Sure throughput goes up with a better “N” multiple stream client and you may need less APs in the deployment to win the deal. But something to consider is the weakest link with BYOD. If you survey with a scaled down client card.

    The BYOD experience will prove to be better for those devices that you really don’t want on the network. Some engineers often survey with the external cards that have stronger RSSI values and think that because they used it the rest of the clients are the same. -WRONG. You need a balanced approach I think. Not too weak and not to strong. If you are just looking at the Zoom Zoom Zoom of the survey-you will miss the Boom Boom Boom that will come to clients in the Room, Room, Rooms!