Wireless Site Survey Fundamentals

So you’re concerned about your wireless network. I’m guessing you’ve had your customers–corporate users, bar guests, or university students–complain about all kinds of vague problems. The wireless drops, can’t connect, or my all-time favorite, “it’s laggy.” The last one always makes me want to pull my hair out, and that’s serious for a bald guy. What if it isn’t complaints you’re receiving? What if it’s a 60,000 sq-ft new building you’re moving into. As an IT professional you want to be certain you have proper coverage and you use a standard 125′-radius coverage zone to help you prepare. Is that right? Wouldn’t it be better if you could model the coverage based on some actual data? What about the “wireless site survey”?

With all these complaints and objectives in mind, you’ve made the decision to contract a wireless site survey, but you’re not quite certain what to expect. There are some things to keep in mind so that you get the biggest bang for you buck. The first is starting with an understanding of the different survey types. The second is what you should receive as a deliverable. If you’re account executive or sales engineer doesn’t explain this, be sure to ask if there’s someone in your VARs organization who can. If you’re spending the money, know what you’re getting into.

Passive Site Survey

Passive Site Survey Heatmap
Passive Survey Signal Strength, Survey Route, & WAP Channels

A passive site survey is an actual, production , radio frequency (RF) analysis. With this type of survey the wireless engineer doesn’t care about building construction material, cube height, or window-encased conference rooms. The technician simply walks every square foot of your facility collecting the RF data for off-site analysis. If you want good coverage in the ladies and men’s restrooms, the technician will go in there for RF samples. The idea is that passive surveys collect actual, production signal.

You should expect that the on-site technician will bring something like the Ekahau Sidekick onto your facility, so you may need to check with the security teams. This equipment is a multi-antenna device that collects all the wireless data; not actual traffic, but radio signals from yours and your neighbors access points, plus equipment that you didn’t know you had. The Sidekick is a wearable, AP-looking piece of kit and is used in conjunction with a Microsoft Surface, iPad, iPhone, or laptop to track the survey path.

Passive surveys are used to verify things like coverage, signal-to-noise, and channel overlap data. If you are having complaints from users in certain parts of your building, a passive survey will help you reveal those coverage gaps. Another use case is to conduct a passive survey after an office reconfiguration. This happens quite a bit in many organizations. Cubes are moved, office are torn down and replaced with conference rooms, etcetera etcetera. These changes to the office environment impact the unseen radio waves coming from your access points and client workstations. Knowing the impact will help your network engineering team adjust to the changing landscape.

Predictive Site Survey

A predictive site survey is speculative and meant to give you an idea for WAP placement, starting channel plan, and expected coverage and wireless performance. For all my surveys I use Ekahau products because they help give some of the best analysis in the industry, though there are other options. If you’re moving into a new building, a predictive survey is the kind you’d like to do. The wireless engineer will take your building floor plan and work with you to understand cube layout, conference room locations, building construction (cinder block, sheet rock, glass, wireless access point (WAP) ceiling height, etc), and overlay this information onto a floorplan that you will provide. Actually, you need to provide a floorplan for all of these studies, but your wireless engineers will tell you that during the engagement.

Passive Survey - Signal To Noise
Signal to Noise Ratio

Within site planning software, like Ekahau Pro, you can set RF patterns based on coverage requirements (support for voice, data, and other time sensitive applications) and vendor radio patterns. If your company has settled on an Aruba CAP standard for example, your site planners should be able to predict with a great deal of certainty what your coverage areas will look like. With the predictive site survey you’ll see how many WAPs you’ll need to provide proper coverage, what the coverage expectation is, and other probable deployment challenges. You’ll also be able to have a rough idea, in most cases, how other RF is interfering with your SSIDs. Modern printers have WAPs built in and will normally beacon on the 2.4 spectrums.

I’d actually recommend for new construction a predictive survey, followed by a passive survey to verify coverage—and other requirements—are met. If you’re working with a value added reseller (VAR) to outfit your building, this may be a good spot for contract negotiations. Hey, VAR, I’d like 3hrs of predictive analysis, and 8hrs of passive survey for this project…go ahead and scope that along with the number of WAPs.”

Many times, when you combine a predictive and passive survey into the project, you’ll come out a little bit ahead in your project budget. Let’s say you’re building out a 65,000 sqft office space and you swag your WAP count at 50 devices. Those 50 devices may be a total cost of $20,000. You reach out to your VAR for some pricing and services, and ask for a predictive and passive survey. Depending on the VAR, and other factors, expect about $15,000 for those services. Base on the predictive survey, the estimation is you’ll only need 30 WAPs which saves your budget $12,000 and you’ll get to see how your WLAN is performing post-installation. Pretty good deal even if it costs you an additional $5,000 on your budget.

There is another type of survey, the active site survey. I won’t go into much detail here because it has all the characteristics of a passive with one difference. On the active side, the WIFI engineer will send test data across the WLAN. Typically it will use something like IPERF to measure throughput across different WAPs and it gives you the ability to see actual network.

What to Expect

If your VAR is to be successful helping you identify wireless issues, you need to engage. To help the VAR help you, make certain you work with them to define your coverage requirements. Many customers I’ve worked with have no idea what they need, so I try to help guide that conversation. Think about your user-base. Do they use soft-phones and stay on a call as they walk from one conference room to another, or are they stationary? Are you only concerned with offering web surfing capability? Are you trying to deprecate your wired infrastructure in your offices? If the VAR isn’t asking these kinds of questions to help guide your wireless needs and tailor a report to help, you need another partner.

You’re spending a lot of money to get information. The old adage, “Garbage in, garbage out,” is something to remember when hiring out this kind of work. I always loved working with my clients to figure out the best approach for wireless connectivity. I remember being on the customer side and when wireless was spotty, it was more frustrating than anything else I’d deal with. Mainly because my other IT colleagues would give me grief for the crappy network! If you have a good partner, work with them and define all your pain points, future vision, and needs. A good partner wants you to succeed!

With that information, your partner ought to provide a thorough report, and, explain the details and recommendations to you. My reports follow a similar framework of this is what I did, this is what I saw, this is what I recommend you do to meet your goal. Multi-tenancy for wireless is very difficult since the air is like the Wild Wild West. There may be things you can do to mitigate your performance issues, and the report should spell out those recommendations. At the very least, you need to see images of signal strength, signal to noise, and channel overlap. This is the absolute minimum! But, your VAR should provide a detailed report and time to review remediation recommendations.

The Heat

In the report you’ll see all kinds of images in the heat map. You would have provided your floor plans and the RF information will be overlayed on that image. You’ll be able to see coverage gaps, signal-to-noise images, and RF interference issues. Don’t let this clutter your independent understanding of their analysis. This report is valuable information but if you don’t know how to interpret it, you can’t tell if the recommendations from your VAR is accurate or they are looking to upsell you.

If you have channel-overlap in your building and are running an enterprise-level WLAN, like Cisco, Aruba, or Juniper, it’s likely you have an issue with the adaptive radios. These devices sample the airwaves and will change RF channels to limit the interference. This is more difficult, again, in multi-tenancy, but if you own the building and are having channel-overlap problems there may be other issues with the device configurations. If you’re VAR doesn’t call out fixes based in logical data, that’s a problem. Your partner should help you understand those nuances as they review the findings report with you and your team.

No Interference

Surveys are incredibly important, especially as the WLAN is bumped up the criticality scale. It used to be wireless was the redheaded step-child of data networking. Now, not so much. We all need mobility as we move from one meeting to another and one client site to another. This means wireless is not relegated to the kid’s table anymore; it’s a mission critical enterprise service. That means it needs to be treated like that and monitored for issues and improvement.

A survey is a good investment into your wireless infrastructure. It makes your enterprise customers happy, their boss’s happy, and senior management happy. Remember to consider your wireless needs and get the most out of your survey investment. It’ll pay off in the long run.