Wireless site surveys and why you no longer need them
Wireless site surveys have always been part of a WLAN installation, but today that no longer should be the case. Capacity, not coverage, is more important.
The site survey -- specifically, the active site survey that calls for the use of a mobile detector to manually measure the strength of a temporary Wi-Fi access point -- has been a common element in wireless LAN installations ever since the first access points appeared almost three decades ago.
While many in the installer community continue to support wireless site surveys, the message today is quite the opposite: This labor-intensive and often expensive activity is, in fact, a relic of the time when access points (APs) were expensive, throughput and latency requirements were low, and the user base was limited to mostly vertical applications, like bar code scanning.
A site survey optimizes for coverage, and today, the goal is quite different. Enterprises want to engineer their wireless LAN (WLAN) systems for optimal capacity. This approach lets users get their jobs done without a slow network getting in the way. Instead of spending money on a site survey, companies can now invest those funds in a more strategic and appropriate goal.
Skip the site survey? Yes, in many cases, that's the way to go.
More options than walking around to measure radio frequency
Wireless site surveys are designed to help determine the optimal locations for production APs. Yet the active walk-around exercise required today isn't the only option.
There are applications that perform predictive modeling of radio frequency (RF) propagation. They can produce interesting results with one important proviso: The data entered -- regarding the site dimensions and building materials -- needs to be accurate; otherwise, the findings are worthless.
Companies should also consider a nominal RF sweep with a spectrum analyzer, which can check for interference and conflicting systems in unlicensed bands. However, much of the data obtained in a site survey, including spectral analysis, is commonly available today in management consoles and Wi-Fi assurance tools that provide continual monitoring of production conditions -- post-installation, of course.
Skip the site survey? Yes, in many cases, that's the way to go.
So, then, why not conduct an active survey? The reasons are numerous.
An active survey yields information from only a glimpse in time and is limited to the sampled locations. Sure, a great deal of effort could be devoted to a survey, but the result is still just a single, temporal snapshot. And surveys aren't cheap; they require skilled labor. The costs involved could range into the tens of thousands of dollars for larger facilities.
The survey does not consider production user locations or how these vary over time due to user relocation, building reconfiguration and especially dynamic mobility. Realistic application loads are similarly ignored, even though these absolutely influence the location of APs.
The survey ignores the RF spectrum and resource management capabilities built into every organizational-class WLAN product being marketed today. Such features as automatic channel selection, varying transmit power levels and adaptive modulation and coding rates -- also known as upshifting and downshifting -- are applied based on real-time conditions. Additionally, more sophisticated features, such as selecting the best channel; beamforming; load balancing; beam steering; band steering; multiuser multiple input, multiple output; meshes; analytics; and AI capabilities are similarly largely ignored.
Most installations today are upgrades, not greenfield deployments. So, significant information regarding real-world usage is already available via management consoles, many of which use analytics to further optimize results. While it could be argued that each new technology generation requires a new survey, that's not the case, as cited in the previous point. Failure to consider all the variables results in a flawed analysis -- every single time.
The survey says nothing about client behavior, which is highly variable. Each type of client -- with respective OS release and specific driver settings -- must be tested during the site survey. This would clearly be cost-prohibitive, and over the course of a given WLAN installation, users will see many iterations of software and hardware.
Today's Wi-Fi installations are much more fluid than they used to be -- with the variables changing over time. While it would be nice to understand coverage characteristics in advance, in reality, a WLAN system's capacity is more important. Thanks to dense AP deployments, analytics and the introduction of other adaptive technologies, it's easier to determine that.
It's also easy to conclude that active wireless site surveys contribute very little value. We'll examine alternatives for spending those site survey dollars more productively in part 2 of this series.
