Just a few years ago, wireless network design was a whole different ballgame. Sophisticated modeling tools from AirMagnet and Ekahau didn't exist, and staffs planning WLAN environments needed to understand how a given antenna would shape a specific amount of transmitted radio frequency (RF) energy and apply the model to each deployment scenario. It was pretty heady stuff. People who "got it" tended to build effective networks, while those who didn't caused a lot of problems for clients with questionable designs. Understanding antennae and RF behavior, or not, was usually the difference.
Fast forward to today, and the wireless industry offers all sorts of graphical planning tools and predictive utilities that promise to demystify wireless network design. These are handy, but alone do not make for guaranteed WLAN success. High-capacity wireless environments are becoming commonplace, where hundreds or thousands of users may all use Wi-Fi in a given space, and complex wireless scenarios (outside, stadiums, retail, and so on) still require knowledge of how to manipulate cell sizes and shapes with different antennae to achieve success, despite the magic built into even the best planning tools.
Most major WLAN vendors feature a slim-profile "flagship" access point with "captive antennas" that aesthetically blend well with most environments. These units are easy to deploy and reduce costs by not requiring add-on antennae, but also constrain design to adding more APs where additional coverage is needed, as each unit has a set omnidirectional pattern that casts signals in all directions. Often, multiple additional APs are needed to fill a long corridor or large space where these captive antenna APs are used. By contrast, a single AP with the proper external antenna capable of shaping the signal into a highly directional pattern might provide the same coverage at a fraction of the cost (if expected user density can be serviced with a single AP).
In multitenant spaces, directional antennae can be the difference between my company's WLAN staying mostly between my walls or being blasted out to the neighbors, where we both suffer decreased performance as a result. Again, "omnis" are cheap and easy, but they aren't always the best choice. Even consumer-tier wireless product lines like Linksys' now offer directional antennae, or the ability to accommodate third-party antennae, as wireless gets ever more pervasive and different use-cases are realized in the enterprise and SOHO spaces alike.
What kind of impact can just changing the antenna really have on a wireless cell? Where a typical consumer-grade wireless access point might have difficulty effectively covering even a medium-size home, swapping out its antennae for highly directional alternatives can add miles to the coverage--potentially lots of miles--but in only one direction. Web searches for WiFi Shootout projects reveal just what sort of insane distances can be covered by highly directional antennae--for fun, in the case of hobbyists who like to roll their own antennae, and for business, where long-range bridge links are needed.
In my day-to-day wireless work, I find myself leveraging the likes of Cisco 3502P access point coupled to its impressive "stadium antenna" that delivers coverage to a single seating area in a large stadium if used right; the Exalt r5005 outside bridge that provides highly directional connectivity to far-away buildings; and simple omnidirectional APs that do well in many indoor spaces. Then there are my sector antennae that cover parking lots and semidirectional antennae that fill the aisles of a warehouse with signal. Each case has an antenna option that works better than others, and only by understanding what a given antenna can do for you can you reap the benefits.
But antenna technology is also a place where WLAN vendors go to differentiate themselves from the competition with performance claims that may or may not make sense for those shopping for coverage. Xirrus, for example, approaches WLAN coverage quite a bit differently, with an architecture that promises a 75% reduction in the number of access points required to cover a given space by using "arrays" that are essentially multiple access points and antennae slathered in special technical sauce and bundled in a single enclosure. If I buy in, then where I use 80 classic access points to cover a residence hall, I should be able to get away with 20 Xirrus units as an alternative. I'd really have to sit down and analyze what the antenna coverage in both options would provide before I could accept or reject Xirrus' unique approach.
Ruckus is another vendor that plays up its antenna technology as a reason for customers to take a second look. Claims of "over 3X increase in performance and range" and "8X expanded coverage" from the Ruckus BeamFlex product line get your attention, but they lack context to me and are in the category of marketing hype for which I don't care. Show me impressive coverage patterns based on the same output power as the competition, but don't require me to read a white paper on a dozen patents that combine to deliver RF magic that the average customer will never understand. The truth is that in capacity-driven networks, performance trumps range, and to talk about them both at the same time can be mighty confusing to even the most RF-savvy. I don't reject Ruckus's claims, but I know they are a lot more complicated to digest than those from others.
Antennae are amazingly simple in many cases, and fundamental antenna theory has changed little since the days of Marconi. At the same time, there are many antenna options in the modern, complex WLAN, and a lot of complicated technology upstream from the antenna feed. The more antenna basics you understand, the better you'll be at design, purchase or support.