I had heard about the fact that modern cars are designed to crumple upon impact. I watched Richard Hammond and other preeminent scientists gesture expressively at the unrecognizable chassis of a late-model car, rhapsodizing about why it was better for the cabin that force is dispersed away from the core of the structure as opposed to transferred directly to the driver. “Sure, I get it… that makes perfect sense. In fact I might even agree with it,” said my normally reasonable brain.
What I didn’t anticipate when I was involved in a minor car accident last week was that a modern collision could be completely invisible. Not the damage mind you. The damage to the car was very real, but I didn’t feel a thing at the driver’s seat. “That sort of damage can easily happen in a garage nowadays,” the body shop owner somewhat morosely mumbled, as we both stood side-by-side staring near our feet with that very grave Minnesota demeanor of a appraising a job that needs doing. But I’m still a little dumbfounded that the collision itself felt nowhere substantial enough to warrant the two-page estimate that now intrudes on my desk.
The truth is… I should not be as surprised. From a wireless networking perspective invisible collisions happen everywhere, constantly. After all, air is a shared medium and every device wants to talk and be heard. Collisions and retransmissions are the most difficult problems a wireless network faces after implementation.
If you could physically listen to the 2.4 GHz spectrum, you would probably hear something akin to a school cafeteria at 12:15 PM. There’s the din of hundreds of kids talking all at once. There’s the piercing ringing of class bells and the muffled announcements of a shoddy 1970’s era public address system. Above all of that, there are the swelling screams of a group of kids as they have to consistently increase their own volume just to be heard by their friends. The noise floor undulates with activity and density, and is changing itself every second.
A similar effect occurs with transmissions in the RF spectrum, but with a very important difference. With 802.11-based WiFi, all devices agree upon a set of protocols which attempt to avoid these collisions that cause devices and access points to misunderstand each other. But there are many factors that can cause these collision avoidance protocols to fail. The average user has no idea these collisions occur, but again, the damage is very real. It can manifest itself in slow performance, dropped connectivity and inconsistent data rates, among other issues. If I’m a user, I grumble that “The WiFi sucks here,” and get frustrated with my loss of productivity. Further, with the explosion of mobile devices, we’ve collectively just doubled or tripled the number of kids in the cafeteria overnight. More generally, If you double or triple anything overnight, whether it’s cars on a road or wireless devices on a network, the infrastructure will bend if not break.
Wireless collisions aren’t completely invisible, however. We have the tools and the expertise to troubleshoot these problematic networks. In fact, it’s a personal mission of mine to eradicate the “The WiFi sucks here” reflex. It doesn’t have to. Even in the most challenging of environments there’s always a design that reaches the goals of what organizations wants to deliver. The goals need to be set appropriately, however, as designing simply for coverage is no longer adequate. Organizations of all types need to be able to have the wireless network capacity to deliver high-performing services to three times more concurrent devices than five years ago and be able to perform even in a noisy, chaotic environment.
That means keeping these invisible collisions to a minumum, because they can result in very significant and visible damage to productivity.