WLAN Design: Learning From Our Mistakes

has the weakest antenna. To support this requirement, we have to take the APs out of the hallways and install them in rooms -- usually in every other room, even in residence halls.

This required us to start doing honest site surveys, meaning we had to create a rig that could get our test AP near its most likely installation location. It also meant we had to go out into the residence halls during peak usage (read: after normal business hours) to get an idea of the actual RF environment that our users experience.

This approach has done wonders for our department. Now we can design around potential holes in coverage, and we can reliably predict what the user experience is going to be.

It's also been a great teaching tool. Students are naturally inquisitive when they see you walking around the halls after hours with a rig and an AP. They perk up when you tell them that you are trying to "fix wireless."

I usually end up with an audience of 8-10 students at a time. I can explain some of the intricacies of wireless networking. I get to tell them how their rogue access points -- personal wireless routers, wireless printers, hotspots, and so on -- are hampering the wireless network. They also get to see that we are taking their complaints seriously, and that we want to make things better.

These after-hours trips are effective on several levels. The students genuinely appreciate the fact that I took time after hours to address their concerns. I get to throw some propaganda their way, and our group comes away with honest data to redesign a suitable network.

One significant drawback here is the sheer cost of purchasing, cabling, and installing all these APs. We had to make some creative decisions to minimize the costs as much as we could.

For instance, all new construction on our campus is going to feature 802.11ac, and all academic areas will see 802.11ac in the next two years. We chose to do this now with "Wave 1" products because they still provide better throughput and device density than 802.11n, and they cost about the same. All the 802.11n technology from these areas will be repurposed in the residence halls and will be supplemented with new 802.11n where needed.

We also won't install wiring in buildings that have a planned renovation within the next three years. This way, we won't be wasting money on wiring just to have it ripped out in a few years. Plus, it's always cheaper to pull wires when the ceilings and walls are open. Instead, we're installing APs over wall plates. We run the risk of APs being damaged, but we haven't had many issues with that to date.

We also plan to install two wires per AP for all new wiring installations in case 802.11ac APs, or future technologies, really need the second uplink. For us, the cost of the labor associated with wiring far outweighs the cost of the wiring itself. It's cheaper to pull a second wire the first time, even if it never gets used, than having to pay for labor a second time to add additional wire.

Finally, we are installing gigabit switches with 10 GbE uplinks in all locations that will see 802.11ac upgrades in the near future. We'll use gigabit switches with gigabit uplinks for all other locations that require new switch gear. Unless you are deploying 802.11ac or you have a high AP density per switch, 10 GbE still isn't cheap enough at the distribution layer to warrant buying a ton of switches with 10 GbE capabilities that will likely never be used.

The choices that we made are projected to save us more than $1 million in capital infrastructure costs over the next three years, though it will slow our adoption of 802.11ac. This is a tradeoff we decided we could live with.

In its ninth year, Interop New York (Sept. 29-Oct. 3) is the premier event for the Northeast IT market. Strongly represented vertical industries include financial services, government, and education. Join more than 5,000 attendees to learn about IT leadership, cloud, collaboration, infrastructure, mobility, risk management and security, and SDN, and explore 125 exhibitors' offerings. Register with Discount Code MPIWK to save $200 off Total Access & Conference Passes.