Tuesday, December 9, 2014

The Tablet Problem...

No, I'm not talking about my ex-wife's habit (we'll save that for another glorious installment of 'Crap Curt Thinks But Should Never Write'); rather, it's a statement outlining a major issue in supporting wireless communications (or wifi, if you prefer) in an iPad-rich (and the like) environment.

As you may know, I'm responsible for the wireless infrastructure (amongst other things) for one of the largest school districts in Georgia. We were one of the first in the nation to support wireless networking - we rolled-out (pun intended) mobile wireless notebook labs in 1998 - and started offering building-wide wifi in 2008. During this time, our number of users has more than doubled...we now support more than 45,000 people on our network (students and staff, combined). This is not an insignificant number.

When we first designed our building-wide wifi infrastructure, notebooks/laptops were the predominate mobile access technologies of the day. Smartphones, such as they were, were relatively rare...and tablets? What tablets? So our hardware placement was designed for notebooks/laptops...which have fairly robust antenna hardware. Computer hardware worked very well in this environment, with speeds no slower than 11 Mbps...though most were near the theoretical limit of the 802.11g standard of the day at 54 Mbps. Life was good, and users were happy.

Then, in April 2010, Apple introduced the iPad to the world. Though I thought the iPad nothing more than an oversized iPod Touch, the rest of the gadget-buying population disagreed with with me (upon actually using one, I had to re-evaluate my opinion - I love my iPad) and purchased them by the truckloads...and brought them to school to use on our wireless network. Good, right? No...not good. Not good at all.

You see, these wonderfully portable devices (along with the proliferation of usable smartphones) sacrifice some things to be so portable...and one of those things is antenna performance. While a laptop may have multiple antennas to ensure the best possible speeds, tablets and smartphone may only have one (to save space), which means a signal must be stronger to get the same speed. Considering that they're more susceptible to outside interference, and that the increasing numbers of "sacks-of-signal-killing-water" - also known as 'students' - negatively impact signal propagation, performance really began to suffer.

Step away from the consumer electronics...
Unfortunately for our users, our infrastructure was designed to support hardware with a greater antenna diversity than the hardware the users were now bringing on-site. You might be thinking: "Just run to the nearest electronics store and pick up another access point. Geez, dude, they're only like $75-$150 and work GREAT! It's not that hard." Well, yes, it IS that hard. When supporting large numbers of users, consumer-grade hardware doesn't work. It just doesn't. So we have to get enterprise-grade hardware...which isn't cheap. When I say "isn't cheap", I mean "in the thousands of dollars". Each. Times a couple hundred. Plus cabling. Plus additional network switches. Plus channel planning to reduce overlap/interference. Plus, plus, plus...

So, as you can see, increasing coverage for these new and wonderful devices was not, and is not, a trivial undertaking. We are currently evaluating the most cost-effective way to provide greater coverage, and faster speeds with the latest wireless tech, to our users with funds from a recently-passed bond initiative. Though I know they'd like to see us provide enhanced coverage sooner than later, it's imperative that we carefully consider the changing landscape of wireless tech and user devices else we run the same risk of designing for today rather than for tomorrow.

As we approach one-to-one devices on our network (we support more than 40,000 unique devices on our network in a 24-hour period) and move into many-to-one (meaning each user has more than one connected device), planning a robust and fault-tolerant infrastructure becomes even more important, if not downright critical. Providing wifi is no longer a secondary consideration; in fact, I might go so far as to say it's taken center-stage as the primary access method. Wires are soooo 2005, you know?

We've encountered many hurdles across the years supporting such a large wifi installation, some we've solved some with great aplomb and others continue to irritate us to this day. Maybe I'll address some of these in the future. Do you have any stories or questions? Be sure to let me know in the comments below.