WiMax hits the big time

10.08.2006
Sprint's announcement this week that it's going to deploy mobile WiMAX in the 2.5 GHz. spectrum that it owns around the country surprised many.

WiMax is a true fourth-generation (4G) wireless technology creating a huge number of opportunities and questions. The answers to those questions will have a profound effect on enterprise mobile computing strategies going forward. (If you've not looked at WiMax yet, you can start with the WiMax Forum's site.

Let's start with the basics. First, while there is no widely accepted definition of 4G, Farpoint Group defines it as a wireless, mobile, broadband, all-IP network or service with support for time-bounded traffic (QoS). WiMax meets this definition. All current cellular services are based on SS7 time-division multiplexing (TDM) technologies that are suitable for voice, but difficult for data; 3G data services are distinct in current cellular implementations, but WiMax puts everything on IP. Interestingly, so does Wi-Fi; yes, metro-scale Wi-Fi is 4G. Note, by the way, that I don't place a throughput constraint on the above, but let's assume a lower bound of 1 to 2 Mbit/sec. Interestingly, I expect up to 100 Mbit/sec from metro-scale Wi-Fi using MIMO, although this approach is just beginning to see deployments.

But WiMax is not, as I continue to read to the contrary, Wi-Fi on steroids. In fact, it has nothing to do with Wi-Fi. WiMax is a big-cell, licensed-bandwidth technology; Wi-Fi is a small-cell, unlicensed technology. They are, in fact, quite complementary. While I still expect to see converged cellular/Wi-Fi services as the most popular option to provision future broadband wireless networks, the Sprint announcement forces us to think in terms of converged WiMax/Wi-Fi and even dual mode WiMax/EV-DO services and devices.

This is because Sprint is one of the two major suppliers, the other is Verizon Wireless, of EV-DO broadband services. In fact, they are in the middle of a major rollout of EV-DO Rev A, which boosts downlink speeds to 3.1 Mbit/sec., and the uplink to 1.8 Mbit/sec. WiMax, of course, uses 16 times the bandwidth of EV-DO in it current form, 20 MHz. vs. 1.25MHz, and thus has much higher peak through -- to about 75 Mbit/sec. But keep in mind that, being a big-cell technology, there could potentially be a large number of users sharing that 75 Mbit/sec. I've believed for some time that 5 Mbit/sec is a good upper bound on what users might see, and I think that holds in Sprint's case. It's worth nothing that the upgrade path for EV-DO includes a "Rev B" specification that offers performance in line with what WiMax can do. But it's safe to say that WiMax will get there first.

But that's still at least 18 months away, and the confusion factor in Sprint's announcement will need to be addressed. Suppose you're an IT manager and you want to roll out data services to the thousands of people in your field force. Will you go with Sprint's EV-DO, wait for WiMax, or perhaps consider the competitive offerings from Verizon (also EV-DO), Cingular (HSDPA, with up to 14.4 Mbit/sec), or perhaps something else, like metro-scale Wi-Fi? Sprint says they will continue to support EV-DO, but should a customer buy that, or is WiMax the way to go?