Giz Explains: Why WiMax and LTE Wireless 4G Data Will Blow Your Mind

Illustration for article titled Giz Explains: Why WiMax and LTE Wireless 4G Data Will Blow Your Mind

3G sucks. Yeah I said it. Try watching YouTube video or hell, loading Giz. Real wireless, ubiquitous broadband for slurping up crazy data anywhere, anytime is coming. Soon. In the form of WiMax and LTE.


We're going to try to keep this pretty simple, as usual, but there are going to be some acronyms and a bit of jargon involved—our previous explainer on mobile terms might be a good place to start, actually, if you're walking into this totally oblivious to mobile tech.

Quickly, though, the current state of mobile networks is that we use 2.5G and 3G networks—mid-second-gen and newer third-gen data protocols. On the Verizon and Sprint side, known as CDMA, 2.5G is referred to as 1XRTT, or just 1X. On the AT&T and T-Mobile side, GSM, the 2.5G flavor is EDGE. Verizon and Sprint's 3G is EVDO, while AT&T and T-Mobile have HSDPA (you might not know that one, since they usually just say "3G").

Second gen wireless was basically just the leap to a digital network, and third gen is a closer attempt at true mobile broadband—kind of. Right now, with their 3G networks, they can all get you typical speeds of around 1 Megabit per second downstream, give or take (though the specs are rated for peak speeds of 3Mbps down on EVDO Rev. A, and 3.6 on HSDPA). 3G has a bit of breathing room left in it—EVDO Rev. B is capable of downstream speeds of 14.7Mbps , while the current HSDPA spec will go up to 14.4Mbps downstream with the right equipment, and depending on how far down the HSPA spec sheet you wanna go, maybe even faster.

But the fourth generation is already on its way. Technically, no wireless technology is officially 4G. But that's what everybody's calling WiMax and Long-Term Evolution, because they both promise crazyfast mobile internet speeds that leave the current 3G in the dirt. In the US, the main WiMax player is Clearwire, which Sprint owns 51 percent of after they combined their operations into one company and actually gave WiMax a chance to live. LTE is championed by AT&T (which makes sense because it was developed initially by companies who mainly build GSM networks like AT&T and T-Mobile's). Verizon also selected LTE, which blew everyone away at first because Verizon isn't in the GSM camp, but it makes sense because Verizon's parent company, Vodafone, is gung-ho for LTE in Europe, where everyone's on GSM.

So here's the crazy thing about WiMax and LTE, which you might not realize from all the smack talk coming out of Verizon and AT&T. I'm probably going to blow your mind right now: "They both use the same fundamental technology," says Barry West, Clearwire's President and Chief Architect. They both use orthogonal frequency-division multiplexing access and they're both IP (internet protocol) based. More simply, you can kind of think of the difference between WiMax and LTE as a software, not a hardware thing (kind of like Macs and PCs using the same Intel chip). Alcatel-Lucent, who makes the 4G wireless hardware, is actually "building hardware that is on a common platform," Paul Mankiewich A-L's Wireless CTO told us. In fact, West told us, at "some point in the future it's possible to harmonize" LTE and WiMax, it just "requires people to be willing to do that."


Here's what the fundamental difference is: Time division duplexing versus frequency division duplexing. Sounds complicated! But it's not. AT&T Labs VP of Architecture Hank Kafka explained it like this: "TDD is like CB radios or walkie-talkies—when one person is talking, the other person can't talk." The same channel is used for downstream and upstream, so the transmission is divided up over very tiny increments of time. Clearwire's West says they currently use a 2/3 downstream and 1/3 upstream split, so 2/3 of the time, you're swallowing data, and 1/3 of the time, you're spitting it. With LTE, Kafka says "it's more like a modem or phone conversation." It separates the available bandwidth into two parts—one operating downstream full time, and one operating upstream—so "you both can talk back and forth at the same time."

Great. But what's so special about WiMax and LTE? And how fast can they really get? Very simply, West told us, "The magic is the channel width." LTE and WiMax use really fat wireless channels, so they can move a lot of data at once. For example, AT&T's Kafka told us that "peak speed for LTE in 10MHz is about 140Mbps and peak speed in 20MHz is about 300Mbps." The thing about them being OFDM is that it makes them more flexible than 3G, since they can use a wide range of spectrum—LTE can use anything from the 1.4MHz channel up through 20MHz—whereas current 3G always uses 5MHz.


Did you see that? 300Mbps? Over the air? Whoooa. Well, don't let your panties get blown away yet. Yes, 4G will be way faster than 3G. But don't expect Asian city internet speeds wirelessly in the next couple of years. Clearwire's Barry West throws a bit of cold water on the ridiculously scorching speeds you might see hyped for LTE: To get to that 170Mbps, "that's like 8.5 bits per hertz and I've never seen a system achieve more than 5 bits per hertz." Huh? Basically, it doesn't take a whole lot of interference to slow your connection down, because it and WiMax use a complicated modulation scheme that you can't have constantly cranked to 11. So real world speeds will be slower.

WiMax is no slouch either, technically capable of up to 72Mbps.

Another thing about those superfat channels is that they don't reach as far out from the tower, and your response drops (obviously) as you get farther away. Which, Alcatel-Lucent's Mankiewich said, is one of the major infrastructure things with 4G: They're going to need to build more cell sites. That's why building out 4G is very pricey. (Not to mention all the money everyone had to spend on the right kind of airwaves to use for 4G.) If you thought 3G rollout was slow, 4G might be slower.


Here's what the real-soon-future looks like: Verizon isn't dicking around, and is doing commercial rollouts of LTE in 2010, while AT&T is following up with their commercial trials in 2011. (AT&T says Verizon "is in a big rush to move to LTE because their 3G technology gives them no room" to increase bandwidth and that red is a stupid color, nyah nyah nyah.) Clearwire has rolled out WiMax to a few cities already, and plans to have 120 million covered by the end of 2010. Verizon says they're getting about 60Mbps in testing, but expect it to be more like cable modem speeds when it launches—like Clearwire has now. For the reasons we mentioned above, and also because there won't be devices that can handle that kind of ridiculous speed—as you probably guessed, battery life being a major reason.


Will one standard eventually beat the other into submission, slinking away into the night, arm and arm with Betamax and HD DVD? Well, LTE does have a lot of momentum—the two biggest carriers in the US are rolling with it, and as part of the GSM family, you can bet all of the GSM carriers all over the world will be on board. But Alcatel-Lucent's Mankiewich says, "there's no real technological reason to pick one over the other." In fact, he thinks no one will "win," and just like now where "multiple technologies exist for economic reasons," it'll be the same thing with WiMax and LTE. So our only real hope for a single, happy standard is that they get together and make sweet, sweet love with some Marvin Gaye crooning in the background. It could happen.


Still something you still wanna know? Send any questions about wireless, Wild Things, or why truckers wear trucker hats to, with "Giz Explains" in the subject line. Original photo up top by Anina Schenker


This is a nice writeup Matt, but you missed an extremely important element. The main problem with cellular network broadband is not the max burst speed but latency. So, you can download at a billion megs per second — that is nice if I am downloading a movie. That that does not help is if I am actually trying to get work done. To truly replace a broadband connection, the latency response rate needs to easily be under 100ms, and under 50ms for many applications. I am not saying whether 4g has that, but I know for a fact that 2.5g and 3g are not really cutting it on latency rates.

I think the major problem with latency on these networks is the distance the signal has to travel before it hits a cell tower. I hope they have found a solution to this (like sending the packets at faster than the speed of light?), but if they haven't, I still think 4g may only be marginally better than what we have now for people who issue laptops running Citrix / Terminal Services or trying to play video games through one of these connections.