Last year at MWC in Barcelona, Intel revealed their plans for 5G network deployment, speaking of a data network that wouldn’t just be faster, but would be better at handling the load of the estimated 50 billion connected devices that will be in service by 2020. This year at CES 2017, Intel backed up their talk by announcing their 5G modem — the latest salvo in what could become tech’s most heated and important competition.
The modem is the kind of thing that will go into 5G-compatible devices of the future. Intel says they’ve implemented wide-band 28 GHz mmWave and sub-6 GHz connectivity, with support for “low latency frame structure, advanced channel coding, massive MIMO and beamforming.” That all sounds a bit arcane, but those features define what the next generation of network will look like.
It won’t just be faster. Of course it’ll be faster — Intel says they’ve achieved transfer speeds in excess of 5 Gbps, while the fastest 4G LTE networks in the United States are around 500 Mbps. But, speed isn’t the issue 5G is trying to address. If the estimate of 50 billion connected devices that Intel cited last year proves accurate, that’s going to be an enormous load for networks to manage — one that would overwhelm current 4G LTE networks. That’s why Intel is implementing massive MIMO (multiple input, multiple output) and advanced channel coding, while adding support for many new bands — so smart devices can have an optimal connection that maximizes speed, while the load is spread evenly across all those bands.
If all goes according to plan (big if), 5G networks would start rolling out this year in very early stages, followed by a gradual advance toward full rollout by 2020. What’s the rush? Well, we’ve been hearing a lot lately about drone delivery and autonomous vehicles. Right now, both are being tested in remote areas, which doesn’t mean much to the network they’re using. But, imagine a full drone delivery program from Amazon or a fleet of autonomous cars from, say, Uber — that’s thousands, potentially millions of devices that need a stable connection and potentially a lot of data. Add that to the millions of smartphones on any given network, and it’s a given no one will end up happy.
Drone delivery and autonomous cars are a big reason why making 5G low-latency is so important. Today, 4G LTE networks by and large are tasked with connecting our smartphones to the internet. If it’s a bit slow delivering a livestream, that’s not optimal, but it’s not the end of the world. For autonomous vehicles, it’s another story entirely. Connected cars of the future will be able to communicate with smart cities equipped with sensors — a city’s smart system could tell a car that a cyclist is nearby, ensuring the safety of everyone involved. But, if there’s lag, the cyclist will actually be ahead of where the car thinks it is — not an outcome that can be allowed.
Part of the way this would work is by reserving bands for dedicated uses. Currently, devices connected to 4G LTE are all muddled together. With 5G networks, it’d be possible to reserve a band just for use with cars, one just for use with delivery drones, and so on. This dovetails with the idea of using smaller cell towers placed all around the city. Instead of everything being routed through huge cell towers, more localized networks could be used, handing off connections to the next one seamlessly once a device moves out of range. Using software, those connections could be managed and prioritized — for example, cars would always be given top priority for network speed, while web browsing would be down near the bottom. In fact, for mmWave bands — the higher speed, shorter range bands — to be viable, this kind of localized setup could be necessary.
By releasing their modem, Intel puts themselves right in the middle of the fight to dominate 5G. They won’t be alone, and they’ll have some familiar rivals. Qualcomm announced their first 5G modem last year, which also uses mmWave technology. Qualcomm came to dominate in 4G modems, to the point where Intel failed to get traction for their mobile chipsets. Samsung is said to be working on a 5G modem, as well.
It’s still early, and these competing modems from Intel and Qualcomm won’t necessarily reflect what 5G modems will look like when they start appearing in devices regularly. A big reason for that is that 5G hasn’t actually been defined yet — 3GPP (the 3rd Generation Partnership Project) is currently in talks with networks about when standards will be established. Until then, Intel, Qualcomm, and the carriers alike won’t know what they’ll be able to market as 5G. The fact that 3GPP works this out is great for consumers — otherwise, any company could create marginally better tech and call it 5G.
It’s going to be a very tumultuous time in tech. For a lot of companies involved, the move to 5G can be seen as a partial reset. When 4G was being established, Intel and Sprint invested heavily in WiMAX, a network standard that was beaten out by LTE — something supported by AT&T, Verizon, T-Mobile, and Qualcomm. It’s no coincidence that Intel has struggled in mobile while Sprint has fallen back into fourth among the major U.S. carriers. With a 5G rollout, Intel and Sprint could be poised for a comeback if they play their cards right. For their part, Sprint owns more spectrum than their rivals, and happens to have a lot of the high-frequency bands that work well with smaller towers — the kind that will be perfect for 5G.
Whoever wins, what consumers should see in the end is the same. 5G is the network that will connect everything to everything. It’s the groundwork that needs to be laid out so that the next great tech transformations — drone delivery and autonomous cars — can actually be viable on a large scale. Thing is, when you start talking about cars relying on connectivity, it’s easy to see why things need to be perfect. It’s not going to be easy.