Makeover for mobile phones

AFTER years of hot air and hyperbole, the fifth generation (5G) of mobile-phone technology has entered its final phase of testing, in preparation for its debut around the world. The Third Generation Partnership Project (3GPP), an industry group for mobile phones, has still to sign off on a 5G reference design that satisfies all its members. But that has not stopped manufacturers from introducing 5G chip-sets and modems for wireless carriers to test. The hope is to get 5G mobile networks up and running in time, at least, for the winter Olympics in South Korea in February 2018. Japan has its own plans for the technology when it hosts the summer Olympics in July 2020. Expect wireless carriers to start rolling out their 5G networks in earnest shortly thereafter.

The race to launch 5G is reminiscent of the rush to do likewise with Wi-Fi in the 1990s, when equipment makers hurried out interim gear both in order to influence emerging standards (and thus lock in their own particular patents) and to get a foot in the door ready for when the technology took off. Which it duly did. Many believe 5G could be an even bigger change. Hype aside, the technology is more than just a faster, better and more efficient network for mobile phones. It makes many things possible that were previously just pipe dreams.

What distinguishes 5G from earlier generations of wireless broadband is its ability to send and receive signals almost instantaneously. The “latency” (ie, the lag between initiating an action and getting a response) that has hobbled mobile phones will be a thing of the past. When 3G phones were the bee's knees, the time taken for two wireless devices to communicate with one another was around 500 milliseconds. That half-second lag could make conversation frustrating. A decade later, 4G had cut the latency to 60 milliseconds or so—not bad, but still an age when waiting for crucial, time-sensitive data, especially from the cloud.

With a latency of less than a millisecond, 5G will make possible all manner of applications that require a rapid response. Sending hazard-warning signals to self-driving cars threading their way through traffic, for example. Or providing instantaneous language translation while someone is conversing over a phone. Or permitting a brain surgeon to manipulate a scalpel remotely from afar. Or streaming ultra-high-definition video to mobile devices without having to buffer the data. Or providing people with virtual rides at home that mimic the experience in a real amusement park. Futuristic, may be, but do not bet against any of these.

New generations of mobile-phone technology have appeared roughly once a decade, starting with the 1G networks for analogue phones in 1981. These were followed by 2G in 1992 and 3G in 2001. The present 4G networks were launched in 2009. Each new generation is marked by the promise of a giant leap in performance, and the threat of a technological break with the past—ie, no backward compatibility. In reality, generations tend to merge, almost seamlessly, into one another, with not all that much, at least initially, in the way of a boost in performance.

When 4G was introduced, for instance, the wireless technologies adopted (Mobile WiMAX and Long Term Evolution, or LTE) offered nothing like the peak download data rates of 1 gigabit per second promised at the time. Turning a blind eye, though, the International Telecommunication Union, the UN agency that rules the radio spectrum, agreed to let wireless operators brand their interim technologes as 4G on the understanding that, over time, they would evolve into the real thing. Most failed to do so. Today, the main survivor, an advanced version of LTE, delivers around 12 megabits per second in America (more elsewhere), and up to 300 megabits per second under special conditions.

Seven years into its ten-year life cycle, 4G has only recently begun to live up to expectations. Its latest iteration, LTE-Advanced Pro, can achieve download speeds approaching one gigabit per second, albeit at a price. So, expect 5G to follow a broadly similar trajectory, starting off with perhaps one gigabit per second and topping out at ten gigabits per second or more as the technology ripens with age.

But if it lives up to expectations, 5G wireless could put some fixed-line internet connections to shame, even at the lower end of its performance range. As such, it could spell the end of wired connections in the street and around the home and workplace. Fibre will still find a role in hauling traffic back from wireless base stations to central offices. But not having to dig up the road to lay “last mile” fibre or coaxial cables for delivering internet access and high-definition television to individual addresses will be a relief for telecoms firms. Beaming such services wirelessly from nearby base stations instead should reduce costs significantly.

None of this will be possible, however, without fresh chunks of spectrum being allocated to the task. In particular, if it is to act as a replacement for fibre, 5G wireless will need to operate at frequencies of 20-60 gigahertz. While these “millimetre waves” provide channels wide enough to support multi-gigabit speeds, they have limited range and cannot pass through walls. Base stations will have to be within sight of one another. They will also need to be placed close together, as a mesh of "micro-cells".

If 5G is really to achieve its full potential, though, two techniques already used widely in Wi-Fi and LTE will have to be pushed to the limit. One is “channel aggregation”, which works by combining wireless signals from several base stations into what is effectively a big, fat pipe capable of beaming data at a far higher rate than would otherwise be possible. Apart from boosting download speeds, carrier aggregation lets mobile operators patch together disparate chunks of spectrum they have acquired piecemeal over the years.

The other technique, “multiple input/multiple output” (MIMO), uses a number of antennae to transmit separate streams of data, with multiple antennae at the receiving end processing all the incoming signals. Today’s MIMO systems tend to have about four antennae at both the transmitting and receiving ends. But there is nothing in theory to stop them from having hundreds or even thousands. The downside is complexity and energy consumption. The upside is reliability, throughput and freedom from interference or jamming. The challenge is to make lots of cheap little radio amplifiers work together as a coherent whole, while adapting seamlessly on the fly to meet the needs of different users.

One way to do this is to slice the network into multiple “logical” networks, each optimised for a different user’s needs. In network slicing, each virtual network is tailored for the device being served at that instant, while all remain part of the same 5G network. Thus, a smart-phone user would get a virtual network optimised for a high data rate; an autonomous car would connect to a low-latency virtual network; and a household thermostat, light switch or other “internet-of-things” device would make do with a virtual network having lower overall performance. Meanwhile, unexpected crucial communications—life-or-death warnings, for instance—would need to interrupt the process and get priority. One trick for doing this is known as “puncturing”, which lets a network listen for mission-critical calls while going about its normal business, and then quickly allocate a connection.

Piling on all these requirements is going to make 5G networks horrendously complicated. The constant reconfiguration will demand huge amounts of computing power—and not just tucked away in distant data centres, but distributed throughout the network itself. Base stations will come to resemble servers as much as radio transceivers.

As this column has noted before, a great deal of haggling by national carriers, telecoms firms, equipment makers and standards bodies remains to be done before 5G can become a practical reality. But the hope still burns that, unlike previous generations of mobile technology, 5G could be a true global standard. If that comes to pass, travellers will be able to use their mobile phones anywhere in the world, without the hassle of having to swap their SIM cards the moment they arrive. And they will be able to use them to do things barely imagined today.