Here's a look at the drivers and application behind 5G mmWave communications.
Recently, there has been one dominating buzzword in any publication about communication networks. It is 5G, for fifth generation wireless networks. While the current 4G LTE networks continue to be expanded and updated, the industry is hard at work defining the next leap in wireless communication. It promises 10Gb/s data rates and under 1ms latency for future users. The breakthrough is achieved by jumping to the millimeter wave frequencies, approximately 30GHz and beyond, where spectrum is plentiful but ill-behaved. I have earlier pondered on the implications for test equipment- higher frequencies, huge bandwidths, new protocols, and overwhelmingly modular due to the multi-channel nature of the beast. I stand by those predictions. 5G offers both opportunities and challenges for test equipment vendors. The opportunity is that new technology waves will enable large swings in market share in an otherwise static marketplace. The challenge is that the measurements are really, really difficult.
5G mmWave communication will rely heavily on beamforming, making mobile communications challenging. Are the results worth the costs? Image courtesy of Keysight Technologies.
5G measurements are difficult because the networks themselves face very difficult challenges. When I say mmWaves are ill-behaved, here's what I mean. mmWave face high attenuation, so active beamforming is a necessity to get the needed gain. On top of that, almost everything becomes a reflector- trees, lampposts, windows, people. So, in addition to simple beamforming, often the base station will have to perform sophisticated bank shots as a user walks behind a post or tree to make a connection. Even with all that, the maximum cell range is in the neighbourhood of 200m, basically two football fields. This means that cell density must be very high, and coupled to sophisticated hand-off mechanisms between cells. And yet, with all these difficulties, operators are racing to define, design, and deploy 5G.
Or are they?
It is with this question in mind that I attended the annual International Microwave Symposium (IMS), this year held in San Francisco. I was there primarily to look at what the test vendors were offering, and with what type of architectures. I will cover that in Part 2 of my IMS coverage. But I also wanted to check on the business case for 5G itself. After all, the investments made by service providers will be in the tens of billions of dollars. And that is only if they can deploy this difficult technology at all. I've seen technology cycles delayed (40Gb/s optical in the early 2000s), so it would not be unprecedented for the industry to take a pause between their major refresh cycles if the value proposition wasn't compelling enough.
So, wherever I went, I asked this intentionally provocative question: Given that peak data rates of 4G are already sufficient for almost all known applications, and coverage is the major dissatisfier for cellular users, why would a service provider commit billions of dollars for something so ill-behaved and only goes 200m?
Here's what I found...
My first stop occurred at a breakfast press event hosted by Keysight Technologies just outside the conference. The keynote speaker was Mark Pierpoint, VP and General Manager of the Internet Infrastructure Solutions Group at Keysight. In full disclosure, I've known Mark for many years, and worked with him before I retired from Agilent. Mark gave a very engaging and thought provoking presentation about new developments in communication networks. Examples ranged from low power ubiquitous communications for IoT (Internet of Things), to satellite-enabled services, to a mmWave bandwidth explosion. While all have sometimes been referred to as "5G", it is the latter that I truly equate with 5G. At the end of the presentation I asked what are the specific applications that require this phenomenal bandwidth that would justify this massive investment. While Mark had hinted at many applications in the long-run (e.g. automated semi trucks arranged in a train line on a freeway), he selected two that would need the bandwidth right away- stadiums and virtual reality. The stadium application is real. There is a very high density of users, and many sporting events cause the cellular network to collapse to a snail's pace. Stadiums, conventions, and indeed everywhere there is a high density of users, the 5G value proposition seems valid. Toss in extra features, like personal instant replay or views from a different perspective, and 5G could be critical in supplying new experiences. It is no coincidence that Samsung has promised their first 5G public demo will be deployed at the 2018 Seoul Olympics. But virtual reality? True, real time VR systems require a lot of bandwidth, but what are the applications? Is the experience that much better than, well, watching a video? In all honesty, I had never put on a virtual reality headset, so I was sceptical. Fortunately, the IEEE was offering VR demo on the conference floor, and I jumped at the chance to experience this myself. Donning the headset, I and two other volunteers sat in chairs that simulated pilots' seats of a spacecraft. I looked above, below, and behind, and saw that I was in a detailed realistic spacecraft alongside two other suited space pilots. Below and to my left was a cup of coffee- still steaming, with the IEEE logo printed across it. We were given two minutes of instructions on how to use the controls before the real simulation would occur.
Here the author experiences VR (virtual reality) for himself. It may appear he is on the floor of a conference, but he is actually piloting a shuttlecraft across the surface of Mars. Fortunately, no people were harmed in the simulation, nor any virtual coffee spilled.
Then the instructor said the simulation would begin...
Whoa! The large windows of the spacecraft lit up – and we looked out onto the red dirt of Mars! Instructions appeared on a virtual heads up display, and we shouted commands to one another, adjusting the controls as instructed. The shuttlecraft lifted, and we flew through canyons reminiscent of the US southwest and eventually landed at the Martian base. It was an incredibly realistic experience.
Is VR the killer app for 5G? I don't know, but the applications go beyond entertainment. Instructional videos can be more effective in a VR environment. Then there is the whole issue of augmented reality.
I pondered this while I strolled the exhibition floor, sadly, sans headset...
More 5G applications
IMS hosted a "Micro-Apps" theatre, which was located in the centre of the exhibition, and consisted of short, rapid-fire presentations. I sat in a couple that were 5G focused. Mike Millhaem, also of Keysight, gave a presentation titled "Now You See Me, Now You Don't!" discussing mmWave signal propagation and implications for networks and tests. Mike showed measured results of 5G channel sounding, and accentuated some of the challenges mmWaves present.
* The above shows a Keysight 5G channel sounding system being rolled in front of a brick wall and window to record the channel characteristics. To the right shows the results versus time. Even small changes in surface cause dramatic effects. This shows the difficulty in predicting mmWave behaviour. Images courtesy of Keysight Technologies.*
I caught up with Mike later in the conference to get his take on 5G applications. He also pointed to the high density stadium case, but added another one, "we don't know." Not being flippant, Mike's point was that it is difficult to predict what new applications can surface when there is a dramatic change in bandwidths. Would we have predicted that 3G/4G would have enabled new taxi companies like Uber or Lyft? I wandered to the National Instruments booth and caught up with David Hall and Charles Schroeder for the same discussion. David's answer to my question dealt with overall capacity. Sure, I could download at 70Mb/s now into my iPhone 5S, but we couldn't all do that at the same time. Here it comes down to pure spectral capacity, not peak rates. In some ways, this is the generalized case of the stadium application- is there enough total bandwidth available to feed a city? Fair enough.
Screenshot of 4G LTE speed test performed on the author's iPhone 5S on the conference floor. But what it everyone wanted this top speed?
My final opportunity to pose the question came during a panel discussion moderated by Eric Higham of Strategy Analytics, with senior panel members from Analog Devices, Anokiwave, Keysight, NI, and Rohde & Schwarz. The panel was titled, "The 5G and IoT Conundrum", where the panelists described the opportunities and challenges represented by 5G.
I played devil's advocate one last time- asking what would drive these 5G investments versus continued advancements of 4G technology.
All participants were adamant that 5G service providers were racing for deployment. Again, the stadium use case and virtual reality were brought up. Also mentioned was the total capacity issue – I might have gotten 70Gb/s download rate, but it would have been slower if everyone in the conference had tried. The moderator described the competitive dynamics that are leading to the race – no service provider can risk not investing in some trials just to be left behind when the bandwidth intensive applications hit. As I described, only somewhat in jest, they all feel they need to belly up to the saloon's bar, and rest their pistol on the counter.
So, what does this all mean? Here's my final takeaways...
Wrapping it up
I think there is enough of a compelling use case in total capacity needed, and ultra-dense applications such as the stadium example, that 5G mmWave will be deployed commercially. But I'm still hedging my bets about the timing of 5G ubiquity. 4G, when aided by massive MIMO (Multiple In, Multiple Out), can achieve new heights of spectral density by using spatial multiplexing. Also, 4G wavelengths below 6GHz behave well compared to their unruly 5G counterparts. But there's one application I will be watching that no one mentioned- replacing the "last 200m" of fibre. Consider this. China has been witnessing a huge migration from the rural countryside to the big cities, a trend predicted to continue until 2035. There are dozens of mega-cities of about 10 million people each, where residents live in dense high-rise apartments. How does one bring internet to these residents without massive levels of new construction? One method would be to bring fibre to the intersection, and then use mmWave technology to connect to each household. The technical issues are massively reduced since this is a fixed communication application, not mobile. Dynamic searching and beamforming isn't needed- the provider already knows the location of their clients. The beams can be fixed. Indeed, the first 5G trial from Verizon is a fixed point application. This video from Starry, an internet start-up, shows how the company is using mmWave communication to compete against the ISP incumbents. If you squint your eyes, between fixed point internet service, ultra-dense locations, and spectral scarcity, you can make the case for 5G. It's not reality yet, but, it is the virtual reality. In part two of this series I will look at the current state of the test industry in supporting 5G. In the meantime, readers, headset on!