Get email delivery of the Cadence blog featured here
Yesterday was my post What Is 5G? which is the first half of my introductory look at what 5G really is, based on Ian Dennison's DAC presentation 5G Intelligent System Design. Today we pick up where that post left off. As a reminder, a 5G network looks like this, with the sub6GHz (and legacy 4G) on the left and the mmWave on the right.
One of the reasons that 5G is so visible and suddenly a part of international diplomacy is that it is one of the highest-profile technologies we've ever seen, and one of the few to operate on an international scale.
When 2G (the first digital cellphones) was introduced, it was somewhat political, but that was addressed in two ways. Many countries with strong 1G mobile companies like the US and Japan (and China, coming from behind) decided they were big enough to use their own incompatible technologies that only their own national champions would bother with. Much of the world, starting with Scandinavia, then broadening to Europe, and then many other countries too, went with a single standard, GSM, which was TDMA (time-division multiple access, where each call was given a time slot in a given frequency band for its exclusive use). One standard in the US was CDMA (code-division multiple access).
It turned out that the way of the future was W-CDMA (wideband CDMA) rather than extending TDMA, which made relatively inefficient use of its frequency band (for one thing, it still allocated time slots for the link in both directions even though normally only one person on the phone talks at once). 4G and the more advanced versions known quas LTE (long-term evolution) and LTE-A (for advanced) gradually migrated to a more uniform architecture where everything is just data and phone calls don't need to be handled with their own bands, they are just digitized data, too. Also, the differences in standards between different regions diminished so that the same smartphone could work anywhere in the world. The frequency bands available varied from place to place based on historical allocations that couldn't easily be undone, but that is fairly easy to handle.
5G carries this uniformity even further. Everyone knows it is a single worldwide standard, and the whole world is up for grabs. Governments and cities around the world all see 5G as necessary infrastructure for future digital economies (not to mention things like autonomous vehicles). Despite a lot of noise being generated in the US, there is no real US company to champion.
The handset market is expected to be a lot more fragmented, as it is today, enabled by application processors and modems available in the merchant market.
I said above that network providers expect buildings to be provisioned with a lot of femtocells. But when was the last time you even used the data on your phone when you were at home, the office, or at a mall? You almost certainly used Wi-Fi. Well, in addition to 5G cellular there is a new Wi-Fi standard. They are calling it Wi-FI 6 and that corresponds to what has also been called 802.11ax. At the same time, 802.11ac was rebranded Wi-Fi 5, and 802.11n was rebranded Wi-Fi 4. Wi-Fi 6 can deliver 10 Gbps, uses unlicensed spectrum, and is free. All the things that you are used to with Wi-Fi, only more bandwidth.
I don't think it is possible to get a definitive answer, but estimates are that 70% of downloads are over Wi-Fi not cellular. Of course, some are still on wired connections. My TV box connects to my router with a patch cable, for example.
It seems that there is likely to be a conflict inside buildings. The networks want you to use their network, so that they make money from you. You, on the other hand, would probably prefer to use free Wi-Fi when it's available, which it certainly will be at your home and your place of work at least. I can't imagine putting a femtocell in my living room and then having Verizon charge me for using it. Conference centers might be more of a issue since they don't provide free Wi-Fi in general. In fact, they are about the only place in the world, besides a few backward airports, where there is no truly free Wi-Fi. Even when it appears to be free to you as a conference-goer, usually it is being "sponsored" (that is, paid for) by one of the big companies attending. In that environment, I don't see much advantage one way or the other between femtocells and Wi-Fi 6 routers.
Another potential conflict is over wired access to buildings. I said in yesterday's post that 5G could be used for fixed wireless, providing last mile internet service (well, last 300m, mmWave doesn't go as far as a mile) without needing to dig up the ground. But even when the ground has already been dug up, there is potential for getting bypassed. Existing TV cables risk becoming stranded assets with limited value, just like many copper phone lines have done, bypassed by mobile. I haven't had a landline phone for about a decade, and I can't imagine my children ever will feel the need to get one, despite the line already coming into their living room. It will be the end of an era when new housing developments no longer install landline phones during construction.
Talking of TV cables, there is likely to be something reminiscent of the fights when cable TV was first installed about what rights existed to use existing poles that were there for electricity or landline phone. No municipality wanted another set of poles installed. Something similar is likely to happen with 5G installations on existing poles and lamposts. One big issue will be whether every network operator needs its own radiohead on each post or whether those will be shared. Whatever the outcome, it is likely to reduce the speed of deployment of mmWave.
I expect there to be a big marketing conflict about 5G. Indeed, it already started, when AT&T upgraded their operating system at CES so that it displayed "5G" instead of "LTE-A" as I reported in my post in January CES: 5G, All Hat and No Cattle.
I think that there will be a lot of 5G hype at the 2020 Olympics in Japan next year. I'm sure that one or two venues will be equipped with 5G but whether anybody has a 5G handset to make use of it remains to be seen. My advice to anyone would be to wait until 5G starts to be widely deployed and second-generation handsets are cheaper.
But I think that the big gap between hype and reality will be that the benefits of mmWave will be billed as 5G, without any mention of the fact that mmWave doesn't go through walls and won't have been widely deployed initially, so mostly that is irrelevant. That's not the performance you will get. There will be a big gap between the headline number promised in the marketing and the actual performance you'll get on your phone.
I saw a similar credibility gap when GSM was introduced. Instead of being billed as making better use of spectrum and so, eventually, cheaper, it was billed as having "digital quality" just like the switch from LPs to CDs. When people found that the quality was nothing close to CD quality, and would often drop out without warning, especially in the early days, they were very dissatisfied.
Cadence is producing its own 5G reference design, currently under development. Some of the board and package level approach was presented at CDNLive EMEA in May. I expect to write about this reference design too sometime later this year. Some aspects of the design are already scheduled at various academic and industry conferences. Keep an eye out for more details.
See the Cadence 5G page. And, I'm sure, many future blog posts here this year and in the years to come.
Sign up for Sunday Brunch, the weekly Breakfast Bytes email.