One of the key selling points of 5G wireless technology to the world's telecommunications providers has been the opportunity to enter new markets and create new sources of revenue from wireless services. The opportunity to build an Amazon, or something like AWS, leveraging much of the assets that telcos already own, has been most compelling of all.
But in a twist that 5G's engineers perhaps should have seen coming, enterprises and industry groups have begun exploring this issue in reverse: If telcos can build the next cloud using x86 servers and micro data centers at the edge, then why couldn't they? All they would need is access to wireless spectrum. And in many countries, including the US, there may be just enough of that unlicensed spectrum hanging around -- including leftover frequencies from the VHF/UHF TV transition -- to pull off an upset.
And just by asking the question, the possibility arises that the very companies these enterprises would seek to compete against in the cloud (even if only in limited capacities) could emerge as communications service providers (even if only in limited capacities), challenging telcos for the rights to provide 5G communications to their production facilities and headquarters buildings.
Cutting the carrier
"If 5G is as important as we claim it to be -- society, infrastructure, etc.," declared Dr. Hans Schotten during a panel session at the recent Brooklyn 5G Summit, "then of course society has to care about making this available to everybody. So we have an issue of participation. If we build a system in a way that network operators completely control when, where, which version of 5G is deployed, then they are the party that decides who will benefit and who will be able to participate in these modern digital societies." Dr. Schotten directs the Intelligent Networks Research Group at the German Research Center for Artificial Intelligence at the University of Kaiserslautern.
Schotten approached the topic from a philosophical standpoint. If all the marketing slides are correct -- the ones that state 5G is, or will be, the foundation not only of global commerce but of human society -- then he argued, the role telecom providers play in that society becomes elevated. They become ministries of connectivity and the facilitators of the economy. This should make the various governments that bestow such power quite wary of the awesome responsibilities they've been given, or that they've claimed for themselves.
Unless, of course, the basic assumption is wrong: that no one voted to give these agencies such awesome power.
"This is a level of responsibility that, from my point of view, was completely underestimated," he told his audience. "There was a big hype saying that 5G is great and will really change the world, and it can provide very significant, added values to industries. But it brings the operators [into] an extremely central, important role for the societies, because they decide which of the different companies in these vertical industries -- and when, and where -- get these benefits."
It isn't the 5G technology itself that produces this outcome, but rather the economic and geopolitical state of the various countries seeking to adopt it. What 5G was actually designed to enable is a concept called non-public networks (NPN). It's the basic concept of 5G New Radio (5G NR) applied to wireless spectrum that isn't claimed by, or leased by, telcos. It's public spectrum, which is why calling it "non-public" may seem counter-intuitive. But the idea is this: If a 5G NR system can use Wi-Fi-like tactics to dynamically claim bands of specified, reserved frequencies while no one is broadcasting on them -- frequencies in the public space -- then private industries can make use of them to provide 5G services in limited areas.
"4G was largely, for all practical purposes, associated with carrier-operated networks," explained Dr. Henning Schulzrinne, speaking with ZDNet, "because of licensed spectrum, equipment costs, historical reasons, and the assumptions made about systems. If you said '4G,' basically, the next question should have been, 'Who's the carrier?'"
Dr. Schulzrinne is professor of computer science at Columbia University, former chief technology officer for the FCC during the Obama Administration, and the co-creator of much of the Internet's multimedia conferencing protocols, including SIP.
"In 5G, that might not be quite as true," he continued. "The carriers would love it to be true -- clearly they see industrial and more enterprise-type applications as a prime way to justify their investment in 5G technology. But people talk about the equipment for 5G, in certain circumstances, actually owned and operated as a separate network, very similar to Wi-Fi, by large industrial manufacturers."
One of the world's centers of research into this application of 5G, Schulzrinne tells us, is the European Union. There, private enterprises are establishing their own NPNs for the first time, detached from carrier networks and providing 4G LTE service -- and soon 5G -- for their own campuses.
Bosch GmbH, the Germany-based machine tools colossus, is one of those enterprises. ZDNet readers may recall last year, when Bosch entered the news for its having advocated a 5G technique called network slicing. This would enable operators to carve out private virtual networks for their industrial clients, using compute and storage capacity from their own cloud servers. VMware is one of the companies seeking to productize this concept, working to create what would be a vSphere-like platform for telcos to manage their digital assets.
For network slicing to work, telcos would have to decide which servers and data centers should be made available for carving up. AT&T is on record with its opinion that this concept is not only unfeasible but, in the end, illegal. Telcos have obligations in their respective countries, its engineers have said, to ensure certain quality-of-service (QoS) levels, and slicing-and-dicing physical networks into virtual ones would violate their contracts.
At any rate, the debate over network slicing hasn't progressed very far in the past year, with the final resolution still up in the air -- as was evident at last April's Brooklyn 5G Summit. Not one to wait for the telecommunications industry to make up its collective mind, Bosch has opted to go a different route, seeking a license from German regulators to operate its own 5G networks, with its own equipment, on designated frequencies within the country's boundaries. German automakers Volkswagen and Daimler (parent of Mercedes-Benz), and machine tools maker Siemens, have followed suit. As a Siemens spokesperson told Reuters, "We can't wait for the network operators to be ready."
These German companies bid for, and won, wireless service licenses. With a forthcoming platform planned by 3GPP as part of the evolving 5G portfolio, even that bidding would be unnecessary. Suddenly the prospects for wholly owning and operating one's own connectivity becomes foreseeable, for industries that would never have considered it before -- for instance, petroleum.
"There are some customer needs where network slicing will help," admitted Geetha Ram, head of HPE's Edgeline business development, during a 5G Summit panel. "But there are other, larger customers: We've jointly worked with a whole lot of oil and gas customers, and they would like their own networks. And it's not about technology for them; it's about the control, and they would like their own private network."
"We should not get scared of the fact that we are talking about a private network," remarked Dr. Simone Redana, who heads the Network and Architecture Research Group at Nokia Bell Labs. "I think we should look at it more from, what we want to achieve with that. We want to achieve a solution that meets the requirements of the verticals."
As industries build out their IT assets and distribute their data center capacity more and more, Dr. Redana explained, their QoS requirements become more specific and, to a growing extent, less negotiable. Data must be stored not only where it's accessible but where it's legal, especially in the EU. His firm is interested in designing and deploying a commercial solution that would provide an alternative to network slicing, whose launch appears to be stuck in a holding pattern.
"There's no particularly good reason why, say, a large manufacturer such as Caterpillar or Boeing would want Verizon or AT&T in their manufacturing plants, messing around with their radio equipment," remarked Dr. Schulzrinne. "They would want their own technicians to run the cables, install it, operate it, and integrate it into their network."
The unlicensed bonanza
The basis of 5G's NPN technology, and the root from which it would sprout forth, is a present-day platform entitled LTE Unlicensed (LTE-U). China Mobile and Huawei joined Qualcomm, Ericsson, and Verizon in formally proposing LTE-U to the 3GPP consortium in 2013. One of the group's originally stated intentions for the platform was to enable carriers to leverage extra capacity in the 5GHz band (initially in the 5725 - 5850 MHz sub-band) for initial bandwidth for facilitating mobile data traffic. Since that time, carriers including Verizon and AT&T have leveraged an offshoot of that platform, called LTE Licensed Assisted Access (LTE-LAA), to boost 4G data speeds. LTE-LAA is a key part of an AT&T transitional platform it calls "5G Evolution," even though 3GPP classifies it as part of 4G.
"Now with Release 16 of 3GPP, there's actually the possibility to take 5G and use unlicensed spectrum," explained Michael Thelander, president and founder of Signals Research Group. "So the idea of an enterprise, another organization, going out and having to get licensed spectrum goes away."
The problem with LTE-LAA, said Thelander, was that it required a carrier to serve as its anchor. With what engineers are now calling 5G NR U (Qualcomm doesn't seem to like hyphens much), the service can be offered as a stand-alone product. Nobody has to license channels or spectrum bands from anyone, since the NR U platform would borrow Wi-Fi's uncanny ability to cobble together long stretches of bandwidth from unused chunks, at just the right time.
Thelander conceded that unlicensed spectrum in the 5GHz territory would probably be subject to interference from Wi-Fi smartphones. (This would include 802.11ac Wi-Fi devices that transmit and receive in 5GHz -- devices which Wi-Fi engineers call "5G," although that's a different "5G" altogether.) But in North America, there's work under way to define discrete sub-bands in the range between 6 and 7.125GHz. Forthcoming Wi-Fi generations may make use of some of this space, but not all of it -- and it's that leftover "quasi-greenfield" spectrum, as Thelander calls it, that may inevitably make NR U feasible for 5G.
It's here where, ironically, the ability to push incumbent carriers out of this space may create an opening for a different category of service providers, which may include Amazon and Google, to make its way in. Both LTE-U and its heir apparent, NR U, are frequently discussed in the context of direct competition against Wi-Fi. Of course, neither company is offering any specific, official comment on this issue, although sometimes from otherwise forthcoming companies, silence can speak volumes.
But we can read quite a bit from these companies' actions. Last April, AWS joined the Kinetic Edge Alliance's effort to build cloud connectivity in so-called edge data centers -- micro facilities located adjacently to wireless towers. And Google continues to make inroads with its Google Fi service, which for years has enabled phones to dynamically select service from T-Mobile, Sprint, US Cellular, and Wi-Fi hotspots. Last February, Google inked an extended deal with Sprint to enable Google Fi service on 5G smartphones, using Sprint's Massive MIMO radios -- a genuine evolution of 3GPP's 4G, on a transitional path to 5G.
This puts both these service providers at the doorstep of being able to offer resold NR U service from Ericsson or Nokia. Naturally, the phones themselves will need to be capable of adapting to NR U's multitude of frequencies, though Qualcomm would certainly see to that. Ericsson engineers have already publicly voiced their willingness to provide such a service, though it has not named any prospective customers.
The components are being assembled for an Internet service provider with some muscle (and there are very few of those) to launch not only a competitive 5G unlicensed spectrum access service for specific enterprises, but also conceivably an NR U service for certain Internet-of-Things devices on the consumer level. Such a service would not only edge out Wi-Fi, but compete against cable/MVNO services such as Comcast's Xfinity and Charter's Spectrum. And if it goes further than a dip in the shallow end, that service may eventually call into question the telcos' rights to be the exclusive providers of all things bright and glorious on 5G.