5G: A Revolution in Networking
A couple of weeks ago I went to the ATIS 5G Symposium in Chicago held in conjunction with the Light Reading Big Telecom Event. ATIS is the chief organization of North American telecom network operators and equipment producers, including all the big names and most of the small ones as well: from AT&T, Verizon, Cisco, and Qualcomm on down to the Texas 9-1-1 Alliance and the Moapa Valley Telephone Company.
Everybody in the telecom business wants to know what 5G will bring, and until very recently the answer was a resounding “it depends” because 5G meant different things to different people. But in the last six months to a year we’ve begun to hear a more consistent story.
Unlike 3G, 5G will not simply be a new radio technology. 4G paved the way for a more comprehensive approach to radio infrastructure by incorporating Internet Protocol into the design of radio networks, and 5G will take that a step farther by incorporating Software Defined Networks (SDN) and Network Function Virtualization (NFV). 4G also includes the Internet’s Integrated Services framework into the radio infrastructure inside something called IP Multimedia Subsystem (IMS) that supports phone calls and media transactions, and 5G will most likely include LTE-U for use on unlicensed spectrum. (We’ve written quite a bit about LTE-U here on High Tech Forum.)
The most forward-looking vision for 5G – and one that not everyone was willing to endorse – came from Tom Anderson of Cisco, who see 5G as a chance to make some overdue and well-needed changes in the nature of the Internet. The Cisco vision (see the presentation here) would make the Internet a more name- and content-oriented network than the server-centric, numbered, firewalled system it always has been.
This is not to say that 5G won’t include a new radio access network; it will, and the new network will be the so-called “millimeter wave” spectrum between 24 GHz and roughly 60 GHz. This is 4 to 10 times higher frequency than today’s Wi-Fi 802.11ac band, which means signals will be more directional but also more prone to interference from tree leaves. The directionality is good for performance because it means less sharing and more exclusive connections between the phone and the tower.
The interference is bad, of course, so the engineering challenge is to work around it. If we can, then smartphones will have gigabit connections just as Google intended. Technologies like beam forming and Multi-user MIMO (a multiple antenna system) point to a solution, but we’re not there yet. There are also power issues with extremely high radio frequencies. But millimeter wave will join the stable of frequencies already in use rather than replacing them, so even if it’s only usable in the urban jungle, that’s a great help.
One speaker who works with the military, Brett T. Walkenhorst, touted cognitive radio with machine learning for making decisions about channel selection, more or less, but didn’t elicit many cheers. Wi-Fi fans have extensively hyped cognitive radio, but it’s better suited to military systems than to commercial ones. When you’re invading another country, your radios need to throw a few elbows to nudge incumbent users off the air, but in peacetime we have better mechanisms such as auctions, licenses, and roaming agreements to make such decisions. It’s rather odd that cognitive radio buffs are also universally upset about commercial interests using “their unlicensed Wi-Fi spectrum”; this is what the Wi-Fi vs. LTE-U controversy is about.
The most definitive voice on 5G is Dino Flore, the chairman of the 5G Radio Access Network work group. He charted the course from where we are to full 5G; this is going to take five to ten years. I won’t steal Flore’s thunder because I plan to do a podcast with him shortly.
4G LTE was a monumental step forward for 3GPP, the standards body that makes cellular standards because it incorporated IP in a system that can scale up to hundreds of megabits per second eventually. 5G is going to be an even more revolutionary step because it will transform the Internet and also give us radio networks that can scale up to 10 gigabits per second.