SAN JOSE, Calif. — 2017 will be a big year for 5G cellular. Engineers will start silicon development in earnest and pick architectures to support massive MIMO antenna technologies, according to Mike Murphy, chief technologist for Nokia in North America, in an interview with EE Times.
Carriers will continue to explore virtualized radio access networks (RANs) and expand work in hot frequency bands such as 3.5, 28, 37, and 39 GHz. Separately, operators will turn on LTE-based Category M and Narowband IoT services, said Murphy, who meets a few times each year with executives from his carrier customers.
Among the hot topics in those meetings, “5G is certainly at the top; virtualization of the RAN has become a hot topic recently; and Wi-Fi/cellular integration with Licensed Assisted Access will be important in 2017,” he said.
The 3GPP group is expected to finish the physical and MAC layer standards for 5G in December. So silicon designs will begin in earnest “a small number of months before that” when engineers decide the specs are “solid enough to go ahead,” he said.
Nokia is among the top three companies making network infrastructure gear such as base stations. All use a combination of off-the-shelf chips and their own ASICs, a trend thatMurphy expects will continue.
Carriers are migrating more and more jobs to applications run in virtual environments on standard x86 servers. But “a 5G small-cell base station on a pole can’t be done with off-the-shelf processors to get the right cost and power target,” he said, suggesting that such chips may need 14- or even 10-nm processes.
After several years, sales of small cells for LTE are now slowly and steadily picking up, he said. They mainly are used in dense urban environments such as stadiums.
Antenna technologies for 5G “are significantly different” than in 4G. Massive MIMO and beamforming techniques will be used to direct signals, especially in millimeter-wave bands. Systems with 8 x 8 antenna arrays will ship this year with 64 x 64 versions not far behind, expanding as millimeter-wave solutions emerge, he said.
“There’s a lot of devil-in-the-details on this topic: options of implementing what is in the antenna, in the baseband, and RFIC chips solutions — it’s a complex topic with a lot of debate on the right approaches,” he said. “We are at the point of picking the fundamental architecture and making decisions on low-level alternatives.”
The move to millimeter-wave bands from 28 GHz and up is bringing in new partners, “some well known and some smaller … and some from the defense industry,” Murphy said, declining to name names.
Next page: Millimeter waves meet cloud-based networks