One of the major engineering challenges of 5G is making Massive MIMO antenna arrays affordable and compact, and so avoid risks of large spikes in TCO, while also optimizing performance as more operators deploy networks in high frequency bands from 3.5 GHz to millimeter wave. One of the significant proof points for the fledgling open RAN movement will be to attract the best Massive MIMO designs to their platforms, and ensure that support for open interfaces, between the radio/antennas and the virtualized baseband units, does not come with performance trade-offs.
In this area, Japan’s NEC is making big strides. Having developed advanced Massive MIMO products for NTT Docomo and Rakuten in its home country, it aims to break into the international radio market for the first time, riding on its early support for O-RAN interfaces. It is as productizing and commercializing its Japanese designs for the wider market, and playing a key role in the development of the Rakuten Communications Platform (RCP), which is designed to generate revenues from a packaged-up, pre-integrated version of the technologies Rakuten has deployed in its cloud-based 4G/5G network.
NEC’s latest moves are to work with Analog Devices (ADI) on a new family of O-RAN Massive MIMO products, and to get closer to the UK government (see separate item) in the hope of gaining business with operators there, as they are mandated to move away from using any Huawei 5G equipment in the core or RAN.
But the O-RAN/Massive MIMO combination is also a hot target for start-ups, such as the newly emerged Xcom, and the more established Blue Danube, and we can expect to see many innovators looking to take advantage of a more open ecosystem to move into the formerly closed cellular market.
But Massive MIMO is the biggest challenge for open RAN, says Yago Tenorio, Vodafone’s head of group network architecture and a leading light in the O-RAN Alliance and Telecom Infra Project (TIP). It is the main reason why it will take open RAN platforms some time to deliver the same performance, for high end systems, as a traditional integrated solution.
Talking in a keynote to TelecomTV’s recent virtual Open RAN Summit, he said: “5G Massive MIMO in particular is a complex problem to crack and will require not just a lot of software development but also hardware, in particular semiconductors… we will need new solutions to be developed in terms of chipsets that are powering those efficient, super-complex radios. It is difficult to say how long that is going to take. Maybe as little as 12-18 months to a full solution. But it can be later if open RAN adoption is low in the industry.”
This is why Tenorio wants as many operators as possible to cooperate within groups such as TIP, in order to accelerate development and provide scale and confidence for vendors. But it also requires large and start-ups suppliers to prioritize open interfaces when they are allocating budget and effort to Massive MIMO, which in turn means they must believe in the potential for strong market demand and adoption in the reasonably near future.
Tenorio talked about Vodafone’s recent open RAN radio hardware RFI, which included Massive MIMO, and said several companies had shone with technologies that complied well with the operator’s criteria. These suppliers included Airspan, GigaTerra, NEC, NTS and Xilinx.
These pioneers will now be joined by Xcom, a San Diego-based start-up set up in 2018 by a group of former senior Qualcomm executives, including former CEO Paul Jacobs and former president Derek Aberle. It aims to test advanced MIMO technologies supporting open RAN interfaces with small cells from Baicell, in an indoor arena in Sacramento, California.
Xcom’s CTO Matt Grob (formerly CTO of Qualcomm) wrote in a filing with the FCC: “We’ve developed a technology to significantly improve the spectral efficiency of systems that use sub-6 GHz spectrum, and this particular filing relates to our being able to expand our trial system.”
He told LightReading that the firm has developed a MIMO antenna design that should dramatically improve overall capacity and performance of a wireless network, and will be particularly targeted at environments with large numbers of users, such as stadiums. The first tests will target the 3.5 GHz CBRS band.
Last year, Xcom acquired M87, a Seattle-based networking start-up led by former T-Mobile executive Cole Brodman. M87 was working on a technology to allow smartphones to transmit communications from nearby users who were outside the range of a signal, perhaps because they were deep indoors. Their signal could bounce off other nearby devices via a mesh network until it reached the Internet.
Jacobs has described his new firm’s mission as giving “everyone’s phones the ability to route traffic like a cell tower”. In an interview with the San Diego Union Tribune when the company was set up, the three founding executives said they would target 5G and IoT applications, and they particularly highlighted solutions which enable large amounts of compute processing power to be supported at device level, to reduce the volumes of data which need to be sent to the cloud (or even the edge cloud) to be processed.
Meanwhile, back with NEC, the collaboration with ADI will initially focus on a new development for Rakuten, but will then seek a wider customer base. The companies are working on a 5G O-RAN Massive MIMO radio using ADI’s fourth generation software-defined radio (SDR) to simplify design and lower power consumption, which means the radio could be used in small cells. The system performs digital pre-distortion (DPD) in addition to digital beamforming to boost performance, said ADI.
The chip company’s VP of wireless, Joe Barry, told EETimes: “Roll-outs are progressing from coverage to capacity deployments with 5G Standalone functionality, and we’re seeing the potential of virtualization of the networks”
He added: “In a traditional RAN, the digital processing is done onsite. With virtualization, much of the network function can be done in the cloud. Most early installs of 5G were non-standalone (NSA) and not virtualized; in the next wave, standalone (SA) installs will be the norm, and will evolve to virtualization (be it vRAN or O-RAN flavor). We are only at the cusp of what is possible with a virtualized 5G network, but early indications are showing real promise.”
The key advantage of open RAN is the broader ecosystem and innovation base it enables, he argues. He said: “The new collaborations in 5G are leading to a vibrant marketplace. In the radio units, we’re seeing innovation in 5G radios that solve the size, weight, power and cost challenges by combining ADI’s technology solutions with silicon and IP partners.”
But he acknowledged that Massive MIMO is challenging, saying: “The first hurdles are size, weight, power, and ultimately cost, which are significant to solve. Then layer on widening bandwidths, which push performance requirements, and frequencies above 3 GHz, which increase the colocation interference risks. It is a multi-dependent variable challenge with many hidden problems. It is innovation at the circuit and architecture level that is successfully enabling M-MIMO for 5G.”
ADI has been seeking to address these issues by increasing channel count density integration, increasing supported radio bandwidths more than tenfold, and reducing the power per channel by a similar amount, and the firm says the results of this work, over five generations of technology, are integrated into its 5G M-MIMO SDR platform.