Engineers at Nokia Bell Labs are ready to talk about some pioneering work they’ve been doing with AT&T on distributed massive MIMO (DmMIMO), which they say significantly increases uplink capacity and speeds in 5G networks – “without requiring an overly complex solution.”
It sounds at least somewhat complex to the average layperson. However, Peter Vetter, president of Bell Labs core research, said it’s really the next logical step after Massive MIMO, which is relatively well established in wireless networking circles by now.
To understand DmMIMO, it’s worth taking a step or two back. MIMO refers to Multiple Input Multiple Output, and it’s been around for decades. Gerard Foschini, who turns 82 today, was the primary contributor to the early MIMO technique when it was first developed in the 1990s at Bell Labs, contributing to the principles that are used today in 5G, according to Vetter. So, in a sense, they’re building on that DNA with today’s announcement.
Massive MIMO also evolved at Bell Labs and refers to a lot of antenna elements – but contrary to the name, they’re not super big in size. In fact, Massive MIMO inventor Tom Marzetta told Fierce a few years ago that he wasn’t really keen on that term but ended up using it because that’s what so many other researchers were using at the time. It’s a semi-long story, and more on that can be found here and here.
RELATED: Massive MIMO gains ground in 5G, but still not reaching its full potential: inventor
Another long story short, Massive MIMO is a friend to TDD networks, which is where Sprint led the way with mid-band spectrum, and that 2.5 GHz spectrum is now in the hands of T-Mobile. Fast forward and other carriers – such as AT&T and Verizon – now hold licenses for other mid-band spectrum, notably the C-band 3.7 GHz and the more newly licensed 3.45 GHz.
The C-band is where AT&T is most likely to implement Massive MIMO and by extension, DmMIMO.
Distributed MIMO for uplink
With DmMIMO, they’re basically using signals in a way that under other conditions might be interpreted as noise.
Typically, antennas in a crowded array will pick up noise from another signal and interpret it as interference. Dealing with that reduces the efficiency of the network, so to get around it, a distributed system will use antenna arrays to coordinate with one another and act like a joint receiver to eliminate cross talk. All of this results in an increase in the uplink capacity, which is becoming increasingly important as users upload content to social media and conduct video conferencing. The process is discussed in this Nokia video.
Gordon Mansfield, who oversees both fixed broadband and wireless access technologies as part of his role as VP/Mobile & Access Architecture at AT&T, said one of the things that has become pretty clear about wireless networks is “they typically trail what we see happening in the wired networks by a few years.”
In wired broadband networks today, “uplink has absolutely become a lot more utilized,” Mansfield said.
Looking forward from a mobility perspective, “we do see the need from a wireless perspective to improve and enhance the uplink. We don’t have a challenge today,” but if you study what’s happening on the wired broadband side, “we do see the uplink becoming more strained in the future,” and that’s where it becomes interesting to work with Nokia on the distributed MIMO concept.
The upshot is their DmMIMO simulations have demonstrated increases in 5G uplink capacity between 60% and 90% compared to similarly configured systems with a single antenna panel, according to Nokia. While other techniques come at the expense of downlink capacity, DmMIMO seemingly would produce sizable increases in uplink capacity without sacrificing performance.
In a press release, AT&T Network Services CTO Andre Fuetsch noted how AT&T and Nokia have a long history of working together on new network technologies and he welcomes this latest collaboration. Newer applications in the 5G-Advanced timeframe, such as Extended Reality (XR), are expected to be more demanding on the uplink compared with regular broadband traffic.
Of note: In addition to uplink improvements, the move to DmMIMO has the potential to alter the configuration of networks by facilitating deployments with smaller antenna arrays per site, according to Nokia. That would lower the weight, resulting in smaller and lighter components and reduced power consumption – things network engineers are always striving to achieve.
Alternatively, DmMIMO also could be used to boost downlink capacity, so operators would have more flexibility in how they deploy the technology.
It's all fine and great to have an idea, but applying it is another matter. “We can proudly say that our prototype works,” Vetter said. It’s working in the lab in Murray Hill, N.J., and they shipped it to the AT&T lab in Austin, Texas, where the companies confirmed it’s working as well.
The next step is evaluating it in real-world conditions, and Mansfield suggested they have the perfect setting to do that at AT&T’s Austin campus, where they can conduct experiments outside the lab and do so in a more realistic environment where mobility is part of the picture. There, AT&T can give Nokia Bell Labs some more real-life feedback before moving to actual commercialization.