Newly funded Cohere wins huge bonuses in beamforming battle | light reading

2021-12-15 00:31:50 By : Mr. Bob Tsai

Since the beginning of the 18th century, visitors to St. Paul’s Cathedral have been able to whisper on the gallery wall inside its huge dome, and can clearly hear it on the opposite side-as if the speaker’s lips were against the listener’s ears Same.

Not the original design intent of Christopher Wren, this whisper gallery phenomenon is still a science tutorial for school children and adults.

"On the acoustic level, this is actually beamforming," said Ray Dolan, CEO of Cohere Technologies.

Since the famous London landmark was first opened for three hundred years, beamforming has become a hot topic in the latest mobile standard 5G.

The idea is to use the wireless signals flowing around any cell site to direct them to the intended recipient, like a whisper through a dome.

Experts say that if done effectively, beamforming can greatly increase coverage and capacity, reduce the demand for mobile network equipment, and increase the value of spectrum. This is what makes Cohere such an exciting adventure.

The Silicon Valley-based company is now 10 years old and claims to have created a version of beamforming that is far superior to any other product on the market.

It is also taken seriously. The test led Vodafone to estimate that Cohere's technology can double the capacity of a 700MHz network compared to using traditional technology. For 3.5GHz-based networks, which are usually associated with 5G, it is estimated that four to five times improvement is possible.

Acoustic beamforming can be observed in one of London’s most famous buildings (Source: Eluveitie)

This is an encouragement for Dolan, who became chairman and chief executive officer in October 2018.

"This is a very cool aspect of our story," he told Light Reading.

"We can make the spectrum more valuable by doubling or quadrupling it."

With the help of Cohere's technology, the suggestion that billions of dollars in frequency investments may be worth multiples seems to attract investors.

"We just received funding, and now the company is well capitalized," Dolan said.

Although Cohere does not intend to share details, Dolan believes that his company has sufficient funds to support growth next year and 2023, by which time the company is expected to become profitable.

The latest round came after Cohere raised US$35 million in 2015 from companies including Telstra Ventures, the Australian telecommunications investment arm. Other service providers are also interested in its technology. Cohere said that Germany's Deutsche Telekom participated in the trial last year, and there is no shortage of interest in the North American market.

Go out to defeat the big gun

What makes Cohere better than anyone else? As Dolan said, the company's researchers basically borrowed technologies used in the radar industry and applied these technologies to wireless communications.

A system called Delay Doppler evaluates time and frequency measurements to create a channel space map between the mobile tower and its users. The beam can then be adjusted for better performance.

"In that world, beamforming is like a lighting director, he not only has a better pencil beam, but also a script so he can count on the actors where they should be," Dolan said.

"When people say you are just another beamforming company, I would say no. We have actually tamed beamforming."

There must be a certain degree of doubt as to whether such a small company can surpass companies like Ericsson and Nokia, which spent a total of nearly US$9 billion on research and development last year.

As the core capability of wireless communication, beamforming is still a priority for Nordic suppliers. Being surpassed by Cohere is like a chess master being defeated by a child prodigy-embarrassing and equally changing the rules of the game.

However, Cohere does have blood. It was founded by Ronny Hadani and Shlomo Rakib, two Israeli scientists who support a modulation scheme called Orthogonal Time Frequency and Space (OTFS), which is considered a potential alternative to the OFDM system currently used in 5G.

Cohere still sees it as a "next-generation waveform." Gabriel Brown, principal analyst at Heavy Reading (Sister Light Reading), believes Cohere may try to position OTFS as a 6G candidate.

Dolan's resume is equally impressive. As a wireless industry veteran, he is the co-founder and CEO of Flarion Technologies, which was sold to Qualcomm in 2005 for $805 million.

"This is basically 90% of the intellectual property in OFDM," he said. "This is 99% of what Qualcomm did. They are worth about $200 billion. I think they got a good deal."

It is not easy to copy OTFS or delayed Doppler things-partly, Dolan said, because Cohere has "patented" its innovations. He believes that at the same time, its competitors have been walking on different beamforming paths.

Nokia’s recent white paper on Massive MIMO (an advanced antenna system that supports beamforming) describes methods including grid of beams (GoB), codebook-based beamforming, and eigenbeamforming.

Want to learn more about 5G? Check out our dedicated 5G content channel on Light Reading.

However, in the less advanced MIMO setup, Cohere has won praise where others have been criticized. Carlos Ubeda, a technical expert at Vodafone, is targeting existing suppliers of 4T4R (four-transmit-four-receive) units.

He stated in a recent Light Reading webinar that these companies cannot handle multi-user MIMO and cannot allow many devices to communicate through multiple antennas. Cohere's technology made this possible and led to the performance improvements touted by Vodafone.

"The capacity of 4T4R today is usually 50% lower than the upper limit. This is because traditional vendors do not support multi-user MIMO, so the peak throughput is achieved by allocating two layers for the same user-single-user MIMO," he said.

"Thanks to Cohere and its delay Doppler-based channel estimation program, we can effectively deal with the challenges of multi-user MIMO and multiplexing two users at the same time. This means doubling the capacity."

But if Cohere is so advanced and difficult to imitate (in law), why hasn't it been acquired by Ericsson, Nokia, or other companies in the wireless business?

James Crawshaw, chief analyst at Omdia (another Light Reading sister company), speculates that Cohere will demand higher prices than the Nordic suppliers are willing to pay.

They may have speculated that they can develop this technology internally at a lower cost and then defeat Cohere in court. In addition, the acquisition of Cohere may herald the embarrassment of the R&D departments of both parties.

Dolan admitted that even cooperation would be embarrassing for them.

"This can be integrated into traditional RAN, but it requires existing people to say please do this for me-please be my brain and decide what I want to eat today," he said.

"This is their nature, so there will be some commercial resistance, but technically speaking, it's very simple."

However, this commercial resistance can be a serious obstacle. Cohere is very similar to other RAN experts, but hopes that opening up RAN will open the door.

If it improves the interoperability between suppliers-as its proponents say-then Cohere can effectively integrate into a multi-vendor network built by many experts. Operators including Vodafone and Deutsche Telekom are very keen.

Therefore, Cohere's argument is that its technology can be run entirely as a software application on the RAN Intelligent Controller (RIC), which is a new platform conceived by the open RAN community. Dolan insists that, separate from RAN equipment and software, it can exist in the cloud and still perform its magic.

This is mainly because the delay Doppler system has been carefully designed to provide 100 milliseconds for the setting of the real-time beam, ensuring that the delay is not a limitation.

"It's a myth that you can't reorganize into the cloud, because the channel changes so fast, you need to put all your smarts as close to the phone as possible," Dolan said.

"This is in the interest of people who want to keep the system shut down."

If he is right, Cohere will be able to partner with RIC platform developers and integrate with various networks. It has established a partnership with VMware and is working with another company that has not been named.

However, not all RICs are equal. Near real-time RIC is the successor of distributed SON (self-organizing network) technology in some traditional base stations, and its running time period is less than one second.

Due to the longer time period, non-real-time RIC is more similar to centralized SON technology. As the name suggests, it is located on a centrally located server. Near real-time RIC is provided with xApps, and the equivalent of non-real-time RIC is called rApps. Cohere doesn't seem to be.

"We are not so focused on non-real time," Dolan said. "The value of our actions is not that great."

There are several disadvantages. The first is that Ericsson launched non-real-time RIC last month, which seems to be resistant to the xApps concept.

"They believe that value and intelligence should be placed on the node," Crawshaw said in a previous interview with Light Reading.

"They don't want the value to be sucked away and stay in the near real-time RIC that VMware may sell."

The second problem is that the interface specification for near real-time RIC has not yet been prepared.

"Non-real-time RIC is far ahead and has a roadmap for commercial deployment," said Brown of Heavy Reading. "For near real-time RIC, it is still too early."

Even if Cohere's performance as an xApp is not worse than when it is integrated with nodes, all this means that it may need to establish partnerships with distributed unit (DU) suppliers, which are part of the RAN that handles baseband processing.

"The way to commercialize it will involve reaching an agreement with the DU software supplier," Brown said.

Living up to expectations, Cohere may disrupt the entire industry. It is difficult for anyone to ignore the large-scale spectrum efficiency improvements achieved entirely through software—whether it is an operator who spends billions of dollars on spectrum or a kit supplier who is worried about commoditization.

The biggest question is whether its effect in this area is as good as during the Vodafone trial, how easy it is to deploy, and whether it is really difficult for competitors to follow suit. By this time next year, the answer should be clearer.

Related articles: Cohere CEO: Open RAN is the ideal choice for enterprise applications Podcast: O-RAN and Cohere Technologies CEO discussion Cohere CEO: This year launched new products and customer names Open RAN People’s Front and Open RAN’s People’s Front Cohere collapsed Open RAN party

— Iain Morris, International Editor of Light Reading

© 2021 Light Reading, part of Informa Tech, a division of Informa PLC.