Lowband congestion would normally force Bell Canada to find new spectrum or build more sites, but its work with Cohere Technologies may hold the answer.
A mobile signal in Canada needs to be a tireless long-distance runner. With a landmass of about 3.5 million square miles, it is the second-largest country in the world, behind Russia. For such a vast territory, it is also relatively uninhabited, with a smaller population than Spain's. Bell Canada and Telus, two of Canada's big three telcos, share a network of only about 11,000 mobile sites across this huge expanse, managing roughly 5,500 each. "The densification of the cell sites is a challenge," said Brad Stimpson, Bell Canada's director of wireless technology. "And so the grid was lowband-oriented, and therefore we had a lot of lowband congestion in the network."
Bell's 850MHz spectrum is ideal for extending a signal's range and helping it to overcome barriers. It is not as good as higher frequency bands, where more spectrum is available, at providing the burst of speed required for data-hungry applications like video. To relieve congestion, an operator would typically bring more spectrum into play or add sites – the densification to which Stimpson refers. Neither is inexpensive, and Bell's recent strategy has been to reduce its capital expenditure. Overall spending fell from $5.1 billion in 2022 to $3.7 billion last year.
Such considerations have piqued the company's interest in Cohere Technologies and its universal spectrum multiplier (USM) software. Already tested by European operators such as Deutsche Telekom and Vodafone, it is essentially intended to provide a more muscular set of legs for the wireless signal, boosting the capacity of the deployed network. "We prided ourselves on having the fastest, best network in Canada, and so it was important for us to find innovations that allowed us to beat this low-band congestion," said Stimpson on stage at the inaugural Mobile Network Innovation Summit (MNIS) in London last week. "That's really where Cohere came in."
Bell seemed to like what it originally saw, having made an investment in Cohere in January 2023 through its venture-capital arm. The startup has also secured funding from Australia's Telstra, besides various investors that do not operate networks, and been praised by Vodafone for the improvements it delivered during field trials in Spain. Cohere owns patents for orthogonal time frequency space (OTFS), a potential 6G technology that has caught the attention of the US Department of Defense, among others, and has an executive lineup with a strong wireless pedigree. Ray Dolan, Cohere's current CEO, previously ran Flarion, a pioneer of the orthogonal frequency division multiplexing (OFDM) technology used in today's 4G and 5G networks. In 2006, he sold that business to Qualcomm for a total consideration of $805 million.
But it would be wrong to assume Cohere has been widely deployed in Canada, and it remains on the periphery of production networks in other parts of the world, too. The challenge for Bell or any operator is introducing Cohere into network products that come from other vendors in a market dominated by a few big players. After removing equipment provided by China's Huawei, to ensure compliance with government rules, Bell divides its current 5G network between Ericsson and Nokia. It is these companies – not the operators – that act as the gatekeepers of the radio access network (RAN). So far, neither has fully opened the doors for Cohere.
Their wariness is not surprising. In a Catch-22 situation, Cohere is unproven in a commercial network deployed at scale. If its technology had an adverse impact on the network, the reputations of Ericsson and Nokia could suffer. More intriguingly, Cohere's objectives are not necessarily aligned with those of the big kit vendors. Huawei, Ericsson and Nokia all depend on selling new radios and other hardware products to operators. A technology that helps telcos to avoid densifying the network or investing in newer equipment might not be in their best interests.
But the mainstream vendors are unlikely to win any popularity contests by opposing the wishes of their customers. And some growing telco pressure on vendors to accommodate Cohere may at last be having an effect. Just over a year ago, Cohere flaunted a common, off-the-shelf server that incorporated its USM technology and could be positioned next to a basestation at a mast site or stored in a data center.
During trials in Mississauga, Ontario, Bell deployed that server at an 850MHz site with a relatively basic radio featuring four transmitters and four receivers. At the time, the spectral efficiency improvement was measured at between 30% and 50%, Ronny Haraldsvik, Cohere's senior vice president of business development, told Light Reading last year. Since then, trial work has progressed, said Stimpson.
Importantly, the trial has involved one of Bell's two existing RAN vendors. Which one has never been disclosed by executives at Bell or Cohere, but Ericsson last year confirmed to Light Reading that it was not working with Cohere on the Bell project. By a very simple process of elimination, that would make Nokia the Cohere partner in Canada.
Ideally, however, Cohere would prefer direct integration with a vendor's technology over the use of an additional server. The spectral efficiency and related benefits are likely to be even greater when USM is built into the basestation as well as receiving all the data it possibly can from the vendor's scheduler, a RAN feature responsible for traffic management and service quality, according to Haraldsvik. Avoiding additional hardware would also probably lower costs for Bell.
In a further possible complication, though, Bell and other telcos are keen to prevent Cohere from becoming an exclusive feature of a single vendor's technology. "What was important for us was a solution that was vendor-agnostic, and so we wanted to make sure whatever we worked with could eventually apply to all the suppliers, all the vendors, all the partners in our network," said Stimpson, echoing the comments of Vodafone's Paco Pignatelli last year. A takeover of Cohere by one of the big kit vendors – surely possible if its technology is so good – would presumably shatter those hopes. Operators might also worry about missing out on the full USM benefits if Cohere were acquired in a defensive move.
In the meantime, Cohere has continued to market new features and polish its offer to entice the interest of other service providers. Branded ECHO (short for enhanced channel insight with holographic observability), its latest feature, unveiled during MNIS, exposes millions of new data points that could assist an operator with the future design and rearchitecting of the network, said Haraldsvik. "We literally generated in 5MHz per sector over 400 million lines of telemetry, which is fantastic," he told the audience at MNIS.
A lot hangs on what Bell can achieve, and how feasible a nationwide deployment of Cohere in Canada proves to be. "We've made a lot of progress," said Stimpson. "There are some steps we have to go through to scale this up to the size of the Bell network. It's more about the industrialization of the software." With other telcos undoubtedly keeping watch, major success for Cohere at Bell could be the international trigger it needs.
Want more Light Reading stories in your Google search results?
