Artemis, developer of the pCell concept, has relaunched its product as it comes to market with a fully fledged commercial offering for LTE and 5G private network and neutral host deployments. CEO Steve Perlman said the company was looking for external investment as it targets growth following the acquisition of its first customers, and launches a new product to market.
The relaunch comes eight years after pCell first splashed down in the mobile radio market. Back in the mid-2010s, a company came along that said it was going to revolutionise wireless networking. It bombastically proposed a new way of delivering wireless signal to devices that it said would have the effect of directing a mobile cell’s full capacity to every device within the cell, with complete uniformity of performance.
The company was Artemis and the product was the pCell. The technology was tighly wrapped up, but simply put (as simply as we can), the pCell was said to work by flooding an area with signal from multiple pWave antenna units, deliberately creating interference. By analysing what happens to signals as they travel through this interference to a connected device, intensive computational processing on a [rack server] then creates a waveform just for each individual connected device. It’s not massive MIMO, whereby lots of antennas create beams to connected devices, and it’s not advanced interference cancellation.
At first the technology was met with some ridicule, not helped by there not being much publicly accessible material detail on how the technology worked. CEO Steve Perlman was heckled at Columbia and Stanford. As for industry, “operators were saying it was too good to be true,” in Perlman’s words.
After a paper was published in 2015, telcos and OEMs did investigate the technology, and in some cases that interest went far enough to move to trials. Such as with Nokia, which put its name to some co-development activity. But according to Perlman that fell away as the sponsoring Nokia CTO Hossein Moiin left the company as the Al-Lu merger integration was accelerated.
And then… nothing. Or not much, in any case – there were some intermittent releases that never seemed to lead to concrete progress. Perlman told TMN that Artemis came up against a sort of institutional inertia. Despite telcos showing interest, they were also faced with the imminent rollout of 5G. They weren’t about to throw over what they knew about cellular networks in favour of an untried technology from a small company sitting in a tech incubator.
Indeed, one Tier 1 operator spent nearly two weeks in the Artemis lab verifying the technology, but accidentally revealed to Artemis’ team that although they verified the technology worked, it would rather wait for one of the major OEMs to copy or deliver the tech, than engage further with Artemis.
pCell rises again, goes private
But then, come forward to the end of the decade and into the early 2020s, we see the emergence of CBRS and the growth of interest in private networks give Artemis a second wind. It made its pivot. The company would move away from trying to sell at a macro level to mobile network operators, and instead exploit the growing trend to explore private network models.
“We envisioned ourselves as a vendor for mobile operators, but we realised we were just the R&D group for these guys,” Perlman said. “Very clearly they saw us as an R&D company that they just wanted to take information from. We realised there was no way in there, and then with the disappointement of the Nokia relationship we couldn’t get in that way.”
But with the rise of interest in private networks, and high profile offerings from the likes of AWS and Azure, along with a number of programmes to open up shared, unlicensed spectrum, Artemis spied an opening.
“When we saw the US was going to come out with the CBRS band, and we saw lots of presentations and the involvement of the hyperscalers in private networoks, we made a big pivot. And for the last five years we have been entirely focussed on crafting what it will take to deliver multi gigabit services in sub 6GHz spectrum, in large scale networks that are going to compete directly with mobile operators and with Wi-Fi.”
So now Artemis is coming back to market with the pCell Multi-Gigabit LTE/5G vRAN. It said that a network powered by pCell can deliver over 1 Gbps in 20 MHz of spectrum, and over 7.5 Gigabits/second in 150 MHz.
It has partnered with CBRS SAS provider Federated Wireless on go-to-market, and, crucially, it has a live, 28-antenna deployment providing service in the 20,000 seat SAP Center, a large sporting and events arena in San Jose.
At the SAP Center, users with devices that support eSIM and CBRS can scan a code or use an app to provision what is effectively a SAP Centre SIM. That means their device can then communicate with the pCell system, which effectively registers them on a separate network, controlled by an integrated, cut-down EPC.
Perlman said that this element of the solution is attracting interest from other venue and event owners. Rather like a WiFi login, it could allow for a sponsor to be the “provider” of connectivity on site, or it could be delivered by the venue owner as a premium visitor experience. The radio server combined with MEC could deliver live streams of multiple stages to attendees and crew at musical festivals, he said, or live feeds to attendees at a race circuit.
Proving the doubters wrong
But at the core of the pCell has been doubts about how its technology can be so revolutionary and deliver on its claims. Perlman said that there are now over 100 academic papers about the technology, which has also been known as Distributed Input, Distributed Ouput (DIDO) as it has developed. The company claims it can deliver 1 Gbps total data rate in 20 MHz, and 7.5 Gbps in 150 MHz.
“With 150MHz of CBRS spectrum, at the SAP Centre, there’s no question it works. We are delivering 7.5 Gbps with complete uniformity across the arena,” Perlman said. And the company argues the system, which requires additional servers to support more radios, but covers the arena with 28 pCell points, is much less expensive than a rival DAS.
It also claims that pCell has 20x the capacity of Wi-Fi in the same spectrum. In a test it ran of 400 phones using Wi-Fi, 23% of the phones could not connect to the Internet at all. Its antennas can also be installed using power over Ethernet, with Ethernet cabling providing the fronthaul connection between the software radio server and the antennas.
Althought the current focus is on the private network space in the growing area of shared and unlicensed spectrum, Perlman says that the radios can work with any spectrum, supporting bands from 600 MHz to 6 GHz, so can in theory support licensed bands. It can also act as neutral host for operators.
“We have good relationships with operators and now we have a system working and deployed they want to partner with us. We are like an ice breaker and at the moment it is CBRS and neutral host that is at the prow for us.”
Perlman also said that the company is opening up for investment. “We have been self-funded but are now stepping up, with our first customers we are now seeking growth investment.”
Explaining how pCell works is not easy. But going by the company’s materials, you can identify the basics, even if a lot still remains in “clever maths” territory. There is a video explainer here, and papers available here.
An earlier paper from the company put the process in the following terms:
“When a device connects to the SDR server, it instantiates an eNB in SDR just for that user.
“Every eNodeB is implemented in the pCell SDR as a software instantiation called a “virtual radio instance” (VRI). The baseband waveforms of all of the VRIs feed into the pCell Processing, which combines the waveforms and produces the complex waveforms that are sent over fronthaul to the pCell antennas. The pCell antennas concurrently transmit the waveforms, and the waveforms propagate through the environment and constructively interfere with each other at the exact location of each equipment device (UE). The combination of these waveforms at the location of each UE results in the synthesis of the baseband waveform that had been output by the VRI associated with that user device. All the users receive their respective waveforms within their own pCells concurrently and in the same spectrum.
“How is [pCell] able to create waveforms that, when they sum together at each user results a clean waveform with the data for that user?
“The complete answer to this question is very long, involving immensely complex mathematics, very carefully designed software and hardware, and new data communications and modulation techniques. The following is a highly simplified explanation: All of the intelligence of the system is in the server, which then communicates to all of the users at once through all of the APs at once.
“Communication begins with the APs exchanging brief test signals with the user devices. By analysing what happened to these test signals as they propagate through the wireless links, the [system] determines precisely what will happen when it transmits data signals from the APs to users, and how the simultaneously transmitted signals will sum together when received by each user device. Then, [pCell] uses this analysis to create precise waveforms for all of the APs that, when transmitted at once will sum together at each user device to create a clean, independent waveform carrying the data requested by that user. So, if there are 10 APs and 10 users all within range of each other, then 10 radio signals will sum together at each antenna of each user’s device to produce an independent waveform for each device.
“In wireless parlance, widely distributing both AP and user antennas achieves exceptionally high “diversity”. Diversity makes each antenna “statistically independent” (its signal paths are different than other antenna signal paths), which is how the system distinguishes each APs signal from those of the many APs that reach a given user. This allows the system to figure out precisely what waveforms it needs to generate so that all the waveforms sum together into a clean waveform for each user. Each of these clean waveforms is an independent channel.”