A Samsung executive has claimed the role of virtualised RAN (vRAN) technology in wireless networks will continue to grow even when telcos move onto 6G and beyond.
The claim comes days after Verizon and Samsung announced the completion of a fully virtualised 5G data session over C-band spectrum, from 4GHz to 8GHz. Samsung is currently the network equipment provider for the US telcos’ 5G network deployment, having signed a $6.6 billion supply contract last year.
“Samsung has been partners with Verizon for a long time,” Samsung Networks vice president and head of advanced system design lab Jeongho Park told .
“Some years ago, when vRAN was just beginning to gain traction, Verizon expressed their desire to shift to vRAN. When we received this request, we were already preparing to launch vRAN. We began making preparations to launch a virtualised distributed unit around 2018, having already developed a virtualised central unit prior to that for a full vRAN solution.”
Samsung launched its first “fully” virtualised 5G RAN last year, which includes a virtualised distributed unit for its virtualised central unit that went into service a year prior in 2019.
“vRAN requires technology, optimisation, and commercialisation experience. We believe Samsung is the only vendor that meets these three requirements, putting us ahead against other companies by 1.5 years in this space,” the Samsung lab head claimed.
Samsung has been one of the more aggressive companies in rolling out vRAN solutions among the major 5G network equipment vendors. Networking powerhouse Ericsson, which calls its vRAN solution Cloud RAN, currently has plans to roll out the first stages of the service in the fourth quarter this year. Nokia, meanwhile, made its full vRAN solution commercially available last year and is trialling the technology with US telco AT&T. The company is also planning to roll out the solution with C-band support later this year.
Huawei, which is still the world’s largest 5G network equipment vendor despite US sanctions, has been the least active in promoting the technology. This is likely due to most carriers outside of China linking it with Open RAN, a term used loosely by different organisations and a concept that the Chinese giant has not been willing to embrace so far.
‘Flexibility is a big advantage of vRAN’
The way vRAN operates is that it delivers network functions as software instead of hardware. Instead of using proprietary hardware-based baseband units and central units, which is what traditional RAN does, vRAN has virtualised distributed units and virtualised central units — software running on commercial off-the-shelf (COTS) hardware. The core network also runs on software.
Through this software, vRAN offers the same network functions as its fully-hardware counterparts in traditional RAN, but virtually. Samsung has touted that this “full” vRAN solution, which means it offers the distributed unit, the central unit, and the core in software allows it to provide flexibility.
“Before vRAN, mobile network operators received dedicated hardware from a specific vendor. In effect, operators would feel, in a way, dependent on the specific vendor. But in vRAN, operators can now use hardware from company A and software from company B as they wish like Legos. This flexibility for operators is the big advantage of vRAN,” the Samsung lab head said.
According to Park, this gives operators two benefits: They can pool their network resources efficiently, which in turn, leads to lower costs.
“In traditional RAN, a dedicated processor is used in the baseband unit to process the signals. But this processing power isn’t needed 24 hours a day. Some coverage areas may not need the full processing power. Processors installed on population-dense areas like Gangnam are usually made to handle the expected data traffic. But sometimes, one population-dense area may have intense data traffic while another similar area has almost no traffic. So this shows that you don’t need maximum specs for every processor.
“In vRAN, software replaces the function of the processors. This software can determine when and where it wants to use its resources. We also offer an orchestrator solution that manages all the virtualised distributed units, which was previously done individually. This leads to an overall decrease in cost of ownership and operation cost,” he said.
It is also important to note that the global 5G network market is just beginning. In the US, though mmWave spectrum support has been deployed, telcos like Verizon are just now grouping it with C-band spectrum to expand coverage. In South Korea, telcos are yet to even deploy mmWave services. The increase in base stations and coverage will require more computing power from networks going forward.
Samsung’s vRAN solutions currently use x86 CPU on the COTS servers, sometimes with accelerators, Park said, while noting that the processing needs would increase for high spectrum from the wider bandwidth going forward.
“Bandwidth becomes wider and more complex as we move from C-band to mmWave. Networks will indeed need more computing power. But at Samsung, we believe CPU products launching up to next year can handle more workload of what they handle right now,” the Samsung lab head said.
“Many companies, including Samsung, are also considering applying GPU. But for now, our commercialization roadmap is drawn based on CPU. We are still looking at the pros and cons of GPU.”
The evolution of cloud will also inevitably be tied with vRAN, Park added.
“Cloud groups processes into one place and is flexible. Its purpose corresponds with that of vRAN. We think in the future, there will be companies that offer vRAN in cloud form,” he added.