Mavenir Launched Open RAN Massive MIMO, When Can It Be More Competitive?

Network Equipment Provider (NEP) Mavenir announced a wide portfolio of Open RAN-compliant radio units in February 2022, with more details revealed during its analyst event that took place June 13 to 14, 2022. The OpenBeam portfolio follows the O-RAN 7.2 interface and supports both licensed and unlicensed spectrum ranges, consisting of macro, micro, Massive Multiple Input, Multiple Output (mMIMO), and Millimeter Wave (mmWave) radios for Open RAN deployments in 2022 and beyond. The radio solutions also address the needs of network operators to be cost efficient with low wind load, lower energy consumption, and built with intelligence and automation. The OpenBeam radios are available to the Open RAN ecosystem (e.g., vendors, operators, and system integrators), focusing on the U.S. and European markets first.

Traditional Radio Access Network (RAN) vendors have already launched 64T64R mMIMO that weighs approximately 20 Kilograms (kg) or even lower (e.g., Ericsson launched a 20-kg unit and Huawei launched one that weighs 19 kg). To be proficient in mMIMO, both Huawei and Ericsson have used Application-Specific Integrated Circuits (ASICs) within the radio designs. ASICs provide high capacity and reduce energy consumption, but have fixed functions and need large volumes of deployments. These customized chipsets are the key advantage for traditional RAN vendors and their high volume of business justifies the investment. In contrast to ASICs, a Field Programmable Gate Array (FPGA) is an alternative type of semiconductor, its flexibility enabling adjusting it after design to meet certain needs, which makes it ideal for Open RAN. However, in most cases, a FPGA is the enemy of energy efficiency and unit cost, which is not ideal for mMIMO. Traditional RAN vendors (e.g., Huawei and ZTE) also invest in Gallium Nitride (GaN)-based Power Amplifiers (PAs) to boost energy efficiency in their mMIMO units. Moreover, these two vendors invest in new materials and new designs to explore their potential to reduce power consumption. For example, Huawei’s MetaAAU integrates software and hardware, and enables a power saving of 30%. ZTE’s V-tooth heat dissipation can improve heat dissipation efficiency by 20% when integrated with new phase change inhibition material. Besides hardware technology, Artificial Intelligence (AI)-based software is often developed to help reduce network energy consumption further. For example, Nokia’s AI solution, AVA, can cut energy by 30% and reduce energy costs by up to 20%. Ericsson’s mMIMO deep sleep feature enables a saving of 70%, and annual site energy saving of 1.2 Megawatt Hours (MWh).

The OpenBeam mMIMO launched by Mavenir is less than 25 kg and leverages Xilinx’s Radio Frequency System-on-Chip Digital Front End (RFSoC DFE) technology to minimize energy consumption. Xilinx, an inventor of the FPGA and Systems-on-Chip (SoCs), has made FPGA more energy efficient by hardening features such as RFSoC DFE; meanwhile, it has invested in other technologies, such as a processor branded AI Engines to maintain the programmability of FPGA. According to Xilinx, AI Engines consume only half of the energy consumption, but can deliver up to 8X the computing power compared to traditional programmable components. Like the traditional vendors, Mavenir introduced an energy-saving software to switch mMIMO into sleep mode when traffic demands are low and claims that this energy-saving feature can yield energy savings of 80+%. However, it is not clear how much the 80% is in terms of Kilowatt Hours (kWh) or cost saving.

Mavenir’s launch of OpenBeam mMIMO is an anticipated step in the broader company strategy, and expands Mavenir’s product portfolio to fill an important gap in the market for sub-6 GHz Open RAN mMIMO. However, there are significant challenges to be addressed in both product design and performance, something that needs to be addressed before Mavenir’s products become strong contenders for both greenfield and brownfield operator networks. The first generation of OpenBeam mMIMO is sub-25 kg, which is a good achievement, but still insufficient to succeed in mMIMO. Companies like Mavenir must produce specifications that are on par with current Tier One traditional vendor products, most of which have reached the third iteration after they were launched, whereas Mavenir has just launched its first version. In terms of size and weight, this is unlikely to happen in the next 1 to 2 years. Regarding the energy saving feature, Mavenir claims energy savings of over 80%, while traditional vendors claim 30% to 70% savings. These percentage figures are subjective and abstract, as neither report what vendor(s) they compare against. Moreover, 5G mMIMO for Open RAN requires not just significant software development, but also hardware, in particular semiconductor expertise for optimizing performance, cost, and energy consumption. The industry needs to invest in semiconductor innovation and develop new solutions in terms of chipsets to power those super-complex and efficient radios, a process that can take months or years to achieve a fully mature solution. Also, more efficient cooling technology should be applied; when a mMIMO product does not require supplement fans, its form factor can be further reduced. Mavenir must accelerate Research and Development (R&D) to match its equipment specifications to Tier One NEPs in the next 2 to 3 years, which may require major effort and investment, because all Tier One NEPs have invested billions of dollars in optimizing their equipment.

According to Mavenir, there have been 100+ Open RAN deployments and trials; this is expected to increase as Open RAN gains more momentum. Many European operators are supportive of Open RAN; for example, Telefónica said 50% of its RAN Capital Expenditure (CAPEX) will be Open RAN by 2025, while Deutsche Telekom, Orange, and Vodafone all signed up for a list of technical requirements for Open RAN. Network operators are committed to contributing to the O-RAN Alliance and bringing Open RAN up to match with integrated systems in performance and in areas like mMIMO. There is no doubt that Open RAN mMIMO will progress over time, but traditional RAN vendors are also investing heavily to maintain their competitiveness. Open RAN mMIMO may gain some market share, but it is uncertain if it can beat traditional RAN mMIMO.