In addition to helping operators cut costs, Ericsson’s AI-enabled approach to 5G transport automation sets the stage for new service monetization
As the industry continues to advance 5G networks, the stage is set for the scaled transition to cloud-native standalone mode of operation that will open new automated capabilities designed to help communications service providers (CSPs) overcome fundamental business challenges—reduce ongoing operating expenses and open up new lines of service revenues. And as CSPs pursue this opportunity for efficiency and growth, it’s imperative that the 5G transport network receives the same level of attention as the radio and core network domains.
Modern mobile networks have reached a degree of complexity that makes a unified approach to network status and performance data observability very difficult. Consumer and enterprise demand for mobile services is growing at an astounding pace, and not just for 5G; legacy generations of cellular have an important role to play in certain geographies and bring tangible benefits to specific types of use cases like low-power IoT and rural connectivity. At the same time, hardware/software disaggregation is bringing to the fore the complexities of integrating and managing multi-vendor network systems. And standardization work for 6G has already started, meaning investments today need to be made as durable as possible as requirements continue to evolve.
And a lack of network observability brings with it very real business and operational challenges, including difficulties in root cause identification, an increasing volume of site visits, configuration errors, skyrocketing energy consumption, and ensuring that optimizations focused on one network element do not cascade into problems that degrade end-to-end system performance. All of these challenges waste time and cost money.
“Network observability is something many operators are struggling with today,” Jari Augustin, Head of Product Line for Transport Automation at Ericsson, explains to RCR Wireless News. “And it is one of the things that has been preventing them from further automation.” Ericsson’s goal, he adds, is to help solve this in part with the recently launched Ericsson Transport Automation Controller.
He distilled the value proposition to three key pillars, all of which enable more automation: minimized complexity; increased operational and energy efficiency; and optimizing transport network performance. The problem though, Augustin points out, is that “today, most of what happens in the network is invisible to the service providers. And being able to see how your network performs, understand what happens and where, and know why certain issues occur become key in solving those challenges.”
Heightened visibility into the network allows service providers to focus on fixing issues quickly and accurately rather than go through the cumbersome process of iterative troubleshooting. “This will ultimately lead to better performance, lower downtime, lower cost,” Augustin says. How? In the case of the transport network, by using AI-enabled solutions to collect, contextualize, analyze and act on the data at hand — effectively “seeing the invisible”— in a predictive manner that delivers wide-ranging efficiencies without compromising performance.
In terms of how transport automation sets CSPs up to deliver (and monetize) new service revenues, the high-level value proposition is reduced time-to-market for network slicing, a logical partition of the network that meets the service level requirements of a particular user or user group while optimizing the operator’s use of network and spectral resources. Discussing this opportunity, Augustin rightly linked the evolution of transport to the evolution of the radio access network (RAN).
In the case of RAN, operators are increasingly investing in disaggregated, distributed radio networks that split functionality between centralized, distributed and radio units. The goal is to use standardized interfaces, particularly for fronthaul, to enable architectural flexibility and support multi-vendor deployments. As an aside, Ericsson is at the vanguard of this movement as evidenced by its $14 billion, five-year win with AT&T.
Discussing the connection (both literal and figurative) between the transport network and RAN, Augustin notes that cloud-based RAN is reaching an “inflection point, especially the fronthaul. Packet fronthaul cannot be an afterthought. It will inherently be a part of the Cloud RAN solution,” he says. “Maximizing investments in both, essentially reaching a situation where the sum is greater than the parts, will require automation to orchestrate data flows and leverage programmability.”
Augustin and his colleagues at Ericsson describe both transport automation specifically and network automation in general as a journey they are taking together with their customers. As progress is made toward the goal of an intelligent, automated network, the importance of network observability cannot be overstated. If an operator doesn’t know where they are currently at, they will not be able to identify their needs and develop a strategy that addresses those needs. Having a high level of observability that is linked up to a larger automation framework creates a strong base to build upon.
