The shift to 5G Standalone—and enterprise service delivery—requires cross-domain network resource management and an OSS/BSS re-alignment
The latest generation of cellular that’s available out in the world today largely follows a 5G Non-Standalone architecture meaning new radios and spectrum were added to existing core infrastructure. What you may have heard called “real 5G” refers to a 5G Standalone mode of operation wherein a shiny, new cloud-native core facilitates the delivery of the advanced enterprise-facing services that, based on industry consensus, will spur long-term operator monetization efforts.
But along with that move to 5G Standalone, operators need to think strategically about cross-domain network management and orchestration that allows the investment in a service-based architecture to be translated into enterprise-facing SLAs that can be dynamically configured and delivered, according to Ravi Sinha, vice president of technology and solution development for Reliance Jio.
Sinha, who also sits in chairmanship positions within the O-RAN Alliance and Small Cells Forum, discussed how this evolution in network architecture and management, as well as how O-RAN principles and AI fit into the mix, during the recent 5G Monetization Forum. In an interview with Hardin & Associates President Lauriston Hardin, Sinha talked through challenges operators will face in the near-term as they transition to 5G Standalone, but also longer-term as 5G Advanced, coming with 3GPP Release 18, and eventually 6G begin to influence how cellular networks are architected and operated.
With 5G Standalone, Sinha said a service-based architecture lets an operator use and re-use the same infrastructure “based on the different kinds of use cases” which can range from automated guided vehicles and robotic control to fleet management and, at a higher-level, more real-time data analysis. “Those things need a very separate approach where the different slices of the network…[are] well-aligned with the service provider maintaining the SLA in terms of jitter, in terms of delay, in terms of throughput, in terms of quality of experience, everything end to end. And that’s also on the fly. Those things will not be supported by Non-Standalone…One day, the major chunk of business will be coming from the enterprise.”
The OSS/BSS of it all
With this transition to 5G Standalone and bespoke SLAs for enterprises, operators also face a transition in how networks are operated and monitored, and how billing systems track and charge consumption. Sinha described this emerging paradigm shift: “There are humongous areas of changes coming because of a new generation of services coming around.” The interactions between RAN, transport, core, and service management and orchestration functions “are very much focused toward AI/ML-driven data collection, data analysis and data model training, and how those things can be pushed across domains…Ultimately it’s all going toward your OSS/BSS.”
The shift to 5G Standalone is predicated on the shift to a cloud-native core, but there are other factors that feed up to the need for OSS/BSS evolution—network disaggregation, currently focused on the RAN and part of the larger story around Open RAN principles and open interface specifications defined by the O-RAN Alliance, adds complexity by adding vendors. Just as the RAN is historically a proprietary monolith provided by a single vendor, the same is often true of OSS/BSS systems.
Speaking last year at the Telco Cloud Forum, MATRIXX CTO Mark Price tied together the future of slicing networks to support niche services and the outlook for operations and billing platforms. “I don’t think we would be reinvent our networks or the business systems around them if it wasn’t for the fact that the business models themselves are changing.” Networks are typically optimized for broad application market segments like voice calling or video streaming, not the specific needs of one manufacturer using one type of industrial robot. “Networks are optimized for throughput,” Price said, adding that “the business systems are optimized to suit that.”
Back to how all this relates to monetization, Sinha said, “It is important for scale and price efficiency that the same network…can be leveraged in many ways…[As] an operator, I can have my connectivity slice going to a bank, many going to the public domain, but the same slice can be re-aligned for a specific need.”
Open RAN, the edge and getting to real time
Open RAN, while still maturing, is prompting operators to distribute radio infrastructure and capabilities in a new way wherein centralized, distributed and radio units can be located in ways that create cost and operational efficiencies while opening up the opportunity for new types of service delivery and network optimization.
This comes as there’s also ramping investment in moving compute functionality out of centralized data centers and to more lightweight, distributed computing nodes; this facilitates latency-sensitive services where data needs to be ingested, analyzed and turned into action more quickly than backhauling it to a remote cloud would allow.
These two trends create an opportunity to think holistically about new network architectures and layer in decision-making compute power for the RAN (and for services) at the network edge. Sinha sees the edge as a platform for network functions and for exposing those functions, and others, to application providers. “Anything in terms of connectivity is going up to the telco layer as well as the telco edge layer and the device layer,” he said.
But Open RAN is not without its challenges (and critics), and has been deployed at scale primarily in greenfield network builds by 1&1, Dish and Rakuten Mobile. That said, major multinational operators have committed to significant future investment in Open RAN systems. In terms of headwinds, Sinha called out management of multi-vendor systems in the RAN, and aligning those multi-vendor systems with other multi-vendor systems at the edge, and in the transport and core domains. “We have not reached to a level where we can have plug and play and pick and choose,” he said. “Most of those challenges will be solved…then we can achieve very much closer to the traditional kind of solutions in telcos.”
Big picture, he said Open RAN will continue to mature and will be aided by the proliferation of Open Test and Integration Center labs and the O-RAN Alliance’s certification of compliance program. “I’m hopeful that 2023 and ‘24 will be the time when you see hundreds of thousands of not only the macro side of it, but I’m hopeful that the pico[cell]s and enterprise solutions can also become compliant to Open RAN.”