How to scale and accelerate Open RAN?

According to Radisys, Open RAN scales comes from simplification, reducing variability and focusing on product integration

Early commercial deployments of Open RAN are beginning to pop up around the world, with Rakuten in Japan and Dish in the U.S. perhaps being the most notable. Despite the progress made, Open RAN is still in its infancy. According to Ganesh Shenbagaraman, head of integrated products and ecosystems and Radisys, there are three dimensions of success when looking to accelerate and scale Open RAN: simplification, variability reduction and a product integration-focused approach.

“Open RAN transformation has just begun for both telcos and enterprise networks and this is being enabled by multiple factors [such as] the disaggregation that is happening in the network architecture, the hardware-software decoupling and cloudification, along with all of the programmability and automation,” Shenbagaraman told 2022 Open RAN Global Forum attendees, adding that telcos around the globe are beginning to dip their toes into larger-scale deployments. “This going to be the turning point where Open RAN becomes a reality,” he said.

First, Shenbagaraman laid out the challenges hindering the large-scale adoption of Open RAN. While multi-vendor integration is seen as the “number one challenge,” particularly when it comes to scale, there are additional obstacles impacting adoption that have to do with achieving end-to-end performance, maintaining interoperability as standards evolve, managing configuration and deployment variability and finally, ensuring that the desired goals are met, which in most cases, involve the creation of new paths to monetization.

Simplified approach

Because there are numerous network variables and considerations — network size and region, coverage goals, spectrum availability, the maturity and availability of the device ecosystem, etc. — Shenbagaraman argued that simplification is key here.

The simplified approach involves a network evolution that is “well-defined” and allows for “adding more spectrum capabilities, band capacities and more vendor diversity.” Flexibility in the network architecture is also an important part of this simplified approach because it allows the network to become more disaggregated, enabling the functionalities of the DUs, CUs, user plane, control plane and so on to be seamlessly moved throughout the network to account for the deployment realties of different regions and cell site areas.

“The network should also allow for the evolution of the network to be completely programmable through the RAN intelligent controller [RIC], which is an essential element of the Open RAN architecture,” added Shenbagaraman.

Finally, employing a united and consistent management layer for the entire network will also ensure that the network architecture is simple enough to be deployed across different regions and at different cell sites.

Reducing variability

As mentioned previously, managing the vendor diversity of Open RAN is one of the technology’s biggest challenges. Therefore, Shenbagaraman encourages operators to exercise “diligence” early in the vendor selection process because having too many software/hardware combinations to work with can became “unmanageable” later on.

However, it takes more than just the operators to tackle the hurdle of multi-vendor integration. According to Shenbagaraman, there also needs to be strong “leadership capabilities of the system integrators to manage all the collaboration and coordination among the vendors … and the vendors also have to play their role beyond the technical compliance [because] seamless collaboration among vendors is required.”

“All of this is leading to a paradigm [that] is going beyond the traditionally thinking, which is the ‘single neck to choke’ approach,” he added.

Multi-stage product integration

Shenbagaraman said that the product integration approach is really about the performance of the network. “The network can perform[ance] [can be] somewhat [suboptimal] if the integration is not done right,” he cautioned. “We recommend a multi-stage integration approach.” He added that node level conformance and IOT and pair-wise and multi-node testing are required key elements to make ensure that the network foundation is done correctly. Similaly, end-to-end hardware/software integration and performance tuning are also very important.

Lastly, Shenbagaraman touched on the need for automation at every stage of the process. “All the way from the development stages, integration, lab testing, pre-production, deployment and post-production deployment optimization,” he said.