By Ladan Pickering, Professional Services 5G Practice Lead, Fujitsu
Adoption of Open RAN infrastructure is heating up, with analysts predicting it will account for more than 5 percent of the overall RAN market this year. As a result, network operators are looking forward to promised improvements in operational efficiency and resiliency due to the freedom to choose the best vendors for each function or service from across the ecosystem. Yet in reality, multivendor network integration is not a simple matter of plug-and-play, despite industry standards and defined O-RAN ALLIANCE interoperability parameters.
As with any network deployment, system integration and testing is an important part of deployment and provisioning. However, unlike legacy proprietary infrastructure, a network built with hardware, virtual infrastructure and networking software from multiple vendors across open interfaces significantly complicates set-up, testing and performance verification. Add in the complexity of the disaggregated 5G New Radio (NR) vRAN architecture, and Open RAN system integration becomes quite challenging and time-consuming.
Facing the Facts
Deployment and integration of an Open RAN network requires a deep engineering knowledge of Open RAN specifications, 3GPP standards, virtualization, RAN transport and software. Challenges range from integration of open interfaces across multivendor infrastructure and various software versions, to dealing with security risks, and integrating the RAN into the service management and orchestration framework — not to mention the management of a large supply chain with varying lead times, which has become particularly convoluted since COVID.
Of all the moving parts involved in O-RAN system integration, arguably, interoperability testing is the pivotal component for success. Failure to conduct sufficient validation testing impacts not only network performance and quality of service (QoS), but also slows build-out and reduces competitive advantage, delaying return on investment (ROI).
Of course, in an open network there is no single vendor to be accountable, so all of the work and responsibility falls on the network operator and a third-party system integrator. In some cases, vendors will have different interpretations of the O-RAN standards, so both the operator and system integrator need a deep knowledge of the standards in order to solve network issues. Integrators possessing that deep knowledge base will be able to work with vendors to create alternate solutions and resolve conflicts.
From lab testing of integrated software and components before roll-out, to field testing of operability and performance before turn-up, the entire process of validating interoperability has become so complex that a traditional manual testing approach is no longer efficient or feasible. At this point, test automation is critical to ensuring interoperability and performance in an open, multivendor network, enabling operators to fully realize the promise of Open RAN.
Furthermore, by introducing automation, network operators can radically improve efficiency of operations and better manage the entire O-RAN lifecycle, from deployment to service delivery. Since most of today’s RAN infrastructure is virtualized, continuous O-RAN software delivery across commercial-off-the-shelf (COTS) hardware adds value to the network with new features and capabilities. This continuous improvement facilitates increasingly automated RAN operations, making traditional manual testing of the network nearly impossible.
Automate, Instantiate, Repeat
As network operators transition to 5G architectures, a considerable amount of capital is being invested in new spectrum licenses, infrastructure, RAN deployment, software and cloud expenses in order to keep the network operational. A number of new over the top (OTT) applications running on the network benefit from advanced network technologies; however, operators are not benefiting from OTT revenue in proportion to their investments.
The O-RAN architecture defines the non-real-time RAN intelligent controller (Non-RT RIC) and near real-time RIC (Near-RT RIC) to allow third-party applications to run in the network infrastructure instead of OTT. Plus, integration of the Near-RT RIC with xAPPs allows low-latency and RAN optimization applications to run at the edge of the network, in addition to multi-access edge computing (MEC) applications.
RIC and x/rAPPs allow operators to host multivendor, revenue-generating applications in a cloud-native environment to unleash the full potential of an intelligent network leveraging artificial intelligence (AI) and machine learning (ML). Yet before new software applications can be introduced, network engineers need to instantiate the software, and then validate its quality and integrity. With RAN automation, a disaggregated test engine can be created to automate software testing across the network.
This decentralized pipeline can perform multiple tests across various topologies, with a test engine instantiated on the edge of the network, or automatically instantiated from the cloud. The test engine detects software delivery and automatically instantiates the correct set of test cases by accessing physical or container-based components.
In this way, network operators can take advantage of continuous software delivery to increase revenue opportunities and optimize network performance without increasing operational costs. With time, the AI/ML engine will automatically narrow down the root cause of problems and potentially repair faults.
Catalyst for Change
The transition to Open RAN architecture is sparking a catalyst for change and innovation, offering greater flexibility, efficiency, programmability and cost reductions. Operators will see these benefits from disaggregation, once the networks are operational and low-cost, revenue-generating applications have been deployed. In the meantime, however, the challenges of service assurance and interoperability testing across open interfaces also bring the need for increased coordination, diligence and transparency. There is no time for a ‘plug-and-pray’ strategy.
Network operators need to fully understand and accept the complexities so they can proactively take control of integrating and validating disaggregated infrastructure from multiple vendors. But they don’t have to face the future alone, because the Open RAN ecosystem is built on shared specifications, use cases, reference architectures and test cases developed with a variety of system integrators and vendor partners.
As today’s network becomes ever more complex, increasing reliance on network test automation is no longer optional; it is the key to enabling the next step in Open RAN evolution. Because getting off to a poor start with Open RAN 5G roll-out not only means that short-term RAN health and QoS will suffer, but long-term network growth and profitability will be in serious jeopardy as well.
Ladan Pickering is the Professional Services 5G Practice Lead at Fujitsu. Her areas of expertise include cloud-native O-RAN test automation, near real-time RAN Intelligent Controller, and xAPPs for operational efficiency and revenue generation. She has held a variety of positions at Fujitsu and other communications network companies, including software strategist, system engineer, technical lead, and technology specialist. Ladan has a bachelor’s degree in computer science from the University of North Texas, and a master’s degree in computer architecture from Southern Methodist University in Dallas. She holds three patents.
Want to hear the latest developments in the Open RAN debate? Join the operators in discussion at this year’s live Total Telecom Congress in London