Manual processes are time-consuming and susceptible to errors, contributing to the time they take to complete. And that was before accommodating all the moving parts in next-generation mobile networks like centralized radio access networks (C-RAN), virtualized RAN (V-RAN), mobile edge computing (MEC), and the cloud.
Automating RAN improvements and optimization
The O-RAN SMO (Service Management and Orchestration) architecture accelerates and automates 5G deployments in next generation multivendor environments. The motivation is simple: implementing a single change to the 2G/3G/4G RAN can take as long as eight weeks, and requires three basic steps:
- collect and analyze data
- reconfigure the network
- monitor changes to measure results
Operations staff repeat these steps until they satisfy customers, or drive tests, measurement campaigns, and key performance indicators show expected results. 5G+ networks have a much faster metabolism than legacy networks and need intelligent automation and operations frameworks like O-RAN SMO to greatly simplify and automate manual processes.
Many mobile network operators (MNOs) use basic automated network operations in their 4G LTE implementations out of necessity. Self-Organizing Networks (SON) standards issued by organizations like 3GPP use computerized algorithms to plan, configure, control, heal, and optimize the RAN. For example, SON can improve call performance by automatically switching bands if a signal is weak. It can also help voice and data move seamlessly between user equipment, hubs, switches, routers, and the internet.
Advanced network automation goes cloud-native
5G+ networks exponentially increase the complexity of network planning, operations, and optimization, especially since most operators expect their legacy networks to coexist and potentially interact. It is impossible for a human being, or even teams of expert human beings with basic automation tools, to manage network parameter designs, tuning, operations, and optimization in terms of smooth coordination among all an operator’s networks in service.
The O-RAN Alliance goes beyond basic network automation by adding cloud-native cross-organizational data collection, ingestion, and artificial intelligence (AI) capabilities to 5G+ advanced network automation. This approach facilitates automated network operations and optimization in cloud-native 5G networks, as well as multivendor integration and exploitation. That’s not to say that SON technologies can’t or don’t benefit from AI. They can and do. However, most SON solutions weren’t born in the cloud, for the cloud, nor are they fully integrated into data ingestion, processing, and management techniques that are also cloud-native.
O-RAN Service Management and Orchestration (SMO)
The first blog in this open RAN software and API series describes O-RAN Open Front Haul and gNB disaggregation and virtualization that is cloud-native. It also discusses the interface between O-RAN xHaul and O-RAN Service Management and Orchestration (SMO), which is the orchestration and management framework defined by the O-RAN architecture. Disaggregated and virtualized 5G+ RAN architecture that supports multivendor environments requires a sophisticated system for programming, operating, and managing its billions of network components.
O-RAN SMO and real-time intelligent control
O-RAN SMO framework hosts the non-real-time radio intelligent controller (non-RT RIC) along with a series of open interfaces that supports network equipment and software interactions. The SMO framework also provides an interface to the near-RT RIC, which provides intelligent control for O-RAN CU/DU network elements. O-RAN SMO and real-time intelligent control work together to create a system that enables efficient and effective orchestration, control, and management of 5G+ RAN deployments and expansions.
RAN intelligent controllers are critical innovations of the O-RAN architecture and represent a significant difference from legacy systems where only proprietary vendors can provide new feature development, control, and optimization for the RAN. An open framework like O-RAN SMO, non-RT RIC, and near-RT RICs allows a healthier and more diverse ecosystem of vendors specializing in discrete elements of the RAN. That ecosystem depends on the notion that open and standard interfaces can be deployed in different network locations, integrate multivendor network equipment, and are compatible with multiple configurations.
Open interfaces and SMO
O-RAN SMO can work with intra-network components that enable multivendor gNB disaggregated architecture. That includes 3GPP interfaces like:
- F1-c/u- connects gNB CU to the gNB DU, allowing signaling between the central unit and the distributed unit. The F1-u transfers application data.
- X2-c/u- connects two eNodeBs
- Xn-c/u- connects ng-eNB to gNB
- NG-c/u- connects multivendor NG-RAN nodes to the 5G Core.
O-RAN SMO also defines native APIs that are open and standard and expose the data for machine learning and artificial intelligence. The O-RAN APIs surface telemetry and data analytics to the real-time intelligent controllers, allowing MNOs to control and automate multivendor equipment in the RAN. For example, using the O-RAN approach, one vendor’s near-RT RIC can interoperate with other vendors’ CU/DU/RU. Multivendor interoperability fosters innovation, faster upgrades, and competition and facilitates onboarding new RAN technologies.
Native O-RAN SMO interfaces are:
- E2- connects the near-RT RIC to RAN nodes
- A1- connects the near-RT RIC to the non-RT RIC
- 01- connects RAN components for management and orchestration
- 02- connects SMO to the O-Cloud.
O-RAN SMO and closed-loop control
The O-RAN SMO architecture offers an intelligent automation and operations framework provides sophisticated systems for programming, operating, and managing billions of network components. The framework includes O-RAN Service Management and Orchestration (SMO), real-time intelligent control, and open interfaces that enable multivendor equipment in the RAN to interoperate with each other. The O-RAN approach allows for innovation, faster upgrades, and competition and facilitates onboarding new RAN technologies.
O-RAN SMO microservices, delivered via real-time intelligent control (non-RT RIC and near RT RIC), enable closed-loop control and management powered by AI/ML. These features ensure an open RAN network’s proper functioning and address any issues during operations. The next blog in this series will explore close-loop control powered by non-RT RIC and near RT RIC and the applications that run them. Between now and then, you can learn more about O-RAN SMO and O-RAN Real-time Intelligent Control in this new whitepaper by Heavy Reading.