Non-RT RIC (O-RAN)

Open RAN (Open Radio Access Network) is a telecommunications initiative that promotes the development of an open and interoperable RAN architecture. It disaggregates traditional monolithic networks by allowing network operators to mix and match components from multiple vendors. Open RAN aims to create flexible, scalable, and cost-efficient networks that support innovation and drive competition.

Open RAN primarily addresses the need for vendor-agnostic networks, enabling operators to deploy RAN hardware and software independently from various suppliers. This opens the door for increased innovation through AI/ML algorithms, real-time network optimizations, and the integration of cloud-native functions.

Explanation of the Open RAN Architecture

The diagram illustrates the Open RAN architecture and its main components and interfaces.

Here's a breakdown of the architecture:

1. Service Management and Orchestration Framework (SMO): The SMO is responsible for the overall management and orchestration of the RAN network. It interfaces with the non-real-time RAN Intelligent Controller (Non-RT RIC) and the near-real-time RAN Intelligent Controller (Near-RT RIC), coordinating the communication between various O-RAN network functions, including external systems that provide enrichment data to the SMO. External systems feed enriched information, such as data analytics or AI-driven insights, to the SMO, which is essential for optimizing network performance in non-real-time.
2. Non-Real-Time RAN Intelligent Controller (Non-RT RIC): The Non-RT RIC is responsible for performing non-real-time functions such as policy management, AI/ML model training, and optimization based on data gathered over time. The Non-RT RIC uses the A1 interface to communicate policies and enrichment information to the Near-RT RIC, guiding its actions in the RAN. It has an important role in handling long-term network optimizations and managing data from external sources for improving network efficiency.
3. Near-Real-Time RAN Intelligent Controller (Near-RT RIC): The Near-RT RIC handles near-real-time control functions, with a response time between 10 milliseconds to 1 second. It interacts directly with O-RAN network functions, adjusting resources such as radio frequencies, performing traffic steering, and balancing network loads in near-real time. The E2 interface connects the Near-RT RIC to the underlying RAN components like the Central Unit (O-CU) and Distributed Unit (O-DU). This interface is used to enforce policies generated by the Non-RT RIC, which allows efficient network operation based on real-time data.
4. O-RAN Network Functions: The core O-RAN network components include the O-RU (O-RAN Radio Unit), O-DU (O-RAN Distributed Unit), and O-CU (O-RAN Central Unit), which handle different parts of data processing and transmission within the network. These units support a disaggregated, modular structure that allows for flexible deployment and upgrades. These functions interface with both the Near-RT RIC and the 5GC (5G Core) for the seamless management of user data, traffic routing, and signaling.
5. O-Cloud: The O-Cloud is a virtualized infrastructure designed to host O-RAN network functions and provide the computational resources required for handling network tasks. This infrastructure is essential for running AI/ML models and other software components that drive real-time optimizations.
6. Y1 Consumers: The Y1 interface provides RAN analytics services, allowing external systems (Y1 consumers) to access real-time network data. These consumers may include third-party applications or external service providers that require insight into the network's status for making informed decisions.

Interfaces in the Open RAN Architecture:

• A1 Interface: This interface enables the exchange of policies and enrichment information between the Non-RT RIC and Near-RT RIC. It supports AI/ML-driven policy enforcement and optimization, enabling more intelligent network management.
• E2 Interface: Connects the Near-RT RIC with O-RAN nodes like the O-CU and O-DU. This allows real-time data collection and optimization actions (such as traffic steering or load balancing) to be performed by the Near-RT RIC.
• O1 Interface: Used for managing and monitoring the various O-RAN network functions, the O1 interface provides the SMO framework with the data it needs to perform operations such as configuration, fault management, and performance monitoring.
• O2 Interface: Provides a communication link between the O-Cloud (which hosts virtualized network functions) and the SMO framework, allowing for resource management and orchestration of cloud infrastructure.

What is Non-RT RIC?

The Non-Real-Time RAN Intelligent Controller (Non-RT RIC) is a key component within the Open RAN architecture. It is responsible for non-real-time functions such as long-term network optimization, policy management, and AI/ML model training. Unlike the Near-RT RIC, which operates within milliseconds to seconds, the Non-RT RIC operates over a longer time scale (seconds to minutes), making decisions that affect the overall performance and efficiency of the RAN.

Key Features of Non-RT RIC:

• AI/ML Workflow: The Non-RT RIC is capable of running machine learning models for optimization purposes. These models are used for training and inference to improve the performance of the network.
• Policy Management: The A1 interface communicates policies to the Near-RT RIC, which in turn guides the Near-RT RIC in real-time network optimization.
• Data Enrichment: The Non-RT RIC can receive and process external data from other systems, known as enrichment data, which can help optimize network performance.
• Interfaces: The Non-RT RIC interacts with other components through the A1, O1, and O2 interfaces to perform its functions.

?

?

Service Management and Orchestration (SMO) Framework

• SMO Framework is responsible for managing and orchestrating the entire Open RAN network, ensuring the effective operation of the Non-RT RIC and the Near-RT RIC.
• O2, O1, and M-Plane Terminations:
• O2 Termination: Handles communication with the O-Cloud infrastructure.
• O1 Termination: Manages the operations and maintenance of RAN elements, including configuration and fault management.
• Open FH M-Plane Termination: In charge of front-haul management, ensuring efficient data transfer between the Radio Unit (O-RU) and the rest of the network.

2. Non-RT RIC Framework

• The Non-RT RIC sits within the SMO framework and is responsible for long-term network optimization, AI/ML model training, and policy management.
• It communicates with external data sources, the Near-RT RIC, and other RAN elements to improve network performance over time. Below are its major functions:
• R1 Service Management and Exposure Functions: Handles the R1 interface that communicates with rApps (applications running in the Non-RT RIC) and exposes their services to other parts of the RAN.
• rApp Management Functions: Manages the lifecycle of rApps, such as launching, updating, and terminating them. These rApps use real-time data to make decisions about the network.
• A1 Termination: The A1 interface connects the Non-RT RIC with the Near-RT RIC, allowing the Non-RT RIC to send policies and enrichment data for near-real-time optimization.
• AI/ML Workflow Functions: Enables the training, deployment, and monitoring of AI/ML models that optimize network performance, such as traffic steering, load balancing, and fault detection.
• Data Management and Exposure Functions: Collects, stores, and exposes data necessary for training AI/ML models and making network decisions.
• External Terminations: Communicates with external systems (such as enrichment data providers or AI/ML service providers) to enhance the quality and intelligence of network decisions.

3. External Systems and Oversight

• External systems provide critical input to the Non-RT RIC to improve its optimization and policy management capabilities. The diagram shows different types of external systems:
• External EI Sources: These are sources of enrichment information (EI) that can be fed into the Non-RT RIC to help it make better decisions.
• External AI/ML Services: These services provide additional machine learning models or data for training and optimization purposes.
• External Oversight: This layer can monitor and oversee the activities of the Non-RT RIC, ensuring compliance with broader network goals and policies.

4. Near-RT RIC, E2 Nodes, and O-RU

• Near-RT RIC: The Near-RT RIC handles real-time control and optimization (within milliseconds to seconds) of the RAN elements, including the E2 nodes and the O-RU (Radio Unit). It receives policies and enrichment data from the Non-RT RIC through the A1 interface.
• E2 Nodes: These nodes (which include O-CU and O-DU) handle specific RAN functions like packet data management and radio control. They communicate with the Near-RT RIC to optimize resource use.
• O-RU: The O-RAN Radio Unit handles physical signal processing and communicates with the 5G Core (5GC) to deliver user data to and from the network.

The primary difference between Near-RT RIC and Non-RT RIC

Key Differences:

Time Sensitivity

1. Near-RT RIC operates on a short time scale and is responsible for managing RAN resources in near real time, typically in milliseconds to seconds.
2. Non-RT RIC works on a longer time scale (seconds to minutes) and focuses on policy management, training AI models, and making optimizations that are not as time-critical.

Functionality

1. Near-RT RIC handles tasks that require quick decisions, like traffic steering and resource allocation to manage immediate network demands.
2. Non-RT RIC focuses on long-term network improvements, such as creating policies based on historical data, training AI models, and managing rApps that help optimize the network over time.

Scope

1. Near-RT RIC typically operates at a finer level of granularity, such as user-level or cell-level optimizations.
2. Non-RT RIC operates at a higher level, focusing on the overall network's long-term performance.

Interaction:

1. Near-RT RIC interacts with xApps for executing real-time tasks.
2. Non-RT RIC interacts with rApps, which provide higher-level control and management of the network's performance.

?