Accelerate 5G Deployment with AI: Cellwize & Intel Joint Demo

Table of Contents

1. Introduction (H2)
2. Overview of CellWise AI Demo (H2)
   • CellWise Chimes Map View (H3)
   • CellWise AI Virtualized DU's (H3)
   • CellWise Chime Ingests Configuration and Topology Data (H3)
   • Open APIs and Cell Level Model (H3)
   • Key Performance Measurements (H3)
3. Performance Analysis (H2)
   • Average User Throughput (H3)
   • Cluster-Level PRB Utilization (H3)
   • Cell-Level PRB Utilization (H3)
4. Cell Load Classification (H2)
   • Unsupervised Learning Techniques (H3)
   • Auto Classification Load Scores (H3)
   • Highly Loaded Cells and New Site Deployment (H3)
5. AI-Driven vRAN Deployment Automation (H2)
   • Instantiation of New VNF (H3)
   • Configuration and Activation of New Cells (H3)
   • Pre-Activation Recipe and Parameter Configuration (H3)
   • Data Fill Enforcement and Network Design (H3)
6. Benefits of AI-Powered Deployment (H2)
   • Accelerating 5G Deployment (H3)
   • Reducing Deployment Cost (H3)
7. Improved Performance After Deployment (H2)
   • Increased Average User Throughput (H3)
   • Decreased PRB Utilization (H3)
   • Enhanced User Experience (H3)
8. Conclusion (H2)

Introduction

Welcome to the Intel CellWise AI Demo. This article aims to demonstrate how you can leverage AI to power end-to-end deployment of a new site while accelerating your journey to 5G.

Overview of CellWise AI Demo

CellWise Chimes Map View The CellWise Chimes Map View provides a visual representation of the demo cluster. The current site cluster contains four virtualized DU's powered by AI running on Intel Xeon Scalable processors.

CellWise Chime Ingests Configuration and Topology Data CellWise Chime ingests the configuration and topology data from Intel FlexRAN. This data is transformed into an abstract data model, which is also available via open APIs. The cell level model contains topology information such as height, azimuth, and location, as well as configuration data such as frequency band, RSI, and PCI.

Key Performance Measurements Intel FlexRAN ingests key performance measurements into CellWise Chime's hourly KPIs. These measurements are aggregated to the cluster level and displayed in CellWise analytic dashboards. The average user throughput of the entire foresight cluster is low, especially during the busy hours, with average physical resource block or PRB utilization of the cluster reaching 70 percent.

Performance Analysis

Average User Throughput

Looking at the PRB utilization at the cell level, it becomes evident that a few of the cells in the cluster reach almost 90 percent PRB utilization during the busy hours. This indicates a heavy load on the network.

Cluster-Level PRB Utilization CellWise AI uses unsupervised learning techniques to auto-classify cell load. The AI auto-classification generates a load score for each cell, categorizing them as low, mid, or high. In this case, the highly loaded cells suggest that a new site should be deployed in that area.

AI-Driven vRAN Deployment Automation

Upon physical installation of the new RUs (Remote Units), an AI-driven vRAN deployment automation flow is initiated. CellWise Chime triggers the instantiation of the new VNF (Virtual Network Function) and creates the new network element. The new DU (Distributed Unit) appears on the map but is currently inactive.

Pre-Activation Recipe and Parameter Configuration A pre-activation recipe is automatically triggered to configure the new VDU (Virtual Distributed Unit) cells. CellWise AI-driven applications perform Rand parameter configuration just in time, minutes before the cells go live. This ensures accurate configuration based on the actual current environment, saving engineering planning effort.

Data Fill Enforcement and Network Design During the data fill enforcement step, the network design is enforced to the new DU, ensuring the cells are ready to provide the best performance for end users. Once the configuration is completed successfully, the new cells will be activated, and the automation process is completed. The entire process takes up to 15 minutes, enabling MNOS to accelerate their 5G deployment and reduce deployment costs.

Benefits of AI-Powered Deployment

The AI-powered deployment approach offers several benefits:

1. Accelerating 5G Deployment: With the automation and optimization provided by CellWise AI, the deployment of new sites becomes faster and more efficient, allowing for quicker expansion of 5G networks.
2. Reducing Deployment Cost: By leveraging AI-driven algorithms for configuration and optimization, MNOS can reduce manual planning effort and minimize the resources required for deployment.

Improved Performance After Deployment

After the deployment of the new site, notable improvements in performance are observed:

• Increased Average User Throughput: The aggregated average user throughput in the cluster increases by 15% during the busy hours, indicating a better experience for users.
• Decreased PRB Utilization: The average PRB utilization is now decreased to less than 60 percent, suggesting a more balanced network load.
• Enhanced User Experience: The reduction in PRB utilization from 90 to 70 percent in loaded cells indicates an improved user experience across the entire cluster.

Conclusion

The Intel CellWise AI Demo showcases the power of AI in accelerating 5G deployment and optimizing network performance. By leveraging AI-driven algorithms for site deployment and optimization, MNOS can achieve faster and more efficient site deployment while improving the user experience on their networks.

ℹ️ Resources:

• Intel CellWise AI Demo

Highlights

• Leveraging AI for end-to-end deployment of new sites in 5G networks
• CellWise Chimes Map View: Visual representation of the demo cluster
• Cell load classification based on unsupervised learning techniques
• AI-driven vRAN deployment automation for quick and efficient site deployment
• Benefits of AI-powered deployment: accelerating 5G deployment and reducing cost
• Improved performance after deployment: increased user throughput and reduced PRB utilization

FAQ

Q: How long does the AI-driven vRAN deployment automation process take? A: The entire process takes up to 15 minutes, allowing for rapid deployment of new sites.

Q: What are the benefits of AI-powered deployment? A: AI-powered deployment accelerates 5G deployment and reduces deployment costs by automating and optimizing the site deployment process.

Q: How does CellWise AI improve network performance? A: CellWise AI utilizes unsupervised learning techniques to classify cell load, optimize resource utilization, and enhance the overall user experience.

Q: Can CellWise AI be used for existing site optimization? A: Yes, CellWise AI can be utilized for continuous optimization of existing sites, ensuring optimal performance and resource allocation.

Q: Are there any specific hardware/software requirements for using CellWise AI? A: CellWise AI runs on Intel Xeon Scalable processors and utilizes Intel FlexRAN for data ingestion.