Abstract: Intelligent 5G L2 MAC Scheduler

Powered by Capgemini NetAnticipate 5G on Intel Architecture.

Key Takeaways:

  • With the objective of improving RAN functions, Capgemini Engineering launched Project Marconi, in collaboration with Intel, to develop an industrialized approach for introducing MP-assisted intelligence to RAN functions.

author-image

Par

Executive Summary

Capgemini Engineering, along with its 5G partner ecosystem, has developed a cognitive 5G Medium Access Controller (MAC), which is first in a series of activities planned as part of Capgemini’s Marconi Project. Marconi includes a series of innovations targeted towards realizing cognitive capabilities as part of Intelligent to RAN Controller nodes. These implementations use capabilities offered as part of Capgemini RATIO ORAN RIC powered with NetAnticipate5g and also integrated with Intel FlexRAN. The key idea here is to enhance network performance in terms of spectral efficiency, QoS, and network resource utilization. The Marconi team demonstrated ~15% improvement in network performance on 5G MAC scheduler and near-RT RIC platform. NetAnticipate5G, leveraging Intel AI software, Analytics Zoo, and optimized framework running on Intel® Xeon® Scalable processors, improved AI accuracy to 55% and reduced AI inference time to 0.64 msec, a 41% improvement, which is a critical requirement for such ultra-low latency and reliable ML-based solutions.

As industries move towards the adoption of Industry 4.0, they are demanding new, more advanced ways to provide services enabled by 5G networks.

As industries move towards the adoption of Industry 4.0, they are demanding new, more advanced ways to provide services enabled by 5G networks. Many of the advanced 5G-based services we foresee are focusing on private 5G networks, industrial IoT, digital twins, cobots, XR-based immersive gaming and many other applications. Most of these services are changing the way users will experience them. It is expected that many of these 5G-powered services will utilize technologies such as augmented reality (AR), holographic calls, immersive mixed-reality gaming, and even remote-controlled robotic arms. These new services will not only add immense complexity and load to network resources, but they will also require dynamic provisioning of advanced QoS traffic classes like Ultra Reliable Low Latency Communications (URLLC), massive Machine Type Communications (mMTC), and enhanced Mobile Broadband (eMBB), to enable the services. Also, it is almost impossible for CSPs to provision resources dynamically for such demanding services as part of the 5G network. This has led companies to start thinking about adding intelligence to their networks, which gives rise to alliances working on intelligent radio networks.

Capgemini Engineering is part of these alliances and is developing RIC platform enablers. We offer our award-winning NetAnticipate5G framework that introduces AI as part of the RAN intelligent controller (RIC) to support autonomous intelligent operations. Also, RATIO is a Capgemini Engineering RIC Platform based on O-RAN Alliance specifications for developing a disaggregated intelligent controller.