Orange and Ericsson Unveil AI-Driven 5G Slicing for Enhanced Flexibility
The telecommunications company Orange, in collaboration with network equipment provider Ericsson, is set to advance 5G slicing technology across Europe. This initiative aims to enhance network utilization and performance of 5G devices through the integration of artificial intelligence.
Orange's 'Service Orchestration and Assurance Platform' will facilitate dynamic end-to-end slice management, catering to various applications such as private mobile networks, geofencing, and fixed wireless access (FWA). The foundation of this system is based on Orange's AI-Network-Brain software, which promises modular expansion and customization to meet diverse customer needs.
Proposed use cases for this technology include critical infrastructure support, predictive maintenance, edge AI applications, robotic control, autonomous vehicles, drone inspections, and real-time video analytics. Orange guarantees upload rates and latency times of under five milliseconds, effectively creating a mobile connection comparable to fiber optics.
Orange emphasizes a new metric termed Quality of Delivery (QoD), which surpasses traditional Quality of Service (QoS) standards. This includes commitments to monitoring, latency management, jitter control, and reliability, extending beyond mere bandwidth considerations. In France, Orange has already introduced 5G contracts tailored for business clients, featuring packages such as 350 GB with Voice over NR (VoNR) and slice security.
Network slicing enables operators to segment a portion of network capacity for specific needs. Each slice operates on the same physical network yet remains logically isolated, allowing for unique resource allocation including radio access networks, transport networks, and core 5G infrastructure. This structure is akin to virtual machines, enabling tailored configurations for different user profiles or bandwidth requirements.
Two years ago, Deutsche Telekom launched its 5G network slicing solution specifically designed for live television production, aimed at reporters and camera operators in the field. This system utilizes the Telekom core network integrated with Mavenir MDCA and orchestrates customer interfaces to ensure high-definition transmissions through dedicated QoS slices and CAMARA open APIs. The service is now available as a premium offering for media companies.
Vodafone also provides 5G slicing solutions underpinned by SLA technology, primarily aimed at niche markets such as live events and sports broadcasting. O2 has introduced similar offerings since October 2023, focusing on business clients with plans to extend into consumer markets for gaming and smart home applications.
Despite the advances in slicing technology, a common challenge exists: the need to predefine requirements and subsequently delete slices when no longer needed. Although programmatic creation and deletion are possible via APIs, this can be cumbersome in dynamic industrial environments where bandwidth needs are sporadic and unpredictable.
A new application is emerging in the business sector, marked by an increase in mobile AI use, which leads to unpredictable data traffic patterns and varying bandwidth demands. To address this, Ericsson's latest technology aims to introduce an intent-based approach to dynamic slice allocation based on real-time demand.
Ericsson's solution automates the creation and replication of 5G slices, allowing the system to detect sudden spikes in demand and allocate resources accordingly. Once the demand wanes, the slice is automatically released, enhancing network flexibility and enabling a larger number of users to experience improved performance concurrently.
An illustrative application presented by Orange showcased AI-driven video generation on mobile devices, which requires significant bandwidth and low latency on demand. Once the video generation process concludes, the associated slice is released, demonstrating the potential for this technology to support not just mobile AI but also industrial applications such as machine control, robotic operations, and video analytics.
While specific references to successful implementations are still forthcoming, two notable test scenarios have emerged. One involved real-time drone operations in Belgium, showcased at the Mobile World Congress, utilizing the experimental Pikeo network, Orange's Telco Cloud, and AWS Outposts. Another collaboration with Schneider Electric focused on industrial process automation, developed in partnership with Nokia and Ericsson.