A Cloud-Native Approach for Orchestrating 6G-Enabled Services at the Computing Continuum

The cloudification of 6G services and the rise of edge and cloud computing have significantly increased the complexity of service lifecycle management, resulting in intricate networks deployed across interconnected infrastructures. The transition from traditional monitoring to modern cloud-native and network observability tools is crucial for providing a fine-grained view of application, infrastructure, and network performance. This paper proposes a novel orchestration methodology that enhances monitoring insights and effectively manages computing and network resources from the extreme edge to the cloud. Our approach encompasses synergetic orchestration mechanisms leveraging multi-agent systems, data fusion from various observability signals, and machine learning techniques wherever possible. Beyond conceptualizing and designing this methodology, we detail a tangible closed-loop architecture capable of supporting and realizing the proposed mechanisms. Finally, we validate a prototype implementation and evaluate its performance through a proof of concept involving realistic experimental trials, demonstrating our solution's effectiveness in orchestrating a 6G- enabled application with stringent latency requirements.