Distributed Intelligent Resource Orchestration
Résumé
The evolution to microservices and cloud-native technologies represents a transformative milestone in the telecommunications sector, offering the potential to unlock the capabilities of 5G/6G advancements. New cloud-based applications can be built from the ground up as containerized microservices, breaking a complex application into a series of smaller specialized and manageable services [1]. Efficient resource orchestration in cloud-native infrastructure is a critical challenge, ensuring uninterrupted service delivery. Kubernetes has emerged as de facto standard orchestration platform. However, its client-server architecture and storage on disk introduced latency which is not suitable for real-time communications. To address this gap, we propose DIRO (Distributed Intelligent Resource Orchestration), an orchestration platform for low latency microservices. Our proposal focus on a new scheduling architecture approach leverages in-memory data grid and distributed resource predictions to reduce scheduling delay.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |