Cloud Native Batch Computing with Volcano - Updates and Future

Explore the latest updates and future developments of Volcano, CNCF's first container batch computing project, in this informative conference talk. Dive into the cloud native batch platform's capabilities, including full lifecycle job management, scheduling policies for batch workloads, heterogeneous hardware support, and performance optimization for high-performance workloads. Learn about Volcano's integration with computing ecosystems like Spark, Flink, Kubeflow, and Ray in big data and AI domains, as well as its deployment in production environments by over 50 users. Discover the recent progress made by Volcano contributors to address challenges in LLM training and inference, and gain insights into upcoming features designed to accelerate GPU/Ascend NPU training efficiency, optimize resource utilization for large-scale clusters, and provide fine-grained scheduling. Get an overview of the latest progress, new features, use cases, new sub-projects, and the future direction of the Volcano community.

Cloud Native Batch Computing with Volcano: Updates and Future - William Wang & Mengxuan Li