Volcano is an open-source project used for running high-performance workloads on Kubernetes.
Kubernetes does not natively provide batch scheduling, which is something required by many workloads, such as machine learning and deep learning, bioinformatics, and other big data applications. This is where Volcano comes in.
According to Volcano, often those types of workloads run on frameworks like TensorFlow, Spark, PyTorch, or MPI, so Volcano offers integrations with them. This allows those workloads to be run with Volcano without much added work. Other integrations include Spark, Flink, Argo, and Kubeflow.
The most recent release as of this writing was 1.7, which was released in January 2023. It added new features such as an enhanced plugin for PyTorch jobs, enhanced scheduling for general Kubernetes services, multi-architecture images, and optimized queue status info.
The project was first open-sourced in 2019,accepted into the CNCF in 2020, and promoted to a CNCF incubating project in 2022. Over 490 developers have contributed code to the project.
