AI Potluck
Back to Gap Map Infrastructure / ML orchestration

Volcano

Volcano
open source / Overall score: 3.2

Batch scheduling system for Kubernetes, hosted by the CNCF, aimed at AI, machine learning and big-data workloads that Kubernetes' default scheduler handles poorly. Volcano adds gang scheduling, queues with fair-share and preemption, and GPU-aware placement, so a distributed training job either gets all the accelerators it asked for or waits rather than deadlocking on a partial allocation.

Openness

5 high confidence
5.0
license
Apache-2.0(OSI)
source
public(the scheduler you run builds from this repo)
core features withheld
no — a 2,861-entry recursive tree carries one root LICENSE and no path matching enterprise, ee, commercial or proprietary

Volcano is published under Apache-2.0 with no commercial edition. A full scan of the repository tree finds a single root LICENSE and no enterprise, ee or commercial path, so there is no paid tier for the core to be gated against.

Adoption

2 low confidence
2.0

Volcano ships as container images and a Helm chart rather than through a package registry, leaving GitHub stars as the only adoption signal. A star is not a use, so the signal shows attention rather than real usage.

Capability

5 medium confidence
5.0

Volcano's own scheduler makes the placement decision for the pods in a PodGroup and admits them as one allocation, choosing the machines and reserving a multi-machine set together. It sits level with KAI Scheduler, which does the same thing, rather than above it.

  • https://github.com/volcano-sh/volcano recorded 2026-09-15

    Repository description and README for volcano-sh/volcano as a cloud-native batch system under the CNCF, the basis for the gang-scheduling and queue feature reading.

Verified 2026-09-15