November 13–15, 2018 - Shanghai, China
Click Here For Information & Registration

To view the Chinese version of this schedule please go here.

Simultaneous translation will be provided for all keynote and breakout sessions.
Wednesday, November 14 • 11:05 - 11:40
Kube-Arbitrator: A Batch System of Kubernetes - Da Ma, IBM & Ti Zhou

Sign up or log in to save this to your schedule and see who's attending!

Feedback form is now closed.
Kubernetes started as a general purpose orchestration framework with a focus on serving jobs. But as it gains popularity, users want to run BigData and AI frameworks on Kubernetes, such as Spark, TensorFlow, etc. When running these batch processing frameworks in a multitenant system like Kubernetes, advanced scheduling features, e.g. fair-share scheduling, gang scheduling, become more important. In 2017, we launched a project into incubation, named kube-arbitrator, to provide policy-based resource scheduling for BigData and AI frameworks to meet these advanced scheduling and resource sharing requirements.

In this presentation, we will describe how we're trying to do batch scheduling in Kubernetes, and our journey building these components out of core.

avatar for Klaus Ma

Klaus Ma

Expert, Huawei
Kubernetes Maintainer, SIG-Scheduling Co-Leader, CNCF Research User Group Tech Lead, Volcano/kube-batch creator. Jilin University master’s degree, majoring in grid computing and distributed system. After graduation, he focus on resource management, resource scheduling in distributed... Read More →
avatar for Ti Zhou

Ti Zhou

Architect, Baidu
Ti Zhou, Kubernetes member, currently serves as software architect in Baidu Inc, focusing on PaddlePaddle Deep Learning Framework and Baidu Cloud Container Engine, helps developers to deploy cloud-native machine learning on private and public cloud.

Wednesday November 14, 2018 11:05 - 11:40
2F Room 3
  • Skill Level Any