Bulk image processing using Kubernetes
Mike Newswanger explains how he used Kubernetes and Google Cloud to burst and extend the capacity of a physical infrastructure for optimizing almost 10 million images in less than two weeks.
| Talk Title | Bulk image processing using Kubernetes |
| Speakers | Mike Newswanger (Elastic) |
| Conference | O’Reilly Velocity Conference |
| Conf Tag | Building and maintaining complex distributed systems |
| Location | New York, New York |
| Date | October 1-3, 2018 |
| URL | Talk Page |
| Slides | Talk Slides |
| Video | |
Mike Newswanger explains how he used Kubernetes and Google Cloud to burst and extend the capacity of a physical infrastructure for optimizing almost 10 million images in less than two weeks. The process was running 3,600 parallel streams with near 100% compute utilization. Mike offers an overview of the infrastructure design, implementation process, performance analysis, and problems encountered along the way. He also shares real-world examples of how cloud compute resources and containers can boost the throughput of a traditional organization.
