Comparison DCOS vs Kubernetes for Docker jobs
Jonathan Meyer edited this page Jan 24, 2018
·
1 revision
In the interest of comparing how Docker jobs are run between DCOS and Kubernetes, we are going to run 1000 jobs through a fresh clusters. We are primarily interested in the overhead of launching containers, so we are using a single node and queueing up all the jobs for the respective scheduler to run as quickly as possible. This will be done on clusters with all agent nodes deployed as m4.xlarge in AWS.
Kubernetes deployment: https://aws.amazon.com/quickstart/architecture/heptio-kubernetes/ DCOS deployment: AWS Node Type: m4.xlarge
- Home
- What's New
-
In-depth Topics
- Enable Scale to run CUDA GPU optimized algorithms
- Enable Scale to store secrets securely
- Test Scale's scan capability on the fly
- Test Scale's workspace broker capability on the fly
- Scale Performance Metrics
- Private docker repository configuration
- Setting up Automated Snapshots for Elasticsearch
- Setting up Cluster Monitoring
- Developer Notes