Docker unveils cross-cloud management, container templates
- 14 June, 2018 06:00
Docker has announced new features for Docker Enterprise Edition and Docker Desktop to port and manage apps on Kubernetes-based clouds, and to build containers via template-based workflows.
Both features are scheduled to appear in Docker editions released in the second half of the year.
Federated application management, as the first announced feature is called, is intended to unify the different ways each cloud provider and operating system runs Docker by way of Kubernetes.
Federation is intended to work interchangeably across multiple cloud providers, and between both Linux and Windows editions of Docker.
Clouds have varying migration, deployment, resource management, security, and replication behaviours.
Federated management is intended to sit between them all and provide “an aggregated view and automated model for deploying, migrating, and replicating applications,” according to Docker’s announcement.
One touted advantage of federated management is that it allows a single application to be partitioned automatically across multiple geographic locations for the sake of compliance in each region.
Each instance of an application can run in a different territory, with data for users of that country retained only in that country’s servers. But management for all instances of the application, across all territories, can still be centralised.
The other announced feature, template-based workflows for Docker Desktop, allows Docker Desktop users to build containers using a wizard-like GUI interface.
Developers do not have to learn Dockerfile or Compose file syntax to build basic containers using common stack components—web services, data layers, language runtimes, and so on.
The resulting containers can be edited manually by more expert users, or by the same user later after they’ve become more accustomed to Docker. It’s also possible for organisations to create and re-use their own templates.
(Reporting by Serdar Yegulalp, a senior writer at InfoWorld, focused on machine learning, containerisation, DevOps, the Python ecosystem, and periodic reviews.)