Mesosphere DC/OS 1.9 takes production container operations to the next level
Keep Calm and Take It To The Next Level message image via Shutterstock
Mesosphere DC/OS 1.9 adds three key new sets of features to simplify the process of building and deploying data-rich applications.
Spoiler alert: Mesosphere DC/OS 1.9 is currently in early access (rc2 was released in early March) but it should be GA before the end of this month.
This release includes quite a few new capabilities for Operators and expands the collection of Data and Developer Services with a focus on:
- Tools for Production Operations – Monitoring and troubleshooting for distributed apps.
- Broader Workload Support – From traditional apps to machine learning.
- New data and developer services.
As of version 1.9, new data services can now be quickly deployed and run elastically on Mesosphere DC/OS, accessible from the DC/OS Universe service catalog.
According to the blog post announcing Mesosphere DC/OS 1.9, the team has broadened workload support with GPU-based scheduling and Pods (thus allowing you to easily launch and manage sets of both traditional Linux and Docker containers within DC/OS). Furthermore, version 1.9 enables operators to give developers access to debugging and associated apps while limiting access to the underlying infrastructure.
Docker 1.12 and 1.13 are now supported — the latter is the default version.
Machine learning: There’s no escaping it
Mesosphere’s Edward Hsu explained that the difficulty of deploying traditional N-tier applications on containers makes teams lose out on the benefits of containerization with legacy apps. But there’s a solution for that!
It’s easier to scale these applications with the help of pods (within DC/OS). Their benefit is that they enable the web, app, and DB tiers to share hosts while using dedicated IPs, always running on hosts attached to their storage volumes. As a result, users can manage their traditional apps within DC/OS to easily scale to any capacity.
With GPU acceleration, neural net training can be 10-20 faster than with CPUs. DC/OS now allows you to isolate and reserve GPU resources for the workloads that need them, but also pool these GPU-enabled resources with the rest of the infrastructure for higher overall utilization, Hsu added.
For more details about Mesosphere DC/OS 1.9, check out the release notes.
The secret behind DC/OS
DC/OS is the only container orchestration platform that already started powering Container 2.0 operations in production environments at some of the largest companies in the world, such as Time Warner Cable and Verizon. Mesosphere CEO Florian Leibert revealed in a blog post that “the secret sauce behind how DC/OS is able to power such robust applications really is […] the open source Apache Mesos technology on which it is built. All of the services running on DC/OS can have their own scheduler and that each scheduler can be specifically optimized for unique kinds of workloads and constraints.”
DC/OS can also run various important services on the same cluster. Jenkins, Elasticsearch and BitBucket are just a few services available in the DC/OS Universe.