Consciously Uncoupling from your Cloud Provider with Ansible

For those that don’t pay attention to Hollywood news – Conscious uncoupling – a five-step process to “end your romantic union in honorable, respectful, and gracious ways” was made popular by Chris Martin and Gwyneth Paltrow. I submit that conscious uncoupling might be something companies should consider when it comes to their cloud provider. After all, it’s not that you aren’t grateful and don’t respect your cloud provider, but why commit to a single cloud provider when there are so many fish in the sea!

All hollywood references aside, uncoupling from a proprietary platform is something the Global Partner Technical Enablement (GPTE) team at Red Hat is undertaking. In 2013, the GPTE team began building a learning platform to allow sales engineers, consultants, and select partners to perform hands on technical training in order to understand how to demonstrate and implement Red Hat’s growing portfolio. The GPTE team began using Ravello Systems in order to deploy virtual environments for trainees. Ravello Systems provided capabilities such as Nested Virtualization, Overlay Networking and Storage that were needed for much of the Red Hat portfolio (particularly the infrastructure technologies) to function properly. The team used Red Hat CloudForms to provide self-service with automatic retirement of applications on the Ravello System.

Fast forward to 2016 and the GPTE team’s Red Hat Product Demo System (RHPDS) runs several hundred applications that are comprised of thousands of virtual machines concurrently. It has been used to teach thousands of Red Hatters and partners in the field how to demonstrate and implement Red Hat’s technologies successfully. However, there have been two key challenges with this system.

First, the Ravello System uses a concept called a blueprint to describe an environment. The blueprint is a concept native to the Ravello System and not something that is supported on any cloud provider. Any logic put into the blueprint is by definition not portable or usable by other teams at Red Hat that don’t use the Ravello system. This runs counter to the culture of open source at Red Hat and does not allow contribution to the demonstrations and training environments to flow back from participants. The team needed to turn participants into contributors.

Second, demonstrations developed on the Ravello system were limited to running only on the Ravello system and could not be deployed in other labs across Red Hat, or even customer environments. This severely limited buy-in from other groups that had their own labs or otherwise felt more comfortable learning in their own way. Many field engineers at Red Hat and partners run OpenStack, oVirt (RHEV), VMware, or other virtualization platform in their labs. These users should be able to deploy demonstration and training environments on the provider of their choice. The team needed to allow re-use of demonstrations and training environments across cloud providers.

The GPTE team wanted to address these two issues in order to increase reuse and spur contribution to the demonstrations themselves from the community of sales engineers. They found an answer in Ansible. By including Ansible in the GPTE platform it will be possible to separate the configuration of the blueprint in the Ravello system from the configuration of the demonstration or training environment. It will also allow automated update of the environments in Ravello any time a code change is made. The result – field engineers can re-use any of the demonstrations and training environments created on the GPTE platform in their own labs and can even share fixes or improvements back. This small change will lead to a greater amount of user acceptance and lower the burden of building and maintaining the technical enablement platform at Red Hat.

If you are interested in learning more about how the GPTE team is consciously uncoupling themselves from a proprietary description and automated the process or if you are interested in deploying the demonstrations in your own lab please check out the Red Hat Demos repository on GitHub where our work is in progress. Contributions welcome!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: