Are you looking for ways to deliver your applications more quickly and with less effort? Do you need to move more services to the cloud?
Red Hat has these same issues and goals. By using our own product OpenShift, we have been able to shorten our release cycle times and support our developers better.
We built and deployed our own OpenShift Enterprise Platform-as-a-Service (PaaS) offering and are running it in Amazon Web Services (AWS). This has allowed our teams to have more control over creating and developing their own applications in the cloud. OpenShift’s automated workflows and tooling help our developers access what they need, when they need it.
We hope our success story can become your success story with OpenShift.
Case Study: IT OpenShift Enterprise
In 2013, Red Hat IT launched a new internal PaaS deployment, available to all associates and based on OpenShift Enterprise.
Before the launch of IT’s OpenShift Enterprise offering, IT did not distinguish much between major enterprise applications and smaller applications that do not require strict enterprise management and change control. Plus our release engineering tooling was falling behind the modern standards.
While management processes are valuable in many cases, our Red Hat associates were asking for a self-service, application-deployment platform for lighter workloads. At the same time, IT wanted to:
- shorten our release cycle times
- support developers
- expand our services to the cloud
Because we needed multiple application security zones and platform customizations (for some enterprise applications), we could not move into OpenShift Online.
IT’s Cloud Enablement team (in collaboration with OpenShift Engineering, IT Operations, and Global Support Services (GSS)) built and deployed a 90-day, proof-of-concept OpenShift Engineering PaaS offering, running in AWS, within 5 weeks. Developers and teams began creating and managing their own applications immediately and moved us towards our goal of running more services in the cloud.
At the end of the proof-of-concept phase, IT delivered additional releases on the platform. IT also made it a production service, expanded the number of available security zones, and added other new features.
The IT OpenShift service currently supports 3 major security zones, each with infrastructure to support 3 different application sizes (small, medium, and large districts). These zones are spread across multiple AWS regions and accounts to ensure higher availability and security. The security zones offer different options depending on the intended customer base and application security requirements, including:
- internal/intranet only
- externally accessible
- IT development
The OpenShift Enterprise PaaS provides a high level of availability through load-balanced broker services, MongoDB replica sets and redundant components at every level. Individual application owners can now deliver application high-availablilty within the platform.
IT team members included both associates who had previously worked on OpenShift Online and associates who had never before touched OpenShift. To facilitate cooperation, the IT team worked closely with the OpenShift Engineering teams, some of whom were former IT team members.
This unique setup allowed for a rapid flow of information between IT and Engineering. The teams met on a regular basis while implementing the project. This collaborative environment made it easier to discuss and deliver feature requests, bugs, and patches.
As we began to form our OpenShift Enterprise support team for our external customers, GSS was able to work closely with IT and use this effort as a training ground for our support associates. Our internal partnership allowed Red Hat to serve our current and future customers even better.
- 31 Amazon Elastic Compute Cloud (EC2) nodes in AWS for Openshift infrastructure (brokers + app nodes)
- 7 EC2 nodes in AWS for supporting Openshift infrastructure
- 1038 Red Hat associates logged into the system
- 1093 Applications deployed
- 618 Domains
- 11 Bugs opened with Engineering
- 5 Feature requests opened with Engineering
- Access Labs
- Partner Enablement Training for JBoss BPMS6 and FSW
- SalesForce.com connector to sync our Partner Center Salesforce data to a SaaS (Software-as-a-Service) Learning Management System
- Conference room reservation mobile application
And many others…
- Red Hat IT’s standard Nagios active monitoring, plus 5 custom scripts to monitor broker services.
- Red Hat IT’s standard Graphite (for data collection and trending)
- Red Hat IT’s standard MongoDB monitoring tools
- AWS CloudFormation Stacks (for provisioning EC2 infrastructure across availability zones in 3 Virtual Private Clouds (VPCs))
- Custom administrative scripts to manage new dynamic cloud services
The Lessons Learned
- Agile (Scrum) methodologies are effective for managing an OpenShift implementation.
- Building services in the cloud is significantly different from building them in a datacenter. But with proper tooling and understanding, the cloud model offers many opportunities.
- Working directly with GSS and Engineering helped us make this project more visible throughout the company.
- Mojo can provide a good place to manage documentation and provide a support portal for our users.
- Developers waited to do most major user application development until the Production PaaS service was available. Users want to know their platform will be supported.
- A flexible and stable PaaS platform is now available to all Red Hat associates.
- OpenShift Enterprise allows us to customize the platform as needed.
- Developers can now code, test, and deploy applications directly into our internal OpenShift Enterprise system with minimal paperwork.
- We now have more services running in the cloud and more experience with the cloud model.
We will soon be moving to OpenShift Enterprise 3, with Docker and Kubernetes technologies built in.
We are also working to integrate Red Hat Atomic into our PaaS offerings.
Read more about how OpenShift is helping customers on our Customers Page.
About the Author
Hope Lynch is Product Manager and Product Owner in Red Hat IT, supporting internal IaaS and PaaS offerings.
The original article was published on the Red Hat Customer Portal Blog.