Attendees of SHARE Atlanta engaged in open and collaborative discussion about Kubernetes (K8s), containers, open source on the mainframe, and much more. From Bring Your Own Device labs to sessions on the environmental sustainability of today's technology and the need for enterprises to incorporate it into their own business continuity planning, attendees experienced a wide range of thought leadership and practical advice for their day-to-day work. According to Brent Ellis at Forrester, SHARE Atlanta was focused on opening technologists' eyes to the flexibility and open collaboration available on the mainframe.
Open Source, the Right Fit for Mainframe
As John Mertic, executive director of the Linux Foundation’s Open Mainframe Project, said during his session, "How Open Source Modernized the Mainframe," the mainframe community has been open to collaboration since its inception and the creation of COBOL. It was built from different languages at different companies by different users and the CBT Tape, an open library of free software distribution for IBM's mainframe environments. "Mainframers have long looked at how they can pull code together to benefit the entire community," he said.
Current open source efforts merely continue the groundwork laid by mainframe communities of yesteryear, he explained. With Open Mainframe Project’s Zowe and cloud, mainframe is driving faster innovation and connecting to other departments and projects in an enterprise-wide effort to maximize the efficiency of the mainframe and leverage it in other areas. Mertic says, "The mainframe community's ethos of collaboration and sharing make it the right fit for enterprises." The Open Mainframe Project is just one effort to open up this collaborative world and allow volunteers to contribute code, offer fixes, and address problems with technology.
According to Mertic, "open source's relationship with mainframe will continue growing over time because it is an enabler for enterprises and it can be the glue that connects mainframe to the rest of the business. This integration is for the good of the business."
Containers, Clusters, and Management, Oh My!
In "The Good, The Bad, and The Ugly of K8S and IBM Z" with Broadcom's Distinguished Engineer Matt Hogstrom, Arron La, UI architect, and Aaron Young, principal software engineer, and "Managing Container Deployments with Kubernetes Using SUSE Rancher" with Michael Friesenegger, SUSE Solution Architect for global IBM Partnership, SHARE Atlanta attendees explored how to embark on a container deployment with Kubernetes (k8s) and how to manage those clusters with SUSE Rancher.
Hogstrom, La, and Young provided a how-to look at k8s deployment, cautioning that it is not a quick 45-minute install on zLinux or zCX. They cautioned that before deploying k8s on an enterprise platform with Red Hat®, technologists will want to read all the documentation and what elements are needed before installing k8s.
For its zLinux k8s install, Broadcom set up three worker nodes for the containers, Red Hat, HTTP server, DSN entries, and more, but soon realized it would need a separate node for the bootstrap system, which after installation could then be transformed into another worker node. They reminded attendees that they need the appropriate amount of hardware to back LPAR so that clusters will boot up, operate, and add users. They also advised that enterprises get signed certificates for installation on the cluster. Once properly installed, k8s can help enterprises administer containers and attract developers to the mainframe, though images will need to be created, backups installed, and limits set on projects so developers don't take up all the space in the cluster for one project.
When installing on zCX, Broadcom found the process easier, but it may be tied to the knowledge gained along the way during the zLinux install. Reading all the installation, power, and storage requirements and documentation can make the process smoother. Once k8s are installed, they can help find service crossover and leverage them for the good of the enterprise.
Michael Friesenegger's session focused more on deploying and managing microservices applications with k8s using SUSE Rancher, which provides a graphic user interface to simplify k8s management. He cautions that SUSE Rancher should be the only application running in a local cluster for optimal performance in managing downstream clusters on-premises and in the cloud, which are focused on the work of the enterprise's operations.
Friesenegger says the tool can ensure clusters are unified under one dashboard for consistency of operations, workload, and enterprise-grade security. SUSE Rancher is CNCF certified and uses standard Kubernetes, which enables it to interact with CNCF certified clusters from anywhere. It can be scaled and built in GitOps for deployment to many downstream clusters.
Additional top-attended sessions included:
- Deep Dive into AI on IBM Z and the Integrated Accelerator for AI
- Mainframe Performance: An Observability Shootout
- SDSF Hidden Treasures and Preview of Upcoming Features!
- MVS Program Opening and Spotlight Speaker
- Watson & Walker's 2023 zRoadshow
- Using the Right Tool for the Job - When to use z/OSMF, Ansible and, Zowe
SHARE Atlanta showed mainframers that the future is bright for the IBM z Systems platform, with k8s, cluster management, and open source innovations still to come. Our events are where our members can spread their wings, learn new skills, and share solutions online and in person. If you missed us in Atlanta, you can still register for the SHARE Atlanta virtual access pass before May 31, 2023. Don’t miss out on all the fun and learning we will have at SHARE New Orleans. Stay tuned for more details regarding registration coming later in April!