The mainframe is the birthplace of virtualization and where it still works best, so it’s no surprise that industrially virtualized Linux took to the mainframe and z/VM like a penguin to water. IBM’s brand new Enterprise Linux Server line of mainframes confirms this.
Yet, it has taken the average organization a decade to begin taking advantage of production mainframe Linux. A recent survey about mainframe Linux (http://ca.com/mainframe/linuxre-search) found that:
• The Integrated Facility for Linux (IFL) specialty mainframe processor, designed to enable Linux to run on the mainframe at a lower cost and with no impact on the cost of traditional workloads, is an important and growing aspect of mainframe data center environments.
• Linux on the mainframe is seen as more cost-effective and “green” than non-mainframe alternatives.
• Provisioning, backup and disaster recovery, the ability to scale virtual machines, security, and availability of applications were all seen as important challenges, but they also all were identified as areas having significant advantages within the mainframe Linux environment (vs. non-mainframe platforms).
Clearly, the future of Linux on System z is bright and getting brighter, as illustrated by the recent announcement of IBM’s first ever mainframe system designed just for Linux. But with such opportunity comes the need for establishing generally accepted approaches to configuring, running and using this environment, along with further innovation and other advances.
The primary virtue of Linux on the mainframe is that it’s cost-effective for production requirements. There’s only one code base and you need only one physical machine to run a nearly unlimited number of concurrent Linux images, saving substantial amounts of time, space, and staff. The advantages don’t stop there; organizations moving to Enterprise Resource Planning (ERP), databases, and Web-based applications also are reporting software licensing savings. Solutions that are licensed by physical machines enable the most licensing savings because of the extreme virtualization possible today. There’s also the benefit of matching the peaks and valleys of numerous concurrent Linux images with an economy of scale that smaller boxes could never achieve.
The first area of managing Linux on System z is z/VM, its virtualization environment using established, proven quality solutions for securing, automating, and provisioning this environment. This brings an immediate advantage, ensuring that each individual instance of Linux can be trusted to run in a reliably secure, available context.
With the assurance of this production-quality foundation, the next step is enabling dynamic creation, con- figuration, modification and even removal of Linux images in a demand-based manner, reflecting the constantly changing needs of the organizations where it runs. While this has often been done as a combination of z/VM-based functionality and local customizations, the opportunity exists to move to a simpler, more dynamic provisioning system.
In addition to excellent security—both within Linux and for Web-based access to applications that are served up by Linux—it’s important to be able to manage the performance of applications that have a mainframe Linux component, to ensure their ongoing availability and acceptable response times, to alert when these are negatively affected, and to drill-down to find the cause when this occurs.
Your system also must have access to data for processing and distribution purposes, as enabled by application development and reporting environments and production-quality data transport solutions.
A final essential component is workload automation, using a single, graphically based solution that can dynamically manage your whole enterprise, including your mainframe Linux images.
A well-managed mainframe Linux environment holds the keys to the future of virtualization, cost control, manageability, environmental (“green”) responsibility, and dynamic adaptability. It looks like this penguin is here to stay.