Enabling continuous development on the mainframe

Why service virtualization is a must-have in your mainframe DevOps toolkit.

It’s no secret that mainframes are an integral and essential part of many datacenters. However, gone are the days when mainframes ran the entire datacenter; the mainframe of today is designed to serve a very different world—and economy.

Mainframes host mission-essential applications that support thousands of applications and devices simultaneously for thousands of users. In fact, mainframes support 70 percent of corporate data, including languages such as Java™ and Linux®, and recently, Docker containers.

However, challenges continue to center around the notion that mainframes are very costly since the MIPS footprint dictates the overall cost of maintaining a mainframe environment. Even though production mainframes have to be sized perfectly – as they have huge MIPS footprints – with capacity to spare to allow for a superior and seamless customer experience, pre(non)-production environments usually do not fall under the same scheme (one with a small MIPS footprint).

Cost reduction is a must to allow for increased profit margins within an organization without impacting brand, innovation, and customer satisfaction.

The push and pull of testing in mainframe environments

In the new world of DevOps, developers and testers alike are forced to share and reserve, in a kind of honor system, the pre-production mainframe environment as needed to allow for the isolated and, more important, appropriate capacity for effective software delivery life cycle  (SDLC) agility in both development as well as testing. And, this reservation process is often flawed.

First, there is no guarantee – or anything to prevent – that once a reservation is in place for one team, that other team(s) are aware of the reservation or will honor the process; they may inadvertently use the reserved environment, thereby impacting any active development and, more important, testing. This can create repeat work and delays in the overall originations of the SDLC to meet deadlines for given deliverables that would otherwise be innovative and competitive.

Second, since preproduction environments are much smaller than their production counterparts, it is highly unlikely that testing can be as effective to meet an organization’s SLA requirements. Testers cannot test “as if” it were production since the capacity is much smaller, which does not allow for proper vetting of IT resources consumption, such as CPU/memory utilization, run-queue, thread exhaustion, garbage collection, and thrashing for the remote/distributed services that depend on the mainframe environment.

Last, all aspects of a given pre-production mainframe environment are not always to scale of a given replicated production mainframe environment, and thus limits usability for mainframe developers and testers alike. And yes, this holds true even across mainframe logical partitions (LPARs).

So you ask, “How can we reduce this cost, and allow for a more agile and effective SDLC to improve on deliverables that meet customer needs as well as overall satisfaction and retention?” The answer is “Service Virtualization!”

Service virtualization to the rescue

An organization can eliminate reservation processes by using copies of virtualized services that are always available. This enables all members of the enterprise’s agile SDLC process to have all the resources needed to be efficiently effective in a timely delivery of business services to meet customer requirements.

The mainframe environment is no longer a limiting factor in the eyes of the distributed developers and testers, since it will seem “as if” it is always available vs being a large roadblock.

In addition to on-the-wire virtualization, one can also virtualize internal mainframe facets in a pre-production mainframe environment and across LPARs to allow for continuous mainframe development and testing of an otherwise unavailable, incomplete, or failing CICS resource.

Mainframe joins the DevOps toolkit

Mainframe virtualization allows enterprises to become more agile and promotes the continuous delivery of business services. With a virtualized enterprise, the mainframe is now just another environment that needs no labeling and is a formidable player in the DevOps continuous delivery toolkit.

Are your mainframe interfaces virtualized? If not, why not?

Omar Agina is a Principal Consultant, Presales for Continuous Delivery at CA Technologies.



Insights from the app driven world
Subscribe Now >
The Sociology of Software >How (Not) to Lie with Data Visualization >DevOps and Cloud Computing: Exploiting the Synergy for Business Advantage >