In the late ’70s and early ’80s, every university in the U.S. had a computer science department that trained students in mainframe technologies. The classic mainframe workforce was composed of systems programmers, Database Administrators (DBAs), and application developers who were skilled to administer a z/OS-type system or one of the subsystems that run on it, such as CICS, IMS, DB2, or WebSphere MQ. Application developers busily wrote application code for transactions and batch jobs. Administrators ran provisioning reports and moved data around on the mainframe.
As the distributed systems, UNIX, and Windows revolutions took hold and computing moved from a centralized to a distributed systems model, universities adjusted what they taught. Over the past 10 years, however, it has become increasingly difficult to find classes to train students in the mainframe knowledge and skills covered 20 and 30 years ago.
Combined with retirement and attrition, this training shortage makes it a challenge to backfill positions to support the mainframe. Meanwhile, global business growth and dramatic increases in transaction volumes are leading companies to rely more on the mainframe than previously projected. The mainframe’s reliability, speed, and performance make it a natural platform for handling strategic transactions in leading companies and industries worldwide, and surveys support that its role is likely to grow.
Intelligent Automation to the Rescue
In many industries—such as banking, insurance, manufacturing, and retail—the mainframe group is typically isolated and segregated from the staff members who manage UNIX or Wintel systems. Some progressive organizations are trying to pull those teams together and do cross-training, but most IT departments still have the operating system team for System z, the CICS team, or the IMS/DB2 team. The challenge for management is in cross-training people to manage the enterprise, regardless of the platform, and to effectively leverage automation and advisor technology so IT staff members don’t need a high level of experience in any particular silo.
One way to address this problem is to reduce the number of skills needed to manage the IT environment by automating redundant tasks. This includes automating system startup, system shutdown, and the handling of routine alerts and alarms. Another approach is to make the interfaces simpler and more intuitive. This involves having Graphical User Interfaces (GUIs) that look more like a Windows Internet Explorer interface, which would be more familiar to someone who has grown up in a UNIX or Wintel world. These capabilities make it possible for fewer people to manage the enterprise than were required a decade ago. For example, one telecommunications company uses mainframe systems to track all of its telephone subscribers across a five-state area. That mainframe touches all the revenue produced from those customers. The mainframe staff manages this with about 60 people. Ten years ago, twice as many people were required to track billing, features, toll call data, and other details.
Everyone Needs a Good Advisor
Advisor technology can take the intelligence that has been traditionally supplied by the user of management products and code that information into the products themselves. These solutions take actions that a human being would have taken in the same situation years ago. For example, advisors, also referred to as wizards, suggest courses of action by watching a trend and how it’s changing. The advisor suggests you take a certain action to avoid that problem. At one time, that data would have merely been presented to a systems programmer, and based on his experience and knowledge, that person would have recognized the appropriate action and performed it manually. By codifying the technology with industrial knowledge that has been built into a tool, you can transfer this knowledge and automate it.
IT organizations have experts, but they’re expensive and they get sick, take vacations, leave the organization, and so on. In the past, companies often would have had a system in which staff members were paired up and one person was always on call. By transferring intelligence and automation into the tooling, the need for someone to be constantly available for phone calls is abated. This increases business availability and reduces downtime, the cost of which in a mainframe system is often calculated in hundreds of thousands of dollars per minute or more.
Know Who Is Minding the Store