So much of today’s computing remains highly dependent on mainframes from the 1970s – especially in government, finance, and insurance sectors. Drafting legislation is no exception. Now, as the skillset for maintaining and supporting mainframes begins to die out, organizations are faced with the critical need to pursue migration options before the technology they are using becomes obsolete. But what are the alternatives to the mainframe? How difficult will it be to execute a successful migration? And what are the implications?
Thanks to advances in technology, the shift towards Cloud-based environments in recent years has opened up windows of possibility that simply didn’t exist four decades ago. In theory, the Cloud offers a solution that need not face the same skill shortage and usability restrictions into the future. But, whether this can hold up in practice remains to be seen.
With the current buzz around infrastructure “virtualization” and “containerization,” you would be forgiven for thinking that these are relatively new concepts. But they were in fact introduced back in the 1970s when IBM undertook highly innovative work enabling applications to run unchanged on different generations of their mainframe hardware and operating system software. Indeed, applications developed back then on IBM System 360 still run unchanged on IBM Z/OS mainframes today. In principle, this containerization/virtualization of mainframe applications bodes well for those looking to transition from mainframe environments to Cloud environments. In practice, however, it isn’t as straightforward as it sounds. The limiting factor is not so much finding ways of transitioning code from physically running on mainframes into the Cloud environment but in finding ways to operate the systems post-transition.
One of the most common patterns in enterprise computing is satisficing the level of automation provided by a system. That is, to get automation to the point where a satisfactory blend is found between full automation and manual interventions. Sometimes, especially where complex workflows are involved, this approach is actually optimal because humans are more adaptable than computerised “rules” when exceptional conditions arise. In complex workflows, exceptions can often be the rule and therefore “human in the loop” is often a great choice when targeting automation.
However, it is the “human in the loop” that causes the issue in simply migrating existing mainframe applications to Cloud environments “as is.” The people who help these applications along on a daily basis – operators, systems people, developers – often possess tremendous amounts of tacit knowledge that keeps the applications running. Personnel with tacit knowledge are notoriously hard to replace, when, for example, they retire, and the technology becomes obsolete without them – as we are seeing with mainframes today.
Measuring your level of automation is key to a successful transition from the mainframe. If your true level of automation is high, then transitioning to the Cloud “as is” may well work for you. If it is low, you may have no alternative but to look at rewriting some or all of your applications into a technology stack you can staff with at least a 40-year horizon.
Most important of all is working with a partner who understands your business processes as well as the technology. In the case of legislatures, this knowledge is hard to find because companies with deep knowledge of both legislative processes and mainframe migrations are few and far between.
At Propylon, we have this deep knowledge. Our LWB 360 product suite is the most comprehensive for legislative environments. In addition, we have accumulated many years of experience in working with legislatures migrating from mainframe environments.