It’s a regular occurrence in the tech sector that the call will go up from time to time that this - delete as applicable - is `the year of the mainframe’ - again. To date, that’s never quite come about in the ‘back from the supposed dead’ way that’s intended by such hyperbole.
That’s largely because ‘wasn’t really dead to begin with’ is the user reality. While the technology had improved in performance and condensed in size, the primary role for the machines had not changed at all. They remain, for those businesses that really need the function, the ‘dreaming spires’ of the back office, running the vital applications of record, the bi-modal model of computing espoused by the likes of Gartner.
That said, that model now looks pretty much redundant, as there are real signs of significant change as developments in the way cloud services are used come together with changes in the way mainframe applications are developed and maintained. And the result is not just a potential role change for the mainframe to change.
Couple that with the fact that two of the leading software and services providers in the mainframe world, Compuware and BMC, have seen the case for coming much closer together in order to take advantage of the change and exploit the hell out of it. The acquisition of Compuware by BMC, announced in March, could easily be read as a sensible consolidation of two players in a gently declining marketplace. But given that they are partners of long standing who have been working well together and ‘living in sin’, why bother ‘getting married’?
Two important words
Two words might explain that last question - ‘cloud’, and ‘edge’, both drivers of change in terms of the way that applications need to work with back office systems. Together they are bringing fundamental changes to bother infrastructure and operational procedures. As compute moves out to the edge to process data where it originates, so the relationship with the back office changes and applications of record need to become far more responsive and collaborative.
Creating the opportunity for mainframe users to get more out of their existing mainframe investments was why Compuware started down the road of bringing mainframe applications development and maintenance into the world of DevOps some five years ago, under the tutelage of then new CEO, Chris O’Malley:
Five years ago we felt that what needed to happen was that the silo of the manager needed to go away. That we needed to bring agile and DevOps to the mainframe platform and that a lot of inventing of tools needed to be done. To make that happen, our dream was that the mainframe will be different only in syntax from any other platform which meant, obviously, that we had to build a bunch of tools.
Over the past five years, new tools have kept coming and the applications have started to spread. By enabling a modern DevOps tool chain and coupling it with the platform’s virtues of reliability and performance, security, and transactional efficiency, users can start to exploit its capability, says O’Malley, accelerated by external factors:
It is the cheapest platform on the planet for the high volumes of transactions, which is a huge part of the digital age. And now the pandemic is causing us to do everything online, so that no cash changes hands. They want you to buy your dinner online before you show up. So that's driving this bursting effect. These are the transactions the mainframe is purpose-built to handle.
Two old chums, one targeted objective
It is O’Malley who is now leading Compuware through the acquisition with BMC, with the former’s capabilities complementary to the latter’s DB2 and IMS tools. He sees changes happening to bring IMS back to life in a big way as all the banks use it at their core. The downside of that, he observed, is that nobody had wanted to touch IMS until the company introduced a virtualized implementation that reduced the number of physical IMS systems required to support applications development.
In addition, customers want data for DBAs and the operations teams and so the coming together of the third and fourth largest vendors in the mainframe software sector seemed a natural step with no overlap in products and capabilities between the two. Compuware covers development and application performance and BMC covers infrastructure performance and security.
In O’Malley’s aspirational view the combination now puts them in a very strong position to catch the cloud/edge computing wave ahead of the current market leaders:
The two top vendors are IBM and Broadcom, neither of which is doing anything near what customers need and demand at this moment, which is to solve these incredibly urgent problems. In its way we've landed at the point where users need a big, bold, aggressive innovator. We've now left Broadcom and IBM in our wake, so far behind so we can't see them anymore, but that's what the market needs. They need an aggressive player to make this mainframe re-invented for the future.
Compuware, of course, has already adopted a DevOps approach to its own development of applications and tools. O’Malley ordered a major rip-and-replace exercise where some 20 tons of hardware was ripped out and skipped, to be replaced by cloud services. Some 80% of the company’s old data center is now empty and the rest is where 15 ‘skinny’ mainframes and their support infrastructure are housed.
That has allowed the firm to benefit from ‘eating its own dog food’. Instead of the classic mainframe roadblock of new and enhanced tools and applications only being available on a three-yearly update cycle, there are now four or five updates a year.
O’Malley sees the mainframe’s traditional role as the core back office system growing into a [starring role’ at the heart of virtualized, distributed data service, providing the necessary processing working out wherever data is generated in significant volume. And as most companies are now moving towards growing a high volume of different logical and/or physical locations where high volumes of data are generated and processed, the role of the back office is likely to grow exponentially.
That, in turn, is going to require the fast, most reliable and most secure systems possible for transaction processing and management, which in O’Malley’s view means only one possible answer – the mainframe. One interesting step he did note is that, while the core market for mainframe systems remains the banking and financial services communities, the growing pressure of the cloud and edge compute requirements, which are considerable in areas such as retail banking, are pushing users to start using the mainframe with new workloads.
What goes around comes around they say, and the mainframe has been going around for many decades now. And while its return to glory has been predicted a few times, it has never quite happened, but maybe it might just this time. The acquisition of Compuware by BMC creates a new business, well-placed to provide a key component of that new infrastructure balance where the essential back office that can work directly with cloud-native consumer-facing applications.