Mainframe: Our Heritage or Present?
Corporate Communications Director
On May 11, we launched our first video blog. We selected mainframe computer systems for the video discussion.
Why mainframe? Again mainframe?
IBA Group has a long history of mainframe software development. This history begins even before the company was officially registered. Roughly 30 percent of our team members develop and support mainframe applications. Why is this subject suddenly so hot today and being discussed everywhere from Bloomberg to CNN?
I invited Andrey Lepeyev, Director of the Mainframe Department at IBA Group, Mark Hillary, our blogger and an author of 17 books on technology and customer experience, and Peter Ryan, a prominent analyst in customer experience management and BPO, to answer these questions.
Mark Hillary: “Big companies are still using these enormously powerful devices.”
Mark Hillary, our blogger and an author of 17 books on technology and customer experience, said:
When I studied software engineering in the late eighties and early nineties, I had to study the difference between mainframe systems and the IBM PCs that were taking over the world at that time. We did use a Prime mainframe, but even 30 years ago, many of us felt that the mainframe was a dinosaur because computing power was coming to every desk.
You might think that this is even more true today, because we all carry around more computing power in our phones than NASA had back when they put a man on the moon. But across the world, mainframes are still out there processing data for airlines, government agencies and highly regulated industries, such as banks, where rapid change is difficult.
IBM released their new Z15 just 7 months ago and it has the power to process a trillion web transactions a day. You might think that large-scale computing power has all migrated to the cloud, but some companies are still using these enormously powerful devices. For many organizations, it is just too difficult to migrate away from their mainframe.
The problem is that computer science students today don’t bother studying mainframes. Mainframes are just something from the history books. Students want to build mobile apps, games, or code for computers that are widely used. I tried using Google earlier today to find a computer science course that included mainframes and I couldn’t find any – if they do exist then they are well hidden.
This means that there is a scarcity of resource for maintenance and improvement. That’s what we are going to talk about today because there are still some areas of the world where mainframe expertise can be found.
Peter Ryan: “Reliance on mainframes has been highlighted by the pandemic. These systems need to be managed more effectively.”
A related problem is that much of the business software developed to help airlines manage ticket sales or banks process transactions was written in COBOL – which means Common Business Oriented Language. It’s a computer language that rose in popularity along with mainframes – and then declined as the mainframes also declined.
So we are in a situation where it is hard to find people who can maintain the machines, but it is also hard to find people who can change the software too. Most software developers with COBOL experience have retired – it is no longer commonly studied in university.
The Covid-19 coronavirus pandemic has vividly demonstrated why we need mainframe and COBOL expertise – many government systems cannot be easily changed without this knowledge and expertise. The $2.2 trillion CARES Act passed in the USA in March included a $600 week increase in unemployment benefits for American citizens, but until the individual states can update their systems to reflect the new law nobody is getting that extra cash.
In Oklahoma it is taking over 2 weeks to process unemployment claims. Newly unemployed citizens have a right to these extra benefits during the crisis and yet because the mainframe systems are so slow and cumbersome – and COBOL programmers are so scarce – people are not being helped.
The last time Gartner counted COBOL experts in the USA was 2004. They counted 2 million back then, but also noted that each year 5% of those experts retire. My math may not be great, but that means they are now almost all retired.
There are mainframes out there delivering critical infrastructure and they are using billions of lines of COBOL code that very few people now understand. The problem is made worse because anyone who studies COBOL today would find themselves focusing on maintenance work – it’s not as inspiring as launching a new app.
This reliance on mainframes and COBOL has been highlighted by the pandemic. These systems need to be managed more effectively.
Andrei Lepeyev: “Mainframe remains unbeaten in terms of security and input / output operation, which is important for transactional applications.”
Currently, we have more than 400 mainframe experts at IBA Group. Initially, we worked on classical tasks. Today, we work on mainframe modernization.
Mainframe remains unbeaten in terms of security and input / output operation, which is important for transactional applications. However, the mainframe platform has been changing and the service model has been changing too.
For example, the IBA’s APPULSE solution monitors operability of business applications that run on System Z and with the help of machine learning offers efficient solutions for problem situations. In addition, the solution implements experience management, which is very important because experts in this field are on high demand.
We are also very active in DevOps for mainframe. It is of great importance because flexibility in the change of functionality is often the most vulnerable part in mainframe applications.
We have been traditionally involved in projects on modernization of the architecture of mainframe applications. Initially, we developed applications from scratch and recently expanded to add open source solutions, such as Zowe. This includes modernization towards GUI (Graphic User Interface) and building of a flexible SOA (Service – Oriented Architecture).
This is our first blog post in a series of video discussions on mainframe systems. Please share your thoughts about the discussion or offer your topics for future videos by leaving your comments or suggestions here.