Skip to main content

The Pandemic and Modernizing Your Mainframe

The press often seems filled with stories about how mainframes were first available in the 1960s and implies that the technology hasn’t advanced since then—ignoring the fact that cars and jet planes are even older, and they can be seen to be updated all the time. So, why wouldn’t mainframes have been updated to give users a competitive edge? Having said that, I do sometimes get a feeling of dread when I see anything referring to mainframe modernization. The articles so often simply explain why organizations should throw away their mainframes and start again on more “modern” platforms. So, it was with mixed feelings that I read the 2021 Mainframe Modernization Business Barometer Report from Advanced.

The report examines the current mainframe market and the challenges facing enterprises worldwide with annual revenues of more than $1 billion. It also explores the impact COVID-19 is having, and will continue to have, on modernization plans.

Getting a baseline first, the report found that the average enterprise consumes 33,286 MIPS every month. And organizations are spending an average of $4,266 per MIPS annually, factoring in the cost of maintenance, software licensing, hardware, operational and modernization costs. This, the report suggests, equates to an average of $142 million a year.

In addition, 78% of enterprises have started at least one modernization program as a direct result of the pandemic, while 60% say COVID-19 has accelerated their cloud transition process. On average, they’re spending $48 million on modernizing the most urgent elements of their legacy systems. Other findings are that more than half (55%) say legacy modernization has helped the business accelerate their digital transformation efforts, while 54% say it has allowed them to be more reactive to market changes. In addition, the report found that 87% of organizations have scheduled at least one legacy system modernization program in the next 12-24 months.

Legacy Modernization Programs

The Advanced 2020 survey found that 64% of organizations use mainframe-based applications that are between 10 and 20 years old. And 28% are between 20 and 30 years old. This, you might conclude just shows the longevity of mainframe applications—although that wasn’t Advanced’s conclusion. This year’s 2021 report found that mainframe-based applications are comprised of an average of 8.86 million lines of code, written in multiple languages, across a single environment.

The report also found that 36% of organizations consider the legacy modernization programs they have completed, to be failures. Similarly, 77% have started but failed to complete at least one modernization program. The figure is 84% for government organizations. And 87% of UK-based organizations have started and failed to complete at least one modernization program, compared to 69% of companies in the rest of Europe and 59% in the USA.

Why did so many modernization programs fail? 33% of respondents pointed toward a lack of the right skills and 28% attributed their failures to poor information about legacy systems in the business. That latter figure rises to 41% for government organizations. 26% said that they have too many time-consuming modernization programs already underway. The survey also found that 59% of respondents failed to get funding from leadership because of past modernization program failures. 42% said it was because there is a lack of understanding of why the cloud offers material advantages over alternatives. The survey also found that 59% of CIOs and heads of IT have confidence in their organization’s digital transformation strategy, a guiding light for modernization initiatives.

Programming Languages and IT Talent

When it comes to the programming language in use, perhaps not surprisingly 75% of the enterprises surveyed said COBOL remains the most prominent language in their mainframe estate. Assembler was found at 66%. Natural was at 41% and ADS/Online at 40%. The survey also asked what data/database structures mainframes use. They found that 76% use Adabas, 65% use Db2, 58% use IDMS, 28% use IMS and 23% use VSAM.

89% of respondents said that they are concerned about having access to the right IT talent to properly maintain and extend their legacy systems. 36% are concerned that people entering the workforce only have “modern” skills (i.e. Java, Python, C# and cloud-native deployment and orchestration tools such as Docker and Kubernetes), while 33% are worried staff are retiring and taking their legacy skills with them. The survey also found that 29% of respondents say their staff don’t want to learn legacy skills, while 28% fear they may lose talent to competitors using more “modern” technologies (i.e. DevOps and cloud ecosystems). In addition, the survey found that 78% of respondents are already targeting—or planning to target—Java when modernizing their mainframe.

38% of respondents said that they prefer a refactoring approach (i.e. transforming legacy applications and databases into equivalent modern ecosystems using automated tooling). 36% said that refactoring will help them reduce the legacy skills gap. 46% of organizations also say that refactoring would allow them to make use of the latest frameworks/architecture, which would support a move to any modern environment.

The survey also looked at people’s attitude to green issues (sustainability, energy efficiency, reducing carbon footprints and minimizing the amount of power required to fuel technologies used).

For those of you who like this kind of detail, the report was carried out online by Coleman Parks throughout April and May 2021. The sample comprised 400 people working for large enterprises in Europe and the USA with a minimum annual turnover of $1 billion. The respondents included enterprise architects, CIOs, heads of IT, application managers and program/project managers in a range of industries.

Advanced clearly have their own agenda, but survey results, in themselves, are always interesting. “Modernizing” and “modernization” are always such difficult words to define. It’s not just about migrating to the cloud. I guess most sites are modernizing all the time by using the latest releases of software. And many have modernized by joining the API economy and making microservices from their mainframe applications, using JSON and REST, available to apps from other platforms, and then available to staff and customers.

While using the cloud clearly offers great business advantages to mainframe-using sites, it’s not the only way of modernizing. As we know, using Zowe allows “modern” programmers to access the mainframe. Still, like I say, survey results are always interesting.