Why now is the time to reframe the Mainframe

Recently, I’ve made the case for why organisations should move to the cloud and shared a few thoughts on how they can go about this. But if it’s that easy, why isn’t everyone doing it?

In this instalment, I’d like to discuss one of the big migration challenges some organisations need to address – and that’s mainframes. 

Mainframes have been a mainstay of organisations for decades – and for good reason. They’re reliable, able to handle thousands of users simultaneously, and host business critical applications continuously. Departing from such a stable system is a major step for organisations where any downtime is unacceptable – and prolonged disruption unthinkable.

So, it’s no surprise that mainframes are still used by 100 percent of the top ten insurers, 92 percent of the top 100 banks and 92 percent of the top 25 airlines.

However, mainframes also have downsides. These are becoming more pressing as time goes on – so that we’re now reaching an inflection point, where the perceived risks of migration and modernisation are outweighed by the rewards.

Here are five areas where we’re seeing this:

1.   Cost: Although the commercial models for mainframes are evolving, they can still be complex and often lack flexibility, meaning that organisations can end up paying a lot for capacity they don’t use. As niche products, there’s also less scope to shop around and reduce costs. Moving to cloud, on the other hand, enables organisations to optimise their spend through pay-as-you-go pricing and select software and tooling from across the market.

2.   Agility: The pace of application development can be constrained on mainframes. Upgrades can take time because they require a lot of regression testing, and work tends to follow a waterfall rather than a more agile model, like DevOps. Cloud infrastructure is far more agile. Incremental code changes can be made frequently and reliably, making the systems easier to upgrade and maintain. The advantages of this came to light during the pandemic, when organisations found themselves needing to deliver new functions, products and features at speed.

3.   Skills: The workforce that understands mainframes is getting older – and smaller. Most mainframes are built using older coding languages such as COBOL and assembler. Young coders entering the industry aren’t gravitating towards these older languages and platforms – they’re bringing their energies and talents to more modern technologies which are easier to build on, such as cloud. 

4.   Data: Mainframe data is often locked away in proprietary databases and management systems, making it very hard to access or analyse across the business. Organisations looking to embrace data and insights more broadly find it far easier to do so with cloud: it’s far easier to securely move data from A to B, and there’s a broader range of tools and techniques that organisations can bring to bear.   

5.   Documentation: As many mainframe systems take shape incrementally over decades, a lot of what is on the system is poorly documented or just resides in people’s heads. This takes us back to the issue of skills and knowhow. Mainframe engineers are beginning to retire from the workforce, taking with them valuable and sadly irreplaceable knowledge.

It’s clear to see why so many organisations are considering the future of their mainframe applications. And while the challenges organisations face are common, the mainframe modernisation and migration journey each takes will be unique.

One option is to work with partners such as AWS and Micro Focus to re-platform the mainframe. In this scenario, the entire mainframe is migrated onto cloud. This typically reduces TCO by 40%, which can then be invested into a longer mainframe modernisation and transformation journey.

The other option is to look at mainframe workloads systematically, and work out what workloads can be decommissioned or rebuilt on new platforms. This is a far more gradual transformation, but can be a better option for organisations that still have life in their mainframe assets.

Traditionally, the second approach has been more prevalent – but this is changing as the Cloud providers and technology partners invest heavily in this space. Thanks to this, organisations are now achieving – and even exceeding – mainframe-level performance in the cloud.

Fundamentally, there is no one path to mainframe migration, but there is a general direction of travel. For many organisations, it’s about “shrinking the core” – peeling off data and functionality bit by bit. It may be that a mainframe remains the best place for some operations, but they are the exception rather than the rule. 

If you’d like to discuss the challenges and opportunities associated with migrating your mainframe applications to the cloud, please do get in touch.