How to Move Mainframes into the Modern World

Mainframes are often treated with an extreme deference that makes their necessary modernisation hard to contemplate, let alone carry out. Ingenious methods used to preserve the legacy of physical institutions while releasing their innovative potential provide inspiration when it comes to the mainframe’s “inner core”.

Let’s use an analogy: In Zurich, a factory that pioneered alternating current for the locomotion industry was moved 60 metres in 2012 to continue its modern mission as a value-added hub for engineering inspiration. The notion of translocation — the movement of artefact, building or even plant — to a more favourable location turns out to be highly apt when it comes to mainframe modernisation. A bold feat of engineering moved the 133-year-old Swiss factory in order to preserve the heritage of ideation that has now been recharged for modern times. In the same way, translocation preserves the mainframe legacy while making it fit for the future. Keeping business value intact – mainframe business logic and applications — while moving it to a future-proofed platform, lowers risk and cost. Crucially, it opens up value generation, too.

Seems impossible? It’s been done successfully many times

Clearly, mainframes — monolithic software structures comprising ancient languages, routines and databases that prop up modern businesses — are not physical buildings. But an engineering mindset that’s open to novel and proven solutions when faced with a seemingly impossible software migration is relevant — and is needed ever more urgently. Programmers who wrote and maintained routines and appended new functionality are busily retiring. And mainframes in their current condition are too unwieldy to spin up agile responses to frequently occurring macroeconomic or climate events that shake the world.

It’s understandable that technology and business chiefs have been reluctant to dismantle and rebuild mainframes. Unfailingly, they host the software that supports an organisation’s bread and butter operations, securely transacting billions of dollars every day. The lack of visible successful modernization examples perpetuates the notion of the untouchable mainframe – often a ‘Big Bang’ rebuild is turned to as a final resort, sometimes with catastrophic consequences.

LzLabs has similarly developed a tried, trusted and low-risk method for relocation – in this instance moving mainframe applications and  business logic to modern Open Source environments. We leave the inner core intact – so we don’t rewrite or recompile or refactor. Instead, using our Software Defined Mainframe, we decompose the inner core into discrete pieces, which we move to the cloud where they continue to process business critical via the SDM. One of the key innovations from LzLabs is that the behaviour of the subsystems is unchanged.

Realize Return on Investment “on the Way”

The beauty of the LzLabs approach is that an ‘intact’ incremental modernisation releases economic benefit rapidly at each micro migration. Confidence and engagement is boosted and the steady flow of funding pays for the ongoing transformation and demise of the legacy platform. Taking a factory-based approach — using the right team, right model, and right governance – incremental modernisation can be a quiet thing running in the background.

Working quietly under the hood and replatforming brings a range of benefits, removing project anxiety and risk — while users don’t even have to know. Unlike a rebuild or refactor approach, functionality doesn’t have to be tested and there’s no question about the behaviour of replacement applications and whether this is palatable to users. The icing on the cake is a potential saving of up to 70 per cent on the legacy platform.

It’s a pragmatic and highly attractive alternative to a ground-up rebuild where the inner core is rewritten from scratch, which increases the risks because of multiple associated dependencies. Just as harmful, rebuilding postpones the realisation of any benefit until the end of a very long project — and sometimes fails to realise any benefit at all. In regulated industries, huge penalties may be levied if customers are negatively impacted by downtime or flawed applications.

Nonetheless, the fear of tackling the inner core has led to the above scenarios — and worse — as technically illiterate executives may commit to project timelines that are untenable and risk the business. Fear is understandable: the inner core touches 90 per cent of the critical processes within an organisation. Typically, no one has looked at it, touched it or recompiled it in 50 years. Worst of all it’s largely undocumented.

Conventional wisdom says that you need to unpick this kernel. But because the task is so daunting,  IT teams tend to nibble around the edges, focussing on fringe use cases in the hope they can make some marginal progress. At LzLabs, we advocate for leaving the inner core intact. This lets IT teams tackle the core head on in a way that allows them to move functionality unchanged, which continues to transact work flawlessly, back to the host. 

Step-by-step mainframe migration, “on rails”

The LzLabs approach allows you to have a phased, low-risk migration, and the comfort of a secure operating model during the transition stage. And, all the while you can leverage existing distributed assets and resources. So current monitoring systems, scheduling capabilities and print and output management continue, business as usual, via the SDM. And in terms of future-proofing, in the cloud, an array of tools awaits that can be deployed to maximise business logic.

Mainframe migration may not be as mechanically straightforward as moving a building on wheels to a better place. But the Software Defined Mainframe lets software engineers deconstruct the mainframe edifice, piece by piece, and relocate functionality intact to a new environment in the cloud. Here it preserves precious business logic, which can be enhanced by an array of software tools and stay competitive in a rapidly changing business world.

Related articles.