Startup advances on plan to crack IBM's mainframe stronghold

Lz Labs announces ‘Gotthard’ the first product release of its “software defined mainframe” solution

Lz Labs is keen to stress it is not a typical startup. It is not young, funky and fun with a dog in the office, I’m told. Instead, the average age is about 60 – many people have been shipped out of retirement from across the globe. And with a large undisclosed sum of anonymous backer money it is hell-bent on solving the decades old global issue of legacy mainframe migration.

Founded in 2011, yesterday Lz Labs announced ‘Gotthard’ – the first product release in its “software defined mainframe” solution. At an event hosted at Microsoft’s Zurich offices, it gathered together partner representatives from Red Hat and Microsoft, Dale Vecchio from Gartner, along with potential new clients, to showcase its offering  

Mainframes – which in the world of modern IT systems seem exceptionally archaic – still run 70% of the world’s financial transactions, are used in 5,000 of the world’s largest organisations and one analyst recently suggested still equate to around a fifth of IBM’s pre-tax profits. Yet despite being a post-war technology which miserably limps behind modern notions of business agility they still work and are well-known for their reliability.

Mark Cresswell, CEO, believes the wider problems around mainframe use are coming to a head. The “number one issue is skills”, he says. All the baby boomers who have been operating these machines for years are about to retire and younger people simply aren’t interested. The second factor is that the waning performance and subsequently increasing price of mainframes are likely to make them far less viable over the coming years. And thirdly “there is an existential threat from born in the web companies”.

Yet the problem of migrating from mainframe is not an easy one to solve. There are three ways to approach this, says Vecchio from Gartner, who has been nearly 40 years in the space. The first of these is package migration, the second is to take the old code and translate it into java and the third is lift and shift, as initially pioneered by Micro Focus. All have their issues but what Lz Labs offers is a new approach to third option.

No solution is really simple though, clarifies Vecchio. “It took 40 years to get into this hole – we’re not going to get out in four months.”

What Lz Labs does differently, says Cresswell, is “we’re the only organisation that allows migration to take place without having to mess with the source code”.

“We took the binary representation of these programs and moved them.” This migration covers the base COBOL code and the data itself. While the software can operate on an x86 server running Red Hat Linux or on a Microsoft Azure cloud.

Lz Labs currently only has test phase clients, who remain undisclosed, but it plans to work on a subscription model, says Cresswell. He anticipates the like-for-like costs of running the software defined mainframe will be one tenth of an actual mainframe – although this doesn’t cover any consultancy or migration fees.

During the event, Lz Labs runs a live demo of a migration in action and prospective clients seem very impressed. Yet it is hard to ignore the problems the company may face ahead. The first of these is that although the technology is very smart, a number of attendees say they would still hedge their bets on its ultimate effectiveness. This is because despite years of testing, when transferring millions of lines of COBOL code, it only take one tiny error to render the whole replication wrong.   

The second is that there is a lot of room for this to infringe on IBM, although a number of Lz Labs employees independently volunteer how rock solid their processes have been. The team has put legal issues into the entire framework of the build, I’m told, so there is no room to tread on IBM’s toes.

So, how is IBM likely to react to this? Traditionally “it has depended on how big you are,” says Gartner’s Vecchio. IBM tends to feel quite safe about large clients and be less bothered by small clients – but in the past it has got quite defensive about those in the mid-range. For these, it tends go “three levels above you” straight to the top of the organisation and/or it drops the price.

Thing have changed at IBM though, he adds. “It is hard to find many people who know or care about mainframes anymore.” Now the overarching message is either cloud or Watson for data analytics.

In fact, IBM may well have responded already with the acquisition of an Israeli company at the start of last month. EZSource’s solution is nothing like Lz Labs’ approach but it does aim help developers modernise mainframe applications for digital business.


Also read:

Startup takes aim at IBM’s mainframe gravy train