Lloyd’s Register (LR), the global shipping classification organisation is using its IBM mainframe to support a speedy mobile app used by thousands of surveyors after turning to Oracle’s Mobile Application Framework (MAF).

The organisation is owned by Lloyd’s Register Foundation, a charity, but runs as a business. Its marine arm relies on ship fleet managers booking surveys for LR to inspect and assess their docked ships in order to make its money.

But until February this year, LR was relying on a website created at the dawn of the internet, running off an IBM mainframe.

It was used for customers to make survey bookings and check existing boat data, and for internal surveyors to monitor boat surveys already made.

LR database administrator Richard Childe said surveyors also relied on out of date laptops to access the site, while docks usually lacked internet access.

As a result, LR came to him with a demand to create a mobile app that would do the job instead.

However, the app needed to utilise the IBM mainframe that contained all the shipping survey data – a real sticking point, said Childe.

Childe, who called LR a "big Oracle shop", went digging into its existing Oracle deployments and found the Oracle Mobile Application Framework (MAF), which was updated just this year to allow customers to create single-source mobile apps for Android and iOS.

He said this ability to make the app suitable for a variety of devices and operating systems was key for LR’s needs.

"When you move into mobile app development, it’s going to clients’ mobile devices and you’ve got not control over what they are using, you don’t know what software they are using on there which could interfere with your app, you don’t know the screen sizes," he said.

But the bigger problem of the IBM mainframe remained. Ordinarily, you want to boost app performance by reducing the number of network calls it makes, usually by caching data on the app.

But Childe explained that was simply not possible with the nature and amount of data contained on the mainframe.

"Every single user of the app is basically looking at individual data, which is volatile and constantly changing," he said. "It’s huge volumes of data. So you’ve got no choice, you can’t copy that entire database onto your phone. It’s just not practical. You have to run queries as and when people need them – it’s on demand, effectively."

This was a problem because the mobile app would become the major way Lloyd’s Register could make money via surveys, because it was designed to replace the website as the central way customers could interact with LR.

Childe faced the issue of speeding up an app that should by all rights be running slowly.

"If you click something on a mobile app and you don’t get a response in two seconds you delete it off your phone," he said. "If you get people doing that, it creates such a bad name for the company."

He set an aim of having data returned via the app in less than one second. After two months of development, Oracle MAF’s boosted the mainframe’s performance to make this possible.

Childe explained it was both MAF’s underlying server keeping up with the speed of his data demands and the app being fast that allowed him to succeed: "I had to use my background in Oracle performance but also I’m dependent on the framework to be able to cope with that. So that’s the WebLogic Application Server being able to cope with that and the mobile app being able to consume it quickly."

While he lacks solid statistics, he claims the app is now in constant use since its release in February.

And he has plans for how it could be developed in future – being able to do everything from sending pictures of repair jobs to GPS tracking.

"Why can’t they have an app that photographs a component that’s looking dodgy and send it back to the database?" Childe asked.

"Why don’t we use GPS to track the vessels’ locations? There’s potential there that’s enormous. It’s how far we want to go with it really."

Read our exclusive E-zine and discover more about System Z…the platform for the modern world here.