Mainframe computer at a crossroads as technology landscape changes rapidly

Last week I spent some time in New York, at the invitation of technology company IBM, for an important anniversary.

In early 1964, Thomas Watson Jr, the son of the company’s founder, essentially bet the company on a new technology, the mainframe.

This was a new era of computing, in which heavy-duty computing muscle was put in the service of some of the most pressing problems, including the space race, banking and management of large sets of data.

At the New York anniversary last week, though, was an interesting sight. The event was held on Park Avenue in midtown Manhattan. When the main festivities were done however, and we got out, there was an advertising van parked across the road.

It was from a competing company, and it bore the legend: After 50 Years of the Mainframe it is Time to Embrace the New Style of IT. Guerrilla marketing, to be sure, but in that tableau lay the very question of the future of information technology.

In an era of telephones that can take pictures, and computers you can hold in one hand and do not require a physical keyboard, it is sometimes difficult to remember just how far we have come in a mere 15 years.

At the turn of the century, it was a wonder when one sent a text from a mobile phone.

Laptops were only then coming into vogue in a widespread manner, and companies in which every employee had access to a desktop computer were rare, at least in Kenya.

A mere five years before that, it was easy to divide computing by size — mainframe, minicomputer and microcomputer (which included desktops). The different computers had unique tasks, delineated conveniently by size — which corresponded to computing power.

But now, in the era of the internet, distributed computing and the cloud, that delineation is breaking down in unpredictable ways.

Because current personal computers are so powerful — the oft-cited factoid is that there’s more computing power in a modern smartphone than was available to the Apollo scientists who crafted the moon landing 45 years ago; the implication is that any computing power needed can be found in these personal devices.

That’s largely true for most personal computing needs, but it is when the demands are greater, and needs not so linear, that the argument returns.

IBM’s bet is that the era of the mainframe is in its infancy. As countries move into offering e-services to their citizens (and Kenya has made bold claims in this direction), vast resources are needed to crunch data on anything from traffic to disease control to climate modelling.

The needs are equally great in the private realm. Banks, for instance, are offering their customers the opportunity to access the full menu of banking from their mobile devices.

At the back end of this is usually a mainframe computer, whose remarkable computing power does everything from calculate interest rates and customer balances in real time, to taking care of security and peak usage.

Can these requirements be met by cloud computing and other ways of harnessing computers together? The jury is out on that one. On the one hand, there are those who say that the ever-increasing abilities of personal devices are just the ticket.

There are others, though, who point to the presence of mainframe computers even in cloud-based services, with these behemoths doing the heavy lifting at what computer experts like calling the back end.

I had a chat with a doctor from Ghana named Elijah Paintsil. He’s currently a professor at Yale University, and is running a remarkable project back home which aims at eliminating mother to child transmission of HIV, from the current 15 per cent to near-zero by 2020.

Tracking women

He told me that one of the biggest hindrances is the lack of ability to correlate data about women who visit health centres for pre-natal care and the ability to test them for HIV.

Often, the test kits will not be available where needed, and, when the expectant woman leaves the clinic, off she goes with her unknown status.

The project aims at tracking these women all the way down to their villages, where other data networks (such as chiefs who know everyone) can be cross-referenced to these visits.

It’s an ambitious, presidentially-decreed project, and Dr Paintsil claims that it can only be delivered using the power of the mainframe.

It’s another company-betting moment at IBM, which, if you remember, sold its personal computer business to China’s Lenovo in 2005, and also began to divest itself of its server business early this year.

It’s now mostly in services, and those of which are ambitious enough — like the Smarter Cities project — to require the computing power of mainframes. Will it succeed? Well, we’ll find out over the next 15 years.

Mr Kantai is the NTV Business Editor.