Listening to Bob Dies it’s easy to believe that network computers have already arrived in the world of prime time computing. The days of having to keep a full copy of every piece of software on every desk are over, he enthuses. A lot of the things that you had to put up with when using PCs you don’t have to put up with anymore. Such sentiments are, however, unsurprising coming from the lips of IBM’s Network Computer (NC) Division chief. And when it comes to hard facts, Dies is less vociferous. Although he claims that a poll of users by the Gartner Group showed that IBM is the clear leader in network computer shipments today, he is loath to discuss specifics. How many Network Stations (IBM’s network computer) have been sold? Tens of thousands have been shipped, he says. Who is using them? In Europe alone we have 1,000 accounts which have installed some number of NCs How much is ‘some number’? More than two, he replies. A 28-year IBM veteran, Dies came to the newly-formed Network Computer Division covered in glory. Under his direction IBM’s AS/400 proprietary midrange computer, which had been struggling against ‘open’ Unix alternatives, was nursed through a difficult redevelopment process that involved rewriting the operating system and changing almost every component in the machine. By the time he left last year, the system was experiencing a significant resurgence in sales and popularity. Now IBM wants Dies to work that same magic on the NC business. But whereas the AS/400 was part of the company’s core server business, benefiting from hefty legacy momentum, the Network Station is having to break new ground in the fiercely competitive end-user marketplace. Not only that, the NC concept is now facing stiff opposition from the NetPC, the cut-down, low-cost version of a standard PC developed by the incumbent industry leaders, Microsoft and Intel, in a bid to kill-off the NC market before it gets started. As a PC supplier itself, IBM was originally supporting both the NC and the NetPC concepts, but now the company has chosen to focus on the NC, targeting general access applications, the internet and Java. The accompanying sales message for these systems is ‘less is more’. When I took the Pentium on my desktop apart I found 11 megabytes of my own, unique data and about half a gigabyte of application software, says Dies. Are you going to have that half a gigabyte on every desktop for every employee, or are you going to put one copy of that on a server and download it as needed? Switching to NCs, says Dies, also brings productivity benefits. Lotus Notes runs faster through my Network Station than on my PC, but it physically runs on a NT Server down the hall. To prove the point, the entire NC_Division has now moved over to Network Stations. You don’t need a Mac truck on your desk to drive an application, says Dies. Market research firms, however, have been encountering mixed messages from users. In a recent survey of 137 senior-level IT buyers and decision-makers, US Internet analysts Zona Research discovered that only 15% were planning to deploy thin-client architectures within the next three years. Of those who were not, 44% said this was due to thin-clients not being PCs, while another 44% said it was a bandwidth and network issue. We expect to see thin-clients deployed in niche markets such as terminal replacement, where enterprises are extending the use of legacy data and applications, says Greg Blatnik, of Zona Research, but, he says, the volume of that market will not be enormous. Contradicting that view, a survey released this month by Bloor Research found that more than 50% of large corporations are examining NC technology, while 18% are already implementing it and 17% are involved in pilot schemes. People have bought the story that a return to centralization gives you lower costs, a better quality of service for the end user, and better control, says Andrew Hammett, director of Bloor Research, In fact, Bloor is predicting that, within five years, the commercial NC market will outnumber PCs in the corporation by two to one. The Bloor study coincides with Dies own prediction that 40% of large accounts will have NCs installed by 1999. The world is moving to NCs, he says. Lots of large companies are either putting NC programs in place and testing the technology or are looking at it in considerable detail. If he is right, he could be the glory boy once more. But if he is wrong, IBM could find itself a sorry loser on the wrong side of Microsoft in the desktop market once more.
This article originally appeared in Computer Business Review