ICL & UNIX - A farewell to arms

In the battle to turn back the tide of IBM's influence, Unix was ICL's primary weapon. But its finally abandoned the OS under pressure from Windows NT.

After nearly 25 years, ICL has thrown in the towel and admitted theL's primary weapon. But its finally abandoned the OS under pressure from Windows NT. Unix operating system has been a failure. From now on, the vendor will concentrate all its operating development efforts on Windows NT, although users running Unix or its mainframe operating system, VME, will continue to be supported.

ICL was one of the first companies to adopt and promote Unix as an international standard and counterpoint to the proprietary environments available 25 years ago, but its hand has been forced. Andrew Boswell, technical director at ICL, says: 'Unix has been a great disappointment to us. The Unix camp hasn't really got its act together. There are still too many versions of Unix. Basically, we've said if our customers wish to use Unix, we'll continue to support them, but future systems will be built around the NT platform.'

The ICL/Microsoft agreement to develop products for a public information network via internet-based information kiosks, has prompted the decision to abandon 25 years of technical development and marketing push for Unix.

But reading between the lines, there are a number of factors which make the deal beneficial to both companies, but not necessarily to their customers or partners.

ICL, once the UK's only mainframe manufacturer, was set up in the 60s and for many years was virtually the only company to supply systems to central and local government and the public utilities. This policy of preferential procurement was ended by the Conservative government in the 80s after more than a decade of protestations from US suppliers such as IBM and Bull (then Honeywell).

But for 10 years and more, ICL had a clear run at the public sector and still has the lion's share of the installed base. It is this that's attracting Microsoft, particularly since it involves internet technology, one of the issues where it's clashing with the US government.

The second advantage to the software giant is the astonishing statement by ICL that it will discontinue development work on Unix and concentrate on NT as its core operating system. Effectively, this means that if the public sector continues to replace ICL kit with ICL kit, they'll be forced down the NT road, like it or not.

As a means of laying the foundations for capturing a significant slice of the UK market without having to do any hard selling, marketing or technical proving of a product, it can scarcely be beaten and Microsoft must be overjoyed.

The vendor has been very clever in its choice of UK partner, which should give it an opening to the lucrative public sector contracts without committing itself to too much. There's nothing exclusive about the ICL deal - Microsoft has similar arrangements with other companies, including Hewlett Packard and Digital, both competitors to ICL in the public and private sectors.

It also has ICL, one of Europe's leading suppliers, endorsing Windows NT and committing itself to it at a time when doubts are being raised about its scalability and robustness as an enterprise wide operating system.

Not that the deal is all one-sided. ICL can claim to be Microsoft's most important European partner in the year when it finally announced plans to float on the stock exchange in the year 2000. This should stand it in good stead with the City which, the US DoJ rulings on abuse of dominant market position notwithstanding, regards anything touched by Microsoft's magic wand as the rainbow leading to a huge pot of gold.

Then there's the business the agreement will bring ICL, which ICL estimates at #500 million, and the fact that it will create 1,000 jobs, which will please the City and government. ICL will train 4,000 staff as Microsoft engineers, which will further its intention of presenting itself as a service company instead of a hardware supplier.

ICL recognised earlier than most companies that there was no future in shifting tin and that the only way for a hardware manufacturer to survive was to look to niche markets and the services sector.

The fact that it's 90 per cent owned by Fujitsu, which manufactures its own and other companies machines, means ICL is free to concentrate on its chosen mission of systems integration and services.

Boswell admits NT's reputation for scalability has been under a cloud, but believes the problems will be sorted out. 'The level of NT scalability is accelerating rapidly. NT 4 is already capable of critical applications and NT 5 will improve on that,' he claims.

Robin Bloor, chief executive at UK consultancy Bloor Research, believes the ICL/Microsoft agreement raises some important issues. 'It's rather like the early days of EPOS (electronic point of sale), which, although technically possible, took a long time to take off with the customers.

The idea that one company that knows nothing about building interfaces has an agreement with another company that knows nothing about building interfaces is absurd,' he says.

Paradoxically, ICL's decision to abandon Unix development has been partially welcomed by some Unix users. Interforum, formerly Uniforum - the independent Unix user group - believes the move might clarify some outstanding issues and work in Unix's favour.

'Fewer competing strands of the OS might mean that a single Unix standard finally appears,' says Peter Gottlieb, an Interforum board member and director of strategy at Unix software house, Tetra. 'One of the reasons why there were so many strands of Unix is that the developers needed to differentiate themselves from one another to make sales. You could see the same thing happening with NT.'

No other company has made such a sweeping commitment to its operating systems development. All the suppliers, including Microsoft's arch-rival IBM, have acknowledged that NT will be a vital OS. IBM, one of the last to endorse Open Systems and Unix, has publicly stated that it aims to be the leading supplier of NT-based systems. In one respect, ICL's effort to force all hardware manufacturers down the Unix route has succeeded.

Having lost the campaign for preferential procurement, ICL adopted a stalling procedure. By convincing the government that only companies conforming to Open Systems standards - in effect Unix - were worthy to bid for state contracts, it forced IBM onto the back foot. Always a defender of its lucrative proprietary-based systems, IBM had no Unix product capable of winning contracts within the government's and, eventually, the EU's criteria.

In the short term, it solved the problem by buying AIX from the US Unix developer, Interactive Systems. AIX was originally used on the RT PC multi-user workstation launched in 1986, but went on to form the basis of a much more convincing Unix box, the RS/6000. Effectively, ICL bought the time it needed to consolidate its position in the public sector.

But ICL had missed the main thrust of IBM's long-term strategy. Far from throwing money and resources at Unix, IBM had a hidden agenda that would remove any hurdles that ICL or any other company could persuade European bodies to put in its way.

Quite simply, IBM decided to make its proprietary systems, most notably MVS, Open Systems compliant. It was aided by the Unix standards-setting body, the Open Group, which decided any OS that supported Unix APIs could be classed as an Open System.

For 25 years, ICL championed the cause of Unix and Open Systems while retaining a commitment to its own proprietary OS and adopting MS Dos, Windows and then NT. This in itself isn't unusual. ICL's competitors from the mainframe days - Bull, Unisys, Siemens Nixdorf, Olivetti and others - all followed a similar strategy.

But none has gone as far as ICL in declaring that the future lies solely in NT. The others may be switching emphasis in terms of R&D resources to NT and away from Unix, but they haven't publicly abandoned it.

ICL may hope to fully ingratiate itself with Microsoft by abandoning Unix and focusing on NT in a bid to steal a march on its rivals, but it's unlikely that Microsoft will be impressed. As far as the IT industry is concerned, Microsoft chief executive Bill Gates is king and ICL, one of the many minor monarchs is paying him lip service.

PARTNERS IN CRIME

ICL's marriage to Unix was founded not on love, but on the principle used by feudal kings of marrying their daughters to the sons of other monarchs to strengthen their hands against rivals.

The competitor in this case was IBM which, 25 years ago, was a well armed antagonist camped on the borders of ICL's installed customer base.

ICL wasn't alone in fearing Big Blue. In France, the national champion, CII-Honeywell Bull, (now Bull) faced a similar challenge as did Siemens Nixdorf in Germany and Olivetti in Italy. IBM represented a considerable threat to the national sovereignty of the IT companies.

The figures speak for themselves - in 1978, US firms supplied 45 per cent of computers in Japan, 60.5 per cent in the UK, 75 per cent in Germany and 83.5 per cent in France.

All Open Systems go

ICL and its European counterparts were sheltered from the storm by their governments' policies of preferential procurement, which favoured the national suppliers. But this safety net couldn't survive forever. The US companies were significant employers in the UK and Europe and were demanding a freer market for their products.

ICL launched its counterattack on IBM by lambasting the proprietary nature of its operating systems which locked the vendor into the supplier. The answer, ICL suggested, was Open Systems, where a universal OS would run across all hardware platforms.

Alan Benjamin, ex-director of corporate communications at ICL, admitted after his retirement that Open Systems had been adopted as an anti-IBM ploy. The operating system was Unix.

On the surface, it was an ideal solution. Hardware suppliers would no longer be able to jack up prices of operating systems and hardware at their whim. An application developed for one Unix machine would port easily to another from a rival supplier and Open Systems appealed to everyone, especially the hard-pressed Europeans.

There was only one problem - it didn't work. There were so many flavours of Unix that portability of applications was just as difficult as under a proprietary system. Unix, a system developed by and for academics, was totally unsuited to the commercial world, lacking such essential features as data security and data integrity.

Despite this, ICL continued to push the system. Now it has dropped all further Unix development, it's finally admitting that, as in so many other things, even as an anti-IBM ploy, Unix was a failure.