Systems integration is suddenly the sexiest business in the computer industry, and after making cautious moves on the continent to establish joint ventures with specialist systems houses, IBM last year pulled out all the stops and turned its Federal Systems Division into a full-scale systems integration business. Now every company that finds that the market has passed its products by announces that its recovery will be achieved on the back of a major new systems integration business. That is the story coming from Nixdorf Computer AG, echoed by Norsk Data A/S, while in the US, Control Data Corp, Xerox Corp, Unisys Corp have all identified integration as one of the most promsing prospects for future growth. Up until the mid-1980s, systems integration had been regarded as a rather dull and pedestrian business, something that service bureau companies did when their core business vanished as all their bureau customers moved over to their own in-house machines. Rows That all changed when General Motors Corp acquired Electronic Data Systems in 1984, although it took three or four years for the rest of the industry to realise that the move – despite all the rows with EDS founder Ross Perot – had been a brilliant one, and that General Motors had picked off the jewel in the systems integration crown. Once IBM had identified the business as one offering exceptional growth potential, the big accountancy firms realised that they already had several of the necessary skills in place – and had the priceless sales argument that they were independent of all hardware manufacturers. Now the business is seen as so attractive that the US trade weekly Computer Systems News looks as if it is preparing to change its name to Computer Systems Integration News – for the past couple of months every other story seems to have had the two magic words – or the acronym SI – in the headline. The step that follows systems integration is facilities management, an activity that has been widespread in the UK market for many years, yet so little recognised in the US that the term was not even used in most of the news stories about Eastman Kodak Co handing its computing operations over to IBM (CI No 1,227). That news made it clear that IBM regards now facilities management as a natural extension of systems integration, but the idea should ring alarm bells in the heads of major users, because IBM is all at once seen to be trying to hunt with the hare and run with the hounds. Facilities management is often a very attractive option for large computer users once they reach the point where they feel that their core activity seems to be becoming ancillary to the all pervasive data processing department. It has been growing in popularity among local authorities in the UK, who feel that running an enormous computer department requires very different skills to those needed to meet the demands of their electorates. Another attraction is the same one as is achieved by privatising unresponsive state monopolies: when the data processing department is an in-house operation, the DP manager may very well be on the board, in which position he is very difficult to kick sufficiently hard to get any action. The picture changes dramatically when when there is an IBM or a Hoskyns Group Plc contracted to deliver defined data processing services at an agreed price under a three-year contract with options on two further years. Ugly The promise that if what the user wants is not delivered yesterday then the options will not be taken up and the contract will be put out to tender in 18 months is a splendid weapon to be able to wield over an intransigent data processing department. The contractor has failed if the switch to facilities management does not lead to a more responsive service and save money over the cost of the in-house facility. And it is not difficult for a competent integrator to deliver a cheaper service: spare time on the in-house computers taken over can be sold to other customers or used for the facilities manager’s own work, and network costs can be shared with other fa
cilities management customers. But when the facilities manager is a major manufacturer wanting to use its own equipment as far as possible in the installation, potential conflicts of interest rear very ugly heads. Does IBM, equipping and running a company’s computer installation with large MVS/ESA systems, pass on the full list price for the hardware – and more importantly the software – to the customer, or does it set a – much lower – internal IBM transfer price? If the former, does it meet the cost of the customer’s insistent and strident demands by cutting costs on the hardware and using 308X machines under MVS/XA instead of 3090s under MVS/ESA and take every measure possible to ensure that that information does not leak out to its customers that still keep the computer department in-house? Almost by definition, the relationship of the customer and the manufacturer of computer systems has to be an antagonistic one: the customer’s interest is in getting its computing done as cheaply as possible, the manufacturer’s interest is in squeezing as much money out of the customer as possible. When the manufacturer effectively becomes its own customer on behalf of the ultimate user, the scope for anti-trust violations becomes legion. Spades Kodak has done IBM an enormous service by becoming its first major commercial facilities management customer – if the great Eastman Kodak Co is happy with the arrangement, why should anyone else worry about it? – but the argument for users confining their facilities management business to companies that are manufacturer-independent are overwhelming. Users need to insure themselves against commercial disaster by insisting that industry standards – MS-DOS, Unix, Open Systems Interconnection, Fibre Distributed Data Interface – are used wherever they exist, not only because they will facilitate handing the work over to another manager or taking it back in-house if things go wrong, but also because the next two or three years are going to see a rapid divergence in the costs of proprietary and standard solutions to the same problem, and very few manufacturers have all the standards in place and implemented. And the temptation for any manufacturer will always be to use its proprietary in house solution rather than buying in a standard one. And the arguments against going to a manufacturer apply in spades when that company is not IBM but a smaller manufacturer with a more limited range – and one suffering financially. The temptation will be to use in-house equipment wherever remotely feasible, and the customer runs the risk that a financially troubled facilities manager may not even be around in five years to clear up whatever mess it may make. All of which seems to make the case for confining the bidding to vendor-independent facilities managers overwhelming.
