If there is one thing that gets up Graham Taylor’s nose it is people referring to ICL Plc as the UK’s only mainframe manaufacturer. This is because Taylor is ICL’s business manager for software, and, above all, ICL would like to be known as the open systems company. Hardly surprising, find me a hardware company above the personal computer level that doesn’t want to claim that tag for itself, after all it is tough to sell proprietary systems today. So what is different about ICL? Well, of all the big hardware vendors attempting to take software seriously ICL does seem to be piling up credibility for shunning the lock-in mentality. It helps that Taylor is also UK Group Chairman of Unix International Inc, it helps that ICL has wholeheartedly adopted Open Systems Interconnect – it offers nothing else, and it will help that ICL is an extremely active member of the Object Management Group. This string of Brownie points cannot simply be born of fortuitous circumstance – could it be that ICL is taking open systems seriously? Of all the companies that have moved from a revenue base emanating from proprietary hardware to one driven by software and services and Unix, ICL, as has been frequently pointed out in Computergram, is the only one that has been consistently profitable while others have floundered. As Taylor puts it ICL is making money out of open systems.
Make money
Others he feels simply saw the open systems route as a kind of Utopia, a way out of their financial and marketing difficulties, which it isn’t. The open systems market is a hard market, very competitive and with margins often dipping into the red. The only way to make money is to be more competitive by exploiting entries into other vendors’ markets. And how do you do that? Well it all comes back to software and related services, which constitutes around 30% of ICL’s UKP1,500m turnover. But, as Taylor points out, its not just a question of revenue contribution – it is how software and services are perceived within the company. In ICL software revenue and profit is seen in its own right, whereas most system suppliers still see software as a pull through from hardware or for hardware. Two years ago ICL software was measured as a hardware return but now it is evaluated in its own right and approximately two-thirds of ICL’s engineering resources are devoted to software. Every system supplier nowadays seems to think if it has a software architecture or a framework people will take its software pretensions seriously. To wit ICL has its OpenFramework, which encompasses user interfaces, application development, application architectures, distributed application services, information management, systems management, networking services across openVME, MS-DOS, OS/2, DRS systems, Unix System V.4, and systems supporting Posix, X/Open XPG3 standards. The unfortunate thing is there are probably a dozen or so such architectures knocking about covering similar areas. OpenFramework differs from many in that it is not pushing product so much as interfaces and is probably most directly comparable to Bull HN Information Systems Ltd’s Distributed Computing Model (CI No 1,634) with the proviso that Bull is approaching the enterprise model of computing from the Open Software Foundation’s diktat, while ICL is working from the OSI’s basic seven layer model, is using X Window interfaces and is embracing moves by Unix International. For example, it is developing an OSI version of Tuxedo with Unix System Laboratories Inc for its distributed application services, although it will also support OSI-RPC, OSI-TP, LU6.2 and APPC connections. Information management is built around SQL and supports Ingres, Oracle and Informix databases.
By Katy Ring
So far so good, but the elements that really separate ICL from other systems suppliers in the software stakes are to be found in the areas of application development, systems management and application architectures. The first fairly secret weapon that ICL expects to launch in 1993 is its answer to the Repository repository being a word it dislik
es intensely. Taylor explains that ICL’s Data Dictionary System is not of the same status as what other vendors refer to as a data dictionary, since it has always been concerned with holding both business and code level information rather than just offering lower-level software engineering linkages. However, sources claim that ICL’s real strength is that its current data dictionary technology is advanced enough for it to work out that you cannot build a tenable repository without the underlying database technology being object-oriented and this is where IBM’s attempts will flounder – DB2 is not an object-oriented database, is it? The second main strength of ICL’s OpenFramework is its Application Architectures, which can be viewed as technical solutions or business solutions and a fast route through OpenFramework for end users, enterprise managers, service providers and application developers. Architectures exist for the retail, financial services and office automation markets to enable collaborators and users to develop their own products. However, Taylor says that many large users are adopting them as blueprints to guide them through an application development life cycle that conforms to existing standards – the overall conformance is to Unix International’s Atlas, but that includes the Open Software Foundation’s Distributed Computing Environment and ICL has provided such architectures to some of its retail users that are using a host running MVS as a server.
Management of objects
ICL’s other software forte is Open Systems Management Centre. According to David Hacker, whose baby the Management Centre is, ICL is concentrating on the management of applications and is currently negotiating to buy in a network management product suite. This distinguishes it from the likes of IBM Corp and Digital Equipment Corp. Second-ly, it perceives system management as the flexible management of objects as promoted by the Network Management Forum, whereas other vendors see it as the management of a large monolithic system. And thirdly, it offers the Unix workstation as the central point of command and control for the computing infrastructure. Consequently, it is not surprising that several hardware manufactures are negotiating to use buy this suite of management products OEM from ICL – certain hardware manufacturers will no doubt spring to mind as the Management Information Centre is exclusively System V.4-based. There will be a VME release early next year. ICL views the Management Centre as a rival to SystemView – it can manage applications on Unix – including OSF/1 derivatives, VME, VMS and OS/2. In fact it can get into MVS but this is not an option ICL is marketing as yet. So at the end of the day, is ICL destined to be an also-ran, me-too open systems company? It has the technological edge in several software areas as its OpenFramework shows but open systems depend on the support of other software developers willing to buy into an architecture and will they buy into ICL’s? After all, most Western software development is done in the US and how many Americans have even heard of ICL? However, those secretive OEM deals with hardware manufacturers may encourage software implementations – will they help to raise ICL’s profile?
