Oracle Corp has been having a difficult time of late – for the first quarter of 1991 it reported its first ever loss, its share price has slumped to around a quarter of its value over the past year and is now hovering around the $6 mark while it is consistently criticised by rival vendors for being all hype and old product – an assumption given credibility by the ButlerBloor database survey. Meanwhile, both IBM and DEC appear to be developing databases to rival Oracle in the third party database market (CI Nos 1,552, 1,564). Could we be witnessing the demise of the largest independent database vendor? Yes and no.The question operates on a variety of levels. First there are the financial issues: how deep a hole is that first quarter loss? Oracle UK’s director of marketing Mike Evans says that Oracle’s expectation is that it will be profitable in the second quarter.

Conventional growth

The root of the problem according to Evans is that Oracle got its cost base wrong – it was planning at an increase of 50% or more and, in the first quarter of 1991, there was a respectable but not dramatic growth rate of 16% in revenues. Historically, Oracle’s first quarter is always its weakest so the fruits of any bad planning are likely to emerge at this quarter. Now Oracle is following a more conventional planned growth rate of 25%. It appears that the company was, among other things, caught out by the slowdown in the US economy – the US market accounts for 50% of of its revenues. Oracle is not banking on any growth in the US for the 1991 financial year, but is hopeful of a 50% growth rate in Europe. One factor that emerged from the first quarter results in the UK that was widely overlooked, says group product marketing manager John Spiers, was that for the first time Unix revenues surpassed VAX/VMS revenues. Over the past couple of years Unix revenue has not grown as fast in Europe as in the US, but Unix is becoming the largest part of Oracle’s revenues – in fiscal 1990, 36% of revenues came from Unix up from 27% last year, 42% came from VMS sales down from 47% in the year before, while revenues from other environments declined from 26% to 22% over the year. Oracle is best known for its eponymous database product, which is being reduced in significance as a source of revenue for the company – its revenue stream is expected to fall as low as 40% of the group’s total sales over the next few years, whereas historically it has dominated Oracle’s revenues. In terms of research and development, 12% of revenues went into the total budget of which 3% was capitalised. Spiers said that a big development effort had gone into graphical user interfaces, networking, CASE and application products with less emphasis on the database management system. However, Spiers strongly refutes suggestions that the Oracle engine is old technology. He says that the current version of Oracle – version 6.0 – has a rewritten underlying layer of code to cope with row-level locking, the mechanism for physical disk input-output and so on. In version 7.0 – due next year – richer functionality has been added on top of this. He claims Oracle is being rewritten gradually and that no code goes back further than five years. However he believes that the rush to push new releases down people’s throats is a sign of immaturity, pointing out that around half of Oracle users still have version 5.0. The mature Oracle apparently has a much more relaxed attitude to the period of time it takes users to move from one version of its database to another.

By Katy Ring

Having pointed out that Oracle is not in the business of introducing gizmos, Spiers moves on to outline some of the key functions to be expected in version 7.0: it will have fully transparent two-phase commit; stored, shared procedures in the database, stored as database objects and those procedures can be triggered in the database so that, for example, business rules can be implemented; it will be able to support massively parallel systems so that thousands of users can run applications under transaction processing mo

nitors; it offers systematic extended national language support, which means that multiple users can access the database using different languages; it also supports very large objects – 2Gb BLOBs in fact. Of course, it is a well-known fact that Oracle is committed to the commercialisation of parallel processing and Oracle can run on the Meiko Computer Surface, the Parsys Super Node and NCube. These systems bring the cost of CPU-intensive transaction processing down in leaps and bounds, and Oracle claims that there has been no resistance to the Unix-based nature of them Spiers says that banks, oil companies and airline companies are trying out the technology in an effort to gain competitive edge. Meanwhile Oracle has gateways from Oracle to RMS, IDMS, IDMS/X, DB2 and SQL/DS – furthermore, Oracle is working on communication with IBM flat file formats such as VSAM, ISAM, on gateways to Software AG’s Adabas, the Siemens Siebas database, and via SQL/Access to relational databases such as Informix, Sybase and, should customers require, even DEC’s Rdb via DRDI. Meanwhile, Oracle’s CASE toolset, which already supports DB2, as well as Oracle, is currently being developed to support ICL’s DDS, and Knowledgeware Inc’s IEW, as well as IBM’s Repository. Oracle sees DEC and IBM as its prime competitors and says its database product is distinct from their offerings in that it is not tied to any particular system. That being so, is Oracle worried by suggestions that both IBM and DEC are gunning for the Unix third-party database market? Taking IBM first, Spiers is of the opinion that the AIX database will not appear in less than three years, which is too late to capitalise on RS/6000 sales, since those users won’t wait for a database. As for DEC, Oracle is clearly not impressed by the recent Information Network announcements (CI No 1,548) because it does not believe users will opt to move applications to a proprietary database, however, should DEC come out with a database that is not Rdb, then it recognises that it would have to sit up and take notice – but even then it has a five year lead on portable software for Unix systems.

Silver-haired

Will Oracle still be operating as a database company in five years? According to Mike Evans part of the answer is no, for he believes that in five years’ time Oracle will be predominantly an applications and software services company – but he is clear that these are market areas that Oracle will enter to exploit its base database and CASE technology. There is no question of it abandoning this market at present. Nevertheless, it is anxious to make a success of areas such as systems integration where it has decided to focus on imaging (CI No 1,557) particularly the integration of image into applications. Evans thinks it is possible that Oracle may be a player in the facilities management market, and a lot more time and effort will be expended upon application software over the next few years. In terms of marketing Oracle’s base technology in the coming years, Oracle appears to be fostering the idea that it is a sensible, safe, reliable company akin to IBM, rather than a go-ahead young thing delivering the latest fad – however, it should be noted that version 7.0 will deliver some object-oriented functionality. Whether this silver-haired, slipper-wearing Oracle will retire and sell out to another company only Ellison knows.