PC and UNIX – is the price right? (September 1998)

Price comparisons of the cost of ownership ofPC and UNIX-based systems show savings in favor of the PC. PDM looks behind the headlines.

The debate continues as to the relative merits of UNIX and PC’s – in fact PDM devoted a whole issue to this way back in 1996 (PDM Vol. 1 No 3). In our previous examination of these issues, we concluded that the PC was fast approaching UNIX systems in terms of hardware performance, and had great potential for cost advantage because of the larger user base. Since then, the UNIX camp has reacted to this by aligning much of its hardware with the PC world so that memory, and many peripherals are now interchangeable. Generic UNIX software (Operating systems, development tools and office automation) is still more expensive but not dramatically so. The only real argument for cost savings is that of application pricing. Typical of such price arguments is that offered is that from Seismic Micro Technology (SMT) (www.seismicmicro.com). Briefly, the cost of a license from this vendor of PC-Based seismic interpretation software is around $10,000 per seat, compared with around $70,000 per seat for an ‘equivalent’ UNIX based system. A comparison of the maintenance costs for a 50 user site give a further price advantage of around $1.5 million for a 50-seat shop over a 3 year period. These sort of figures make seductive reading in this cost-conscious world, but of course they are not entirely un-biased.

Several protagonists would back these arguments. Microsoft first and foremost. PC-based vendors such as SMT above, and the generic software vendors such as Intergraph who made a move from their own bespoke hardware and software combination some time ago. On the Microsoft/Compaq stand at the SEG, a Digital 40100, with 4 x 400MHz Alpha processors and 16GB of main memory on display at the SEG was quoted as being "1/10^th the cost of equivalent UNIX system" - but again this is mainly because of application costs. Whether such differentials are sustainable is questionable. After all the inroads made by NT are vulnerable to price-cutting from the UNIX camp. Should this happen this would drive down profitability of sector, and could end up rather messily. Collective suicide is actually rather improbable so the price argument may be overstated. The overstatement is also evident when one reflects on what really drives down software costs. Microsoft’s success has come from piling them high and selling them cheap. Whether these be operating systems, development tools or office automation. In short in the mass market. This effect does not really exist in a narrow vertical market such as E&P IT.