The vast majority, 88%, of data processing managers have been evaluating massively parallel processing for commercial data processing this year, says Gartner Group Inc, which asked 30,000 managers about parallel processing. It explained the interest as a factor of the high cost of using legacy systems for stuff like decision support and data warehousing. It reckons users can expect a return of up to seven times on their investment in parallel processing over a five year period – five times over two or three years. But Informix Software Inc believes customers are still far more likely to choose symmetric multiprocessing over the massively parallel variety for their high-end requirements, because the latter represent a substantial administrative overhead.