 Suppliers of IT and IT systems tell you how valuable your data is. They tell you how it helps you to understand your customers and ultimately increase sales. The burden of ever-increasing data grows over time, however. Every new IT system generates more data to be recorded and requires further and larger systems to analyse that data. It is a never-ending upwards spiral that locks in small and large business alike. Faster than the capacity of technology can increase, the appetite is fuelled. For the giant IT suppliers this is heaven-sent, for only with aggressive replacement of old IT systems can they meet their even more aggressive targets and Wall Street expectations. The “more for more” argument is easy to sell. All your suppliers are aligned to it – the hardware manufacturers, the software vendors and the consultancies or systems integration houses are all on the same upwards track. It makes sense, and doesn’t rock the boat. It will carry on forever. Or will it? Indeed, can it? In Canary Wharf, for instance, it certainly cannot. The power supply network is already at its limit because of the enormous data appetite of London’s newest financial district, and as a result no new system can be deployed without decommissioning an old one first. This issue exposes the ultimate problem for the continued expansion of IT systems – the amount of electricity used to power the swelling ranks of servers, and, even more alarmingly, the power used to cool them. When figures are quoted indicating that the average data centre uses as much power as a medium-sized European city, and the largest require the local fire brigade to hose the buildings down on the hottest days of the summer, someone must call time on this practice. One solution is to revisit the software used to organise the data rather than merely installing new hardware for storage. It is possible to organise your data differently so that it uses less storage and less processing power, to deliver higher performance and better business flexibility. This might sound hard to believe, but it’s true. The analogy with old and new building practices is instructive. High-rise buildings that used to rely on the tower-block model of a central supporting pillar had limitations – the base of the pillar had to get wider and wider to support a taller building until no more floor space was provided by adding a new floor. Compare this to today, where the tallest buildings are now designed with “exoskeletons” to provide strength through a carefully planned structure. Most large data analysis systems are still being built with the equivalent of the old-fashioned tower block, the relational database management system (RDBMS), which explodes data storage and processing power demands so that smaller and smaller increments in performance and effectiveness are gained at greater and greater expense, not to mention the demands on air-conditioning and power supply. The RDBMS was the great mainstay of online transaction processing, enabling many users to edit and maintain individual customer or transaction records – but when it comes to deriving business value from the analysis of many millions of these records with rapidly changing questions, an RDBMS just can’t do it efficiently, and the cost-performance curve hits a brick wall. One solution is Sybase IQ, a database engine purposely designed for business intelligence. Compared to using an RDBMS, IQ can reduce storage footprint by up to 90% and increase load and reporting performance by 10 to 100 times. Another solution is Teradata, a powerful solution for analytic work but one that uses a lot of hardware – a bit like stacking multiple engines in line to make a car go fast, compared to which the IQ solution is akin to installing a single jet engine. Another data enterprise provider, Netezza, uses some of the design principles behind IQ but optimises efficiency when deployed on a smaller scale owing to its use of proprietary hardware. Whatever the solution, power consumption and cost efficiencies have moved right to the forefront of data management. Implemented properly, the latest software architecture offers that elusive “more for less” factor – truly a step change in performance and reduced cost. But, as with any disruptive technology, there is a rub. The software has been produced that can deliver these promises, but to whose cost? Your hardware supplier and managed service partner, whose business models are based on ever-increasing numbers of servers, will surely not be in the vanguard of this impending revolution. Alf Franklin is sales manager for Sybase UK’s cross-industry group. He holds an MA in mathematics from Lincoln College at Oxford and an MBA from the London Business School |