Logicalis

Logicalis helps Heriot-Watt University’s Institute of Petroleum Engineering speed up its research

Company Profile

Heriot-Watt Institute of Petroleum Engineering is a leading international Institute with a reputation that stretches worldwide. The Institute provides research and teaching excellence and is the only institution in the UK to receive the top 5* research rating in both the 1996 and 2001 Research Assessment Exercises with an exemplary quality audit rating for its relationship with industry. The Institute has major research programmes funded by both industry and research councils across the whole field of petroleum engineering.

Customer's Business Issues

As one of the UK 's leading centres of research into oil reservoir evaluation and management, it was important for the Institute of Petroleum Engineering to maintain its high standards. To do this, it requires more and more intelligent computer-based applications that are capable of generating sophisticated mathematical models and crunching vast quantities of data.

Unfortunately, the technology that the Institute was using to support its prediction of fluid dynamics in oil reservoirs was reaching the upper limits of its capability – taking almost twenty-four hours to process the simulations that researchers wished to run.

In today's increasingly competitive research culture this was simply not fast enough for the Institute's requirements. In fact, such periods of delay not only slowed progress but meant that workstations dedicated to that area of the field were unavailable when simulations were running – restricting other elements of the work as well.

Consequently, the Institute wanted a new solution that would provide both the desired speed and power. Furthermore, to ensure cost-efficiency, the Institute decided that any proposed replacement should be an Intel platform running a Linux operating system.

The need for the replacement system provided the perfect opportunity for the University to further accelerate its work on fluid dynamics in oil reservoirs in other aspects. If it could find hardware which was better in every way, then researchers could use more detailed models, introduce additional factors and achieve even greater accuracy. Subsequently, for the purpose of selecting the right system at tender, a benchmark was constructed to test every proposed solution.

The business challenge that Logicalis met then, was to provide a solution that best fitted all the needs of the researchers – as set by the benchmark – whilst providing the pre-requisite reliability and lower cost that the Institute desired.

The Logicalis Solution

Logicalis was successful at tender simply because it provided the best overall solution, not least because of its status as the leading business partner for IBM Linux solutions in the UK . It built the entire solution on established IBM technology and crucially – thanks to the expertise of the Logicalis team and the company's ability to provide a turnkey solution – the entire project could be implemented in just four weeks.

The solution itself consisted of an IBM Linux platform featuring IBM xSeries servers clustered together to act as a single, high performance computing environment. The additional advantage of this design set-up has been the removal of any single point of failure, which proves fundamental when running many different resource intensive applications simultaneously. Furthermore, the world-renowned stability of the Linux operating system has further ensured safety when conducting this calibre of work.

Why Logicalis?

The solution not only met the Institute's core requirements but also offered two large advantages. Firstly, it was highly scalable, allowing the department to expand it as work developed. Secondly, it was massively parallel, meaning that researchers could execute algorithms many times over simultaneously. This greatly improved the speed with which the Institute could obtain results because researchers could be working on several aspects of a study at the same time.

The time it took to run simulations was also dramatically improved. Simulations that used to take a full twenty-four hours have now been reduced to around fifteen minutes – a marked difference enabling research to be conducted at a much faster rate.