New ways to handle huge data problems
Posted on 13 November 2002
Now the University of York is taking the lead in a £3 million project which will link networks of computers and use 'grid technology' to manipulate vast amounts of information as it looks at aircraft maintenance by remote control.
The project, known as DAME (Distributed Aircraft Maintenance Environment), will be headed by Professor Jim Austin of York who will lead a team of more than 30 researchers, students and academic co-investigators based at York, Leeds, Sheffield and Oxford universities.
It is funded by the Government's e-science research initiative and involves major industrial partners including Rolls-Royce, Data Systems and Solutions (a joint venture between Rolls-Royce and SAIC, the global research, engineering and IT company), and Cybula Ltd, which is a University of York spin-off company.
DAME aims to build a grid test bed, or test system, based on a network of computers, for distributed diagnostics, and will apply it initially to maintaining civil aircraft engines. 'Distributed diagnostics' are a system for finding faults in systems run on many computers. The technology could also be applied to the medical field, pharmaceutical and other areas of manufacturing, and to transport problems.
The White Rose Computational cluster, which has grown from existing collaboration between York, Sheffield and Leeds universities, will provide a vital resource for DAME in building a £2.8 million computing infrastructure. Oxford will add its own resources to this facility.
Professor Austin said: "Scientists are facing massive increases in the amount of data available. In order to process, analyse and store this information new computing hardware and software needs to be developed - this is at the heart of e-science.
"This research will also benefit business, commerce and education in the long run.
"The essential research theme is real-time intelligent feature extraction, advanced data mining and state of-the-art decision support tools, distributed on a global basis."
"For example, we will be looking at data generated by 100,000 Rolls-Royce engines in flight (each producing over 1Gb of data per flight). The size of the databases needed to handle the information, and the need for distributed access to these data, make it a particularly demanding challenge for grid technologies. The research will capitalise on the experience of data grids in the USA and will address performance issues such as large e-scale data management."