IGC-Eddie3 high performance storage arrays

We combine the latest computational and experimental technologies to investigate how our genomes work to control the function of molecules, cells and tissues in people and populations. For more than half a century our research has been dedicated to understanding human genetic disease. Today we continue to apply our clinical and scientific expertise, harnessing the power of complex data, to improve health, and the lives of patients and their families. The IE3 high-performance computing cluster provides the analytic power to analyse rich and complex data that underpins most of our research and discoveries, for example in the mechanisms that drive cancer progression. The storage arrays requested here are needed so that data can be moved to computer processors and results received back quickly enough to efficiently make use of the computing power available and to be able to analyse the large quantities of data needed for these studies - for example the analysis of hundreds or thousands of whole genome sequences.

The IGC-Eddie3 (IE3) cluster underpins all the research at the MRC Human Genetics Unit where we combine the latest computational and experimental technologies to investigate how our genomes work to control the function of molecules, cells and tissues in people and populations.

All seventeen core HGU groups and most affiliate groups make use of IE3, encompassing >200 individual researchers at all career stages. For at least nine of the research groups (Ponting, Semple, Taylor, Marsh, Ewing, Khamseh, Colhoun, Vallejos, QTL) IE3 can be considered to represent the principal research tool. It is a dedicated resource optimised and exclusively for the use of the MRC Human Genetics Unit and associated Institute of Genetics and Cancer (IGC) research groups.

IE3 is a modular system comprising compute units and high-performance storage arrays. The high-performance storage arrays provide a temporary data storage area with sufficient speed and capacity to provide data to, and receive from, the >2,500 CPU core compute cluster. This is essential for genomics research which heavily depends on large-scale data movement, e.g. parallel analysis of thousands of whole genome sequences or RNA-sequencing from thousands of single cells. Efficient large-scale data movement is also essential for high content image analysis, central to the 4D cellular medicine initiative at the HGU.

The current high-performance storage arrays are end of life and operating on an extended warranty. Individual disk failures are now becoming common and synchronous failures risk the loss of data and protracted system down-time if the storage infrastructure needs to be repaired. The proposed replacements are model updates of the existing end of life hardware components that have had an excellent track record in delivering reliable, sustained high-performance storage for the continuously running cluster over the last 5.5 years.