Computing/storage for mechanical engineering (CFD, FE/A), for product design. This would be utilising both our larger clusters and your licences.
Structuring, managing and running analytic on your data using Open Source products and often via a Cloud. We can advise and help in understanding your environment and suggest options or propose a strategy. There will be an element of understanding the new service options you can begin to use. This may inform the sequence of which department you decide to start analytics with. The process will benefit from being planned out to help set expectations.
Big Data & Analytics of your data is more down to which file system you need to solve your desired problem statement(s) and then which tool is most refined for your domain. We can advise across the spectrum of open-source options from Lustre, CEPH, and Swift - as well as added layers of Hadoop and applets. It is closely linked to how you set out your file structures, how fast is your scratch storage, how tiering can help, or how might Non Volatile memory help you. i.e. Some Genomics codes like Spark under Hadoop.
Storage of your data. To manage costs you would typically form a tiered storage to archive policy, relative to what you are wishing to achieve:
Scratch Fast IO & disk for you to load your data, compute and store the results. Your decision of how long to keep in scratch storage is driven by the problem being resolved, volume of through-put and how often you may wish to run 'what if’s' on it.
Data Store 30 days to 6 months typically – depending on how many users wish to access and will they each be running analytics or historical comparisons.
Tiered archive On site, off site, active archive.