Much effort and concentration has been put into devising training regimes for a number of different technologies in distributed and high-performance computing (Jandric, Artacho, Hopkins, & Fergusson, 2008; Fergusson, D., Romano, van der Meer, & Atkinson, 2008). On the whole, however, these have tended to concentrate on the computational aspects of research tasks rather than the data-related aspects. There have been a number of reasons for this including the immaturity and extra complexity of the data field, the more discipline-specific aspects of data usage compared to computational patterns, and the focus of providers on the “easier” problem of providing distributed computation resources (Fergusson, 2006).