In clusters working on large amounts of in memory data, the approach is often to load the data, then move the code (e.g. a java class implementing some data pro

In clusters working on large amounts of in memory data, the approach is often to load the data, then move the code (e.g. a java class implementing some data procesing interface) to the data as required, rather than move the data to the code.