The mapping system we use will only take into account the structure of the data type, and the name of the node that is to be allocated. The drawback of this is that we do not take into account which nodes already exist. The result of this will be that for unbalanced structures the nodes will not be divided out as equally as possible, some processors may have many more nodes than others, and will therefore be required to do an unfair amount of computation. This is the disadvantage of a static allocation scheme, as opposed to a dynamic scheme, which would take account of the imbalance and re-distribute the nodes as they are allocated. A dynamic allocation scheme would be a valuable addition to the whole system, but is not a problem that will be considered here.