Materialization refers to the pre-computation of aggregations across dimensions in a data cube. Full materialization refers to pre-computation of all aggregations across all dimension combinations. This can increase the speed and efficiency of queries but is not scalable to high dimensional data. Alternatively, only the most commonly queried aggregations can be pre-computed, which is referred to as partial materialization. An iceberg cube is a materialization and is a nod to the fact that much of the data is actually "below the surface".