Parallel Data Cube Construction: Algorithms, Theoretical Analysis, and Experimental Evaluation

Data cube construction is a commonly used operation in data warehouses. Because of the volume of data that is stored and analyzed in a data warehouse and the amount of computation involved in data cube construction, it is natural to consider parallel machines for this operation. This paper presents two new algorithms for parallel data cube construction, along with their theoretical analysis and experimental evaluation. Our work is based upon a new data-structure, called the aggregation tree, which results in minimally bounded memory requirements. An aggregation tree is parameterized by the ordering of dimensions. We prove that the same ordering of the dimensions minimizes both the computational and communication requirements, for both the algorithms. We also describe a method for partitioning the initial array, which again minimizes the communication volume for both the algorithms. Experimental results further validate the theoretical results.