Actually, a lot of attention focusing on the problem of computing privacy-preserving OLAP cubes effectively and efficiently arises. State-of-the-art proposals rather focus on an algorithmic vision of the problem, and neglect relevant theoretical aspects the investigated problem introduces naturally. In order to fulfill this gap, in this paper we provide algorithms for supporting privacy-preserving OLAP in distributed environments, based on the well-known CUR matrix decomposition method, enriched by some relevant theory-inspired optimizations that look at the intrinsic nature of the investigated problem in order to gain significant benefits, at both the (privacy-preserving) cube computation level and the (privacy-preserving) cube delivery level.
