A System of Privacy Preserving Distributed Spatial Data Warehouse Using Relation Decomposition

The article presents relation decomposition as a method of preserving data confidentiality in new distributed spatial data warehouse architecture that we proposed. Each node in new architecture has its own privacy preserving module which can communicate with other modules. It enables parallel query processing through all privacy preserving modules and as a result it shortens the time of query processing in DSDW. Procedures processing basic SQL as well as typical OLAP queries for protected data were also implemented. Those procedures realize execution of queries in such a way so that the greatest possible number of processing stages could be executed parallel. Additionally the privacy preserving module was designed to enable spatial data mining in DSDW. In order to do those two density-based algorithms of distributed clustering were used. First of those is applied if data warehouse nodes and privacy preserving modules are placed on internal servers of a corporation, second one is used if nodes and modules are on external servers of third party companies. Tests of effectiveness of executing queries as well as data mining, both using privacy preserving based on relation decomposition are finally presented.