Online analytical processing (OLAP) is at the core of business data analytics. OLAP is part of the business intelligence software provided by e.g. IBM, SAP, Oracle and Microsoft for the decision support systems of many corporations. In contrast to online transaction processing (OLTP), OLAP queries can be very complex and compute intensive because they may require data aggregation over a large portion of the database. Therefore, many traditional OLAP systems build pre-aggregated OLAP data cubes that are stored in addition to the OLTP system and are updated in batches (e.g. once a week). We study the use of parallel computing to speed up OLAP systems. Our goal is to provide performance increases that are sufficient to allow for real-time OLAP on large databases. Such real-time OLAP systems will allow corporate decision support systems that operate on the latest OLTP data and support real-time decision making on big databases.
This project is part of a long term collaboration with IBM Canada. Our team won an IBM Canada Innovation Impact of the Year Award. IBM Canada gives one such award per year.