EFFICIENT OLAP FOR BIG DATA

Abstract

OnLine Analytical Processing (OLAP) faces many challenges in the era of big data. The exploded volume of the immutable data hinders OLAP system achieving high service level agreement with traditional indexing mechanisms. Further, the traditional OLAP system can hardly support new analysis queries such as cohort query or its variant in user behavioral data. In this dissertation, starting from the data indexing in storage layer to query processing engine, we present a general framework for emerging OLAP system in a bottom-up manner.