Interactive Big Data Exploration on Modern Hardware

PI: Alfons Kemper, Viktor Leis, Thomas Neumann (TU Munich)
Project Collaborator: Alexander van Renen
Project website: To be announced

Currently, Big Data exploration is carried out outside of database systems with distributed frameworks such as Hadoop/MapReduce, Spark, or Flink. These systems, however, are batch processing systems that cannot support interactive response times because of their need to transfer huge volumes of data to the processing units. To achieve interactive response times we propose to deeply integrate big data exploration functionality into the core of database systems. In this project we want to exploit the latest and emerging hardware trends to scale database systems threefold: (1) in functionality by incorporating advanced exploratory capabilities, (2) in performance to provide interactive response times for complex exploration workflows and (3) in data volume that can be efficiently processed.