Presentations given by group members at external events

Using the In-Memory Columnar Store to Perform Real-Time Analysis of CERN Data

Name: Using the In-Memory Columnar Store to Perform Real-Time Analysis of CERN Data
Start: 2014-10-02T13:15:00-07:00
End: 2014-10-02T14:00:00-07:00
Location: Other Institutes

by Emil Pilecki (CERN), Manuel Martin Marquez (CERN)

Thursday 2 Oct 2014, 13:15 → 14:00 US/Pacific

Moscone South 102 (Other Institutes)

Moscone South 102

Other Institutes

Description

To help physicists at CERN analyze their data more efficiently, a study was done of the use of database technology to store and analyze data recorded by the experiments at the Large Hadron Collider. Because reconstructed physics objects are described with hundreds of variables, this physics analysis database has very wide tables. Each physicist typically requires a unique combination of variables to find the events of interest for a particular analysis, making it difficult to choose appropriate indexes for this data. This presentation shows how Oracle’s new in-memory columnar store helps improve performance when analyzing this type of data.