Theta-join SQL Operators Optimization on Distributed Systems

This project aims to design algorithms to accelerate theta-join SQL operators on distributed systems.

Project Details

• Algorithms:
  • “Equal” SQL operator acceleration algorithm
  • “Less than” SQL operator acceleration algorithm
  • “More than” SQL operator acceleration algorithm
• Environment:
  • Memory: 40 * 50G
  • CPU Cores: 40 * 7
  • Platform: Spark 2.2.0 + JDK 1.8
• Data Set:
  • TPC-H
  • 1G, 10G, 100G
• Performance:
  • 5 ~ 400 times faster than Spark SQL.

People

Zijian Li

Xiaofei Zhang

Wenjie Liu

Associate Professor