The Business Intelligence on Hadoop Benchmark Q4 2016.

As more enterprises run BI on Hadoop, a critical question arises: "How "fit" are SQL-on-Hadoop engines to satisfy BI workloads?"

In the first study of its kind, we answer that question --- with performance measures derived from real-world testing on actual workloads --- across the latest engines: Impala, Spark, Presto and Hive. 

Screen_Shot_2016-09-26_at_1.49.10_PM.png

Our evaluation looks at how well these engines satisfy 3 key criteria: 

  • Performs on Big Data
  • Fast on Small Data
  • Stable for Many Users

...with the latest data updated for Fall 2016.