InfiniDB Pushes Performance in SQL-on-Hadoop Query Performance Benchmark

InfiniDB, which recently introduced InfiniDB 4.5, has announced the results of a new benchmark test conducted by Radiant Advisors to compare the performance of leading open source SQL-on-Hadoop query engines, including InfiniDB for Hadoop 4.0. 

In the benchmark, which was conducted in Q1 of 2014, the SQL engines were installed in various data file encodings in three Hadoop clusters on identical Amazon EC2 servers.  Piwik was the application workload selected for the performance benchmark from the open-source web analytics application,  The Piwik schema included three tables representing user visits and actions.  A base of 3 years of initial data was provided, and then was duplicated to represent 100x user activities over the same 3 years. In total, this schema represented approximately 1TB of user data. After loading data files, all queries were manually run independently at least 10 times in each cluster, SQL engine, and file encoding.   

According to Jim Tommaney, CTO of InfiniDB, one of the key takeaways from the benchmark is that InfiniDB was able to handle the widest range of analytic queries – with large performance improvements. According to the company, the InfiniDB engine was the only SQL engine that completed all reporting, ad hoc, and analytic queries in the benchmark, and was the fastest in 6 out of 10 queries.

Separately the company recently announced InfiniDB 4.5, with performance enhancements, expanded ETL support , new Hadoop capabilities, and improved administration and monitoring capabilities. “With 4.5 we are announcing the availability of our enterprise management console. It is a nice layer top that allow for administration from a central point, allows for a real time and historical view of system metrics, real time and historical view of SQL activities on the system. It is a very powerful tool,” said Tommaney.

To register for an April 23 presentation of the benchmark results and receive a copy of the report, go to