In this edition of Talking Data, Jack Vaughan and Ed Burns discuss recent news in benchmarks and the "SQL-ization" of non-SQL databases. First off is a look at Hadoop specialist Cloudera's recent report on Impala query engine performance.
Impala is important because it is a means to place SQL analytics on the Hadoop platform, thus opening up the mysterious world of Hadoop to the more familiar ways of ubiquitous SQL. Vaughan says all benchmarks rightly have their detractors. But, he suggests, it is interesting to see early versions of Impala performing on par with or better than established, SQL-only analytic alternatives.
The writers also discuss the resurgence of SQL in the face of very vibrant Hadoop and NoSQL activity. Some of that SQL upswing is due to Hadoop advances, such as the aforementioned Impala. Others take the form of NewSQL databases that seem well-tuned for cloud-based data management efforts, according to Vaughan.
Listen to this four-minute podcast and read more on these topics in Cloudera Impala bests DBMS-Y analytical engine in bake-off. During the podcast, you will:
- Learn about Impala benchmarks,
- Find a discussion of benchmarks' drawbacks, and
- Discover that SQL is reappearing in NoSQL settings.
This was first published in February 2014