MemSQL Introduces dbbench for Database Benchmarking

Date 2016/1/26 9:07:46 | Topic: Product News

MemSQL has introduced a new open source benchmarking tool that generates continuous loads defined by a flexible configuration. Unlike many other tools that run for a fixed period and then complete, dbbench maintains an ongoing specified workload upon which additional performance tests and measurements can be conducted, so users can find bottlenecks quickly through active benchmarking. This approach fits today's modern workloads that combine continuous data ingest with concurrent queries.
"As we pursue real-time solutions with our customers, we have seen the need to benchmark use cases with continuous load and stress testing," said Ankur Goyal, VP of Engineering, MemSQL. "dbbench allows us to model these intensive environments and ensure MemSQL achieves the highest performance possible on any given deployment platform."

"While there are many great database benchmarking tools out there, we found that modern engineering requires a more robust, active benchmarking approach to match the flow of real-time data," added Alex Reece, performance engineer, MemSQL, and the creator of dbbench. "dbbench runs continuously, allowing developers to ask difficult questions about how their system performs under heavy load and identify bottlenecks, a necessity for today's complex workloads that combine transactions and analytics."

The Internet of Things (IoT) is a perfect example where companies need to capture an ongoing stream of data and deliver insights and action based on querying that data. As IoT expands, greater volumes of data from disparate sources engender more questions about the meaning and value of that data. Performance engineers can use a question-driven methodology with dbbench to investigate these large, continuously running workloads. In just a few simple lines of configuration code, engineers can define and run a complicated workload with a wide range of concurrent reads and writes. This flexibility enables quick iteration and improvisation of new workloads for deeper analysis. Overall, dbbench reflects a shift in performance benchmarking towards a more active approach.

With dbbench, MemSQL continues its efforts in making other tools available as open source such as MemSQL Streamliner, an integrated Apache Spark solution for real-time streaming, and MemSQL Loader for natively importing data from HDFS and S3.

To download dbbench, visit https://github.com/memsql/dbbench



This article comes from Software Development Tools
http://www.softdevtools.com

The URL for this story is:
http://www.softdevtools.com/modules/news/article.php?storyid=3888