MemSQL announced version 6 of their flagship data warehousing platform at Strata Data Conference last week. MemSQL 6 features new extensibility capabilities to enable ML, large-scale improvements for analytical queries, and a more broad set of online operations. Developers will use this tool for ML functions with live data and SQL. They are also able to build operational applications with no need for disparate system-support via a real-time data warehouse.
MemSQL 6 includes the ability to run ML algorithms in a distributed SQL environment, enhancements to online operations, and increases to query performance to deliver up to “80 times improvement” from previous versions, according to the provider. For organizations wanting to bring machine learning functions closer to live data, version 6 supports real-time scoring with existing or new models, DOT_PRODUCT for image recognition, and new extensibility capabilities that enable functions, such as K-means in SQL.
The platform utilizes modern chip architectures including Single Instruction, Multiple Data for performance and can fire as high as processing a billion rows per CPU core per second. In addition, MemSQL now offers more efficient query isolation that is aimed at improving concurrency on large volumes of data and many users.
In a press statement, the company’s CEO and co-founder Nikita Shamgunov concluded: “As machine learning use expands within companies, a smarter database is needed to maximize learning. MemSQL 6 now has capabilities such as extensibility to enable more ML functions, so customers can meet the needs of their most ambitious challenges.”
A trial of MemSQL 6 is available now, with general availability coming in October.
Latest posts by Timothy King (see all)
- Modern Data Integration Strategies Must Evolve with Cloud Adoption - January 11, 2018
- Key Takeaways from Gartner’s Market Guide for Data Preparation - January 11, 2018
- Top 5 Questions to Ask When Evaluating Data Integration Tools - January 5, 2018