With access to larger amounts of memory on both processors and via Flash cards, interest is starting to rise in in-memory databases that make it easier to combine transaction processing and analytics within a single application.
Looking to become one of the primary drivers of that trend is MemSQL, which today released version 4.0 of its namesake database, which adds integration with the Apache Spark in-memory clusters. In addition, MemSQL is now offering a free community edition of the database to make it easier for IT organizations to get started developing applications for in-memory databases.
MemSQL CEO Eric Frenkiel says in-memory databases and Apache Spark are natural complements, because even though MemSQL is significantly faster than Apache Spark, organizations are using Spark to process large amounts of Hadoop data that can then be fed into MemSQL. The end result, says Frenkiel, is a new framework for processing massive amounts of Big Data in real time.
Other new MemSQL 4.0 features include support for geospatial capabilities, an enhanced optimizer for processing SQL queries and an updated disk-based column store coupled with an in-memory row store.
Frenkiel says the time is fast approaching when more IT organizations will be looking for a general-purpose, in-memory SQL database that is not optimized for only one particular class of applications.