In-memory databases are clearly the next big thing in enterprise IT. The challenge is going to be figuring out where and when to use what type of in-memory database.
Looking to become one of the major players in this space, EXASOL AG has released version 5 of its namesake database that, in addition to being much faster, now includes a “Preference Analytics” capability and support for analytics applications written in Python, Luna and R in addition to Java.
Despite all the hype surrounding in-memory databases, Graham Mossman, a senior solution engineer for EXASOL, says in-memory databases, for the foreseeable future at least, are going to be used primarily for analytics applications that need to correlate data from multiple types of application environments. As such, the EXASOL database is specifically optimized for analytics rather than transaction processing applications by combining columnar storage, compression and massively parallel processing capabilities in a single database platform.
New features in EXASOL 5 include an in-database analytic tool called Skyline to enable "Preference Analytics" that help users remove subjective elements from the analysis process in a way that is intended to increase the confidence in the results generated by queries, says Mossman.
Mossman adds that EXASOL has also significantly improved the caching and index creation techniques it uses to make sure that the EXASOL database remains the single fastest in-memory database for analytics applications.
While in-memory computing will certainly transform enterprise IT in the years ahead, it’s unlikely that most organizations are going to adopt a single approach. Instead, there will more than likely be a multitude of in-memory databases in use for different classes of applications across the enterprise. The next big challenge, of course, will be finding a way to actually manage both new and old database architectures that are going to continue to be used side by side for years to come.