2016 Data Analytics Forecast: Top 5 Trends to Watch
At the Strata + Hadoop conference today, MemSQL released a 5.0 upgrade to its namesake in-memory database that advances the goal of integrating transactions and real-time analytics.
Gary Orenstein, chief marketing officer for MemSQL, says MemSQL 5.0 makes it easier to accomplish that goal by invoking a Hybrid Transaction/Analytical Processing (HTAP) model that provides concurrent support for OLTP and OLAP queries. In addition, with this release, MemSQL is adding support for a new code generation technique based on low-level virtual machines (LLVM) that makes it possible to compile queries in a way that allows end users to interactively explore data sets running in memory.
With this release, MemSQL is also adding one-click integration with the Apache Spark in-memory computing framework along with support for pluggable authentication modules based on, for example, Kerberos.
In general, Orenstein says, two macro trends are coming together to transform the way databases are deployed and managed in the enterprise. The first is the rise of multi-modal databases such as MemSQL that can support, for example, both rows used in transaction applications and columns used most often in analytics applications. The second is the shift toward real-time analytics, which is starting to sharply reduce the need to run analytics as a batch application running on a separate data warehouse. The end result from a database administration perspective is a much less complex environment that enables applications to run much faster by running in memory.
Naturally, it may take IT organizations some time to appreciate the implications of in-memory databases that can simultaneously support multiple types of applications. But as business leaders become more accustomed to being able to interact with data in real time, it’s now only a question of when IT organizations are going to adopt these modern database frameworks.
