Tableau adds in-memory data engine Hyper to Tableau 10.5, launches Tableau Server for Linux

Tableau is adding in-memory technology to its upcoming 10.5 release to speed up query times 5x and ingest large data sets faster.
Written by Larry Dignan, Contributor

Video: CEO Adam Selipsky outlines Tableau 10.5's new data engine Hyper

Tableau said its in-memory data engine, called Hyper, is generally available and will be included in Tableau 10.5. Hyper will be able to boost query speed by 5X and extract data and large data sets faster.

With the move, Tableau gets into the database game. Typically, Tableau is extracting data from multiple data sets and joining them together.

What Tableau is hoping to do is speed up time to insight and visualization. Tableau is also releasing Tableau Server on Linux and the ability to embed multiple visualizations in a single view with Viz in Tooltip.

In 2016, Tableau acquired HyPer, a technology that grew out of a Technical University of Munich project, to speed up its analytics tools. Now, the technology will be released in the Tableau 10.5 upgrade to provide the ability to handle larger data sets faster.

Read also: Tableau unveils high-scale Hyper engine, previews self-service data-prep and 'smart' capabilities | Tableau extends its footprint | Tableau details its natural language query plans | Tableau is the data visualization program you need to learn (ZDNet Academy) | Tableau acquires startup ClearGraph for natural language processing

Tableau CEO Adam Selipsky said the company integrated Hyper in about 18 months for Tableau's desktop version. He added that by making data ingestion faster it'll be easier to get insights across a broader spectrum of users.

"Hyper changes the way customers work with data. A million row extract goes from 5 minutes to 10 seconds. Multiply that on all scenarios and you can refresh data without thinking about it," said Selipsky.

In a nutshell, Hyper enables Tableau to work with billions of rows of data. That amount of data crunching means that a company can collapse something like 30 workbooks across many units to one.

Selipsky said Tableau is still the Switzerland of data with multiple connectors to various databases. However, there's the possibility that enterprises will do more database work inside of Tableau instead of other tools.

The key points about Hyper include:

  • Hyper is an in-memory data engine designed to add new fields to visualizations, apply filters faster, and deliver complex dashboards within Tableau.
  • The database leverages multi-core processing and uses parallelization techniques to optimize queries.
  • Tableau may be able to entice customers to leverage Tableau's Data Engine and move traditional database workloads.
  • Hyper has had a seven-month, pre-release program and nightly performance testing with 62,000 workbooks.
  • Customers won't have to migrate data to use Hyper in Tableau 10.5.

As for Tableau 10.5 and Tableau Server on Linux, the company said it will be able to cut support costs. In a nutshell, customers won't have to maintain Windows and Linux environments for Tableau. Tableau Server on Linux can also be a public cloud option.

Selipsky said Tableau has been asked about a version for Linux, which can be more cost effective in cloud deployments.

Linux distributions supported include CentOS, Ubuntu, Red Hat Enterprise Linux, and Oracle Linux.

The update for Viz in Tooltip will allow customers to embed contextual visualizations inside the tooltip without code. Tableau customers can also streamline dashboards with more interactive tools.

Previous and related coverage

Hadoop 3 confronts the realities of storage growth

While Hadoop was designed around commodity infrastructure, Hadoop 3.x confronts the reality that too much cheap storage can get expensive.

Amazon, Salesforce would love to move from Oracle databases, muzzle Ellison, but can they succeed?

It must be time for those Oracle database negotiations to start. A report highlights how Amazon and Salesforce are aiming to run their businesses on their own databases.

Editorial standards