Weekly Update

Tableau goes “Elastic”…and so do database models

Apple may have used yesterday to make announcements about Apple Watch and its new iPhone 6 models, but data discovery darling Tableau used yesterday to announce its own iPad initiative. As covered rather comprehensively by Gigaom’s Derrick Harris, Tableau will be releasing an iPad-specific data visualization app sometime in 2015.

The project, named “Elastic” and announced at the annual Tableau Conference yesterday, is about creating an app conceived for the iPad from the get-go and aiming it at small business – or even individuals.

Deja Viz
On the platform front, Elastic is reminiscent of Roambi which, though it now runs on the Android platform, was initially an iOS-only offering. Functionally, in aiming at the SME and consumer audience, Elastic seems to be muscling in a bit on territory occupied by Excel, of all applications.

But Elastic wants to be more than PivotTables on your tablet. Tableau’s thesis here is that data isn’t just big, it’s personal. Meanwhile, most tools for looking at data are business-oriented, so a “digital life” approach to data discovery is needed now.

Could this lead to a BYoDD (bring your own data discovery) phenomenon, where personal data apps make the round trip back to business settings? That probably wouldn’t be such a bad thing, if so.

Bridging the SQL-NoSQL chasm
This week’s news doesn’t end with Tableau. In fact, important announcements have piled up from Informatica, MapR and FoundationDB.

Let’s start with FoundatioDB’s announcement. The company, which markets an eponymous Key-Value store NoSQL database, has announced the general availability of its SQL Layer, previously available in Beta form, which implements a relational database on top of the Key-Value platform.

According to FoundationDB, SQL Layer offers an ANSI SQL-92-compliant interface, fully-schematized tables and true ACID transactions. But, unlike NewSQL products, the underlying key-value store is still there and available. FoundationDB identifies this chameleon-like capability as support for “multi-model databases.” But maybe we should identify the category as “OmniSQL.”

The NoSQL engine that could
Meanwhile, MapR announced yesterday that it has set new records in data ingestion rates for the open source database OpenTSDB, running on MapR’s Hadoop distribution and MapR-DB, its Apache HBase-compatible NoSQL database. In fact, MapR says that using only 4 nodes of a 10-node MapR cluster, it was able to push OpenTSDB to an ingestion rate of over 100 million “data points” per second.

MapR’s point? NoSQL databases can work well for write-intensive operational applications, not just read-heavy workloads.

Informatica gets Big Data-friendly
Finally, there’s an announcement from Enterprise Information Management juggernaut Informatica, a vendor often associated with moving data between relational databases (RDBMSes) and data warehouse BI platforms. But contrary to its RDBMS typecast, Informatica offers a Big Data Edition of its platform. And now it’s working to get it into more Hadoop-enthusiastic hands.

Informatica is today announcing that it will offer 60-day free eval versions of its Big Data Edition, installed by default in both Cloudera’s and Hortonworks’ pre-built virtual machine images (the Cloudera QuickStart VM and the Hortonworks Sandbox). And for those ready to buy, Informatica will now offer subscription-based pricing, which is more harmonious with Big Data world procurement practices than the Enterprise licensing more typical in RDBMS circles.

Much as Tableau’s forays into the personal world may circle back to business, enterprise and NoSQL database vendors are venturing into each other’s worlds, and bringing once-exotic features home. As each type of database gains capabilities of the other, and as data management tools increasingly integrate different database types, the distinctions are blurring and becoming less important.

All of this is really another facet of technology’s consumerization. Now we just need a version of HBase that will run on Apple Watch, with a data viz app to match.