When it comes to analytics in the enterprise, there is no shortage of options when it comes to management systems. There are, of course, traditional row-based relational databases, but now we're also seeing the rise of columnar databases and the open-source Hadoop data management framework.https://o1.qnsr.com/log/p.gif?;n=203;c=204663295;s=11915;x=7936;f=201904081034270;u=j;z=TIMESTAMP;a=20410779;e=i
This rising diversity in data management systems is creating new challenges for IT organizations that have traditionally relied on one SQL-based platform to manage that data. To specifically address this issue, Teradata today rolled out version 14 of its namesake database system, which now includes support for both row-based and columnar data formats, while at the same time tightening the integration between the Teradata database and the Aster Data appliance platform that Teradata acquired earlier this year.
According to Randy Lea, vice president of Teradata's Aster Center for Innovation, what makes the Teradata approach unique is that the company is creating a common management framework for row and columnar databases running on a Teradata machine and MapReduce deployments running on Aster Data appliances. Beyond lowering the cost of managing all these environments, this approach also makes it possible for IT organizations to leverage SQL against Hadoop environments versus requiring customers to learn how to master the MapReduce interface that was developed alongside Hadoop. As Lea notes, SQL is the language of business, so in order to effectively bring MapReduce into the enterprise, data in those systems needs to be accessible via SQL. What Teradata is doing, says Lea, is creating an analytics management ecosystem where all the data sources are logically, not just physically, integrated. To that end, Lea notes that customers can now even elect to use row and columnar formats side by side within the same application.
At the same time, there's no doubt that MapReduce represents a major step forward both in terms of the types of analytic applications that can be developed and the actual cost of managing large amounts of data. But like every other emerging technology, Hadoop needs become part of the rest of the enterprise fabric, which means the emphasis on data management is now shifting quickly towards frameworks that encompass multiple types of data management engines.