What is the future of Hadoop?

Until about two or three years ago Hadoop and Big Data were almost synonyms. In order to store all your data, structured, semi structured and in a very few cases truly unstructured, a Hadoop cluster was inevitable. 

While the core of Hadoop is very good in storing data cheaply, as it was designed for, doing analytics on the data was a whole different ball game. All kinds of Apache projects were started and start-ups were founded to make analytics in Hadoop easier and more performant. But using a tool for a purpose it is not designed for will in the end never be completely successful.

This realization has hit the market as well, the merger of Cloudera and HortonWorks and the take over from MapR by HPE were not made out of luxury. 

But does this mean your investments in Hadoop were a waste? Or is there a way you can still use Hadoop as it was intended, and use a tool that is from the start designed to analyze data in a fast and efficient way? Or would you like to do even more and use data that is not in Hadoop but that we need to optimize the analysis? And of course, preferably do all of this in a language you already are familiar with, like ANSI SQL.

Yes, there is a way where you can use a database that leverages your Hadoop investment, but gives you all the benefits of a MPP, true columnar, shared nothing design. Using ANSI SQL for all the interactions with the data.

With Vertica you can do all of this and more. Just imagine to be able to access data that is stored in Hadoop directly from you front end BI tool. Using in-database machine learning algorithms on data that is in Hadoop combined with data that is not in Hadoop.

With Vertica you can use Hadoop as it was intended, storing the data in an efficient and cheap way. And use Vertica, a purpose build analytical database, doing all queries, advanced analytics and machine learning.

#Hadoop #SQL #Vertica #BigData