Hadoop’s Legacy: No more fear of data

A couple weeks back, the Apache community pulled the plug on a slew of dormant Hadoop projects. Comrade-in-arms Andrew Brust gave the rundown in his ZDnet post. The tagline of Hadoop being dead is hardly new; it's dated back at least 5 years when Gartner announced a drastic slowdown in new Hadoop projects.

There's no question that activity for implementing Hadoop in the raw has practically ceased. That doesn't mean that the underlying technology is dead: Surviving projects continue to live on in the Cloudera Data Platform, which unlike previous Hadoop generations, is a coherent platform packaged through its own binaries rather than a string of disparate zoo animal open source projects.

The more important point, however, is that Hadoop opened the world of what we used to call "Big Data." Before Hadoop, and the associated discoveries with file systems that could live on commodity hardware and almost perfectly linear scale-out processing that Google published as research papers, the notion of going outside the data warehouse to bring non-relational data -- and lots of it (in the terabyte and petabyte) range was practically unthinkable. Hadoop introduced us to the art of the possible, and given that today, there are a variety of paths for analyzing and building machine learning models on all that data, and that relational data is no longer the limitation, are all thanks to Hadoop.

For our take on how this happened and what it made possible, check out our detailed post on ZDnet. And stay tuned, from all our research over the past decade, we'll be putting together a timeline on exactly how this all unfolded.