The other day I ran across an article on Infoworld.com: Cloudera pitches Hadoop for everything. Really?”
Of course, the article starts by mentioning the expression about hammers and nails. This is an old story and it appears that it is getting ready to repeat itself. Like it’s been said: “those who forget the past are doomed to repeat it”.
Hadoop has been the biggest star of the big data story. I have to say that it is revolutionizing data processing and for good reasons. Many seem to point to the use of cheap clusters based on commodity hardware. I personally prefer to attribute it to the large amount of data that has different requirements from traditional data processing.
The traditional data processing needs are still there and still growing. Getting rid of “silos” of data has proven extremely difficult. It also relies on getting rid of years of investments and re-writing many proven applications.
Instead of trying to fit everything into Hadoop, it is much better to have an overall strategy that takes into accounts the different needs of different data sets and make sure the overall architecture accommodates exchange of information between all of them.
Cloudera want to become the “enterprise data hub” powered by Hadoop. Like the article mentions, “Hadoop i still seen on all sides as a bucket of parts..”. Maybe it is a bit early to talk about an enterprise data hub based on Hadoop.
Of course, if all you have is a hammer, everything looks like nail