Things you have to know before using Hadoop
Hadoop is very accessible; this is one of the best things about Hadoop and it is what everybody has experienced. A lot of Hadoop users say they have downloaded it and they could use it in an hour and they could run MapReduce tasks in an hour too. But at the same time it is very difficult to tweak and customize it for specific purposes. And to that end talent in this area is also very rare and so people are looking forward to such a time when the ecosystem has evolved to such an extent that one doesn’t need talent or a stage where people don’t have to think about the algorithms, implications, and other low-level stuff.
With Hadoop, another core principle is modified the core data as little as possible. There are two reasons for it; basically you can store as much data as required since data storage is scalable. Hence, it is possible to keep a copy of the original data quite comfortably, without having to worry about the cost, and make any number of sub-copies with as many tweaks as necessary. And second, during this exploration, when modifying data, surely you will be removing something that you might want later.
In this arena agility is the key to success; the faster you can go, the faster you can adapt to the changing environment, the better. And finally, drive towards generalized components; there is quite a debate on this issue with Google keeping the underlying architecture a secret and FaceBook trying to reveal it, but it is best to keep to generalized components. Even if one has really fast and innovative architecture, it will only be advantageous for a short period of time in this space. This because technology is improving really fast and one might not be able to keep up with the ecosystem. However knowing about the tools available and how they would impact one’s business is a lasting strategic advantage. Hence when thinking about this, use as much as existing product as you can so that you can focus your own resources on the things that are important to your business.