Overall Satisfaction with Hadoop
My present company uses Hadoop and associated technology to create a data pipeline using open source tools. Apart from that we also consult for projects which could potentially use Hadoop. Apart from that, I also work as a consultant for HDP. We actively help in installation and setup of hadoop clusters.
- Hadoop is open source and with a wide community already present, the usage is much easy for individuals, startups and MNCs alike.
- Hadoop works well for commodity hardware and that makes it easier to avoid pricey clusters.
- Hadoop takes parallel programming to next level and helps processing of multi terabytes (even petabytes) of data easier.
- While Hadoop MR parallelizes jobs involving Big Data, it is slow for smaller data sets
- OLAP (analytics)is easier, however, OLTP (transactions) is a problem in most cases.
- People using Hadoop have to keep in mind that small proof of concepts may not scale as expected.