From the experience of a naive developer!
December 01, 2015

From the experience of a naive developer!

Anonymous | TrustRadius Reviewer
Score 9 out of 10
Vetted Review
Verified User

Modules Used

  • Hadoop Distributed File System

Overall Satisfaction with Hadoop

I used Hadoop for my academic projects for processing high volume data of my data mining project.
  • It was able to map our data with clear distinction based on the key.
  • We were able to write simple map reduce code which ran simultaneously on multiple nodes.
  • The auto heal system was really helpful in case of multiple failures.
  • I think Hadoop should not have single point of failure in terms of name node.
  • It should have good public facing API's for easy integration.
  • Internals of Hadoop are very abstract.
  • Protoco Buffers is a really good concept but I am not sure if we have checked other options as well.
  • Nope
Processing of big data has been the ultimate need for the me choosing Hadoop. Big data is massive and messy, and it’s coming at you uncontrolled. Data are gathered to be analyzed to discover patterns and correlations that could not be initially apparent, but might be useful in making business decisions in an organization. These data are often personal data, which are useful from a marketing viewpoint to understand the desires and demands of potential customers and in analyzing and predicting their buying tendencies.

I think Hadoop processes it very efficiently.
I think Hadoop has multiple flavors which people can customize to use as per their requirement. But I would choose hadoop based on following factors:
1. Number of nodes decision based on parallelism we want.
2. The module we want to run should be able to run parallely on all machine.