Win a copy of Mesos in Action this week in the Cloud/Virtualizaton forum!
  • Post Reply
  • Bookmark Topic Watch Topic
  • New Topic

Hadoop in Practice:Is the way to program different from the traditional Java applications?

 
Qunfeng Wang
Ranch Hand
Posts: 434
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
I'm fresh new to distributed systems. Is there any differences when programming with these systems? It might not be the library or framework we use. It's about patterns, the way to think?
 
Alex Holmes
Author
Greenhorn
Posts: 21
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
Hi,

Hadoop's computational tier, MapReduce, is definitely a different programming/processing paradigm from your traditional procedural/OO multi-threaded programming. It excels at processing and joining together large volumes of data in parallel. Its programming model, which is designed to minimize constructs which can lead to performance issues in distributed systems such as distributed locking, requires you to represent your code in map and reduce functions, allowing you to filter, project, aggregate and join your data together at very fast rates.

If you're new to Hadoop then I'd recommend reading "Hadoop in Action" which provides a good background on Hadoop, and gives you a leg-up on useful Hadoop use cases.

Thanks,
Alex
 
  • Post Reply
  • Bookmark Topic Watch Topic
  • New Topic