0

If I use a single node setup with Hadoop where it has only one disk, I can easily use it to play around with Hadoop and stuff, but will there actually be any benefit to using Hadoop in this case over another paradigm (processing in Python without using mapreduce, etc.)

2 Answers 2

0

The power of Hadoop comes from the fact it uses a distributed file system (HDFS) and distributed programming model (MapReduce). Processing cannot be parallelized in a single node cluster with one disk, so - because of Hadoop's overhead - simpler paradigms will probably be faster.

Even when using a full-blown cluster, you won't see the benefit of Hadoop if the date set is too small or the algorithm is not suitable for distributed processing (e.g. finding a median value).

0

Hadoop is completely based on Distributed file system(HDFS), and the real power lies here.

But at the same time its important to learn the framework and understand its components. That is the reason Hadoop provides 3 modes of installation.

Answering to your exact question,single disk does not benefit from HDFS part of hadoop. But again it will help in understanding MapReduce Component of Hadoop.

If you have goof configuration system, I would suggest go for Pseudo Cluster setup,having two virtual machine interacting with reach other.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .