Hadoop vs Spark

Spark POC : Hadoop vs Spark

SimilarWeb is doing POC with Spark with the goal to assess it as Hadoop MR replacement.During the POC dedicated cluster was deployed and several jobs where re-implemented in Sparkin order to compare their performance with existing Hadoop MR and HIve jobs.Similar Web kindly allowed me to disclose many details of the POC in this meetup.

In meetup I am going to present:

a) Short overview of Spark - how it works, just to be on the same page. It is advised to read some materials before the meetup.b) Explain what is a nature of jobs, and how they are implemented in Spark / Scalac) Benchmark results and our attempt to explain them.d) Conclusions and further questions.

Meetup will be very technical and focus on our understanding of Spark, Hadoop internals