How Spark is complementing Hadoop

Spark is in memory analytics tool designed for complementing Map reduce jobs for low latency and iterative jobs.

Spark is next generation distributed processing engine for Bigdata .

Spark never been competition to Hadoop jobs in fact it will increase scope of Bigdata project for different type of job which are difficult to do in Hadoop

Spark became good tool for online near realtime analytics with straming support.

Spark is very fast and follows the top down approach for streaming and analytics.

Spark is independent of file systesm and can integrate with any Hadoop file system or no SQL DB or Cloud platform.

Spark has very good SQL tool called Spark Sql which can read and write to almost any DB.

Spark supports Machine learning with Mlib and supports R with SparkR.

Spark integrates with all Hadoop ecosystem tools very well.