What dependency does Apache Spark have on Hadoop?

[Monica Shiralkar]

Does Apache Spark has any dependency on Hadoop.If so What dependency does Apache Spark have on Hadoop?

thanks

[Theo van Kraay]

Absolutely none. The fact that Spark is able to run over Hadoop was done merely for backwards compatibility as Hadoop was the 'big thing' in distributed processing at the time. In 'standalone mode', Spark can actually run over any distributed file system... it doesn't have to be Hadoop.

[Monica Shiralkar]

Thanks

Theo van Kraay wrote:Spark can actually run over any distributed file system

For what reason does it always require a file system. I can understand it is required for reading data from file system but in case of streaming, why would the file system surely be required?