Comment by tlipcon

Just a small note: Kudu shares no code or provenance with Hadoop. Only commonality is ecosystem (eg you can use spark or Impala to query it)