4 ms·

How do these customers plan on doing historical analysis of raw data without Hadoop?

• agibsonccc 10y ago
  Spark is still "hdfs". These guys just use spark sql instead of map reduce jobs. SPark is heavily used for batch workloads.