Ah - that's a good point. Usually when people say "Hadoop", I assume they're referring to HDFS, but there is (was...) the Hadoop MR system that ran on top of it that has been almost entirely replaced by Spark.