Spark Hadoop Connections and Differences

1. Spark and Hadoop are computing frameworks

2. Hadoop consists of 3 parts: MapReduce, Hdfs, Yarn. Among them, MapReduce has gradually withdrawn from the historical stage and is replaced by Spark. So saying that Hadoop is dead is not true

2.1. The bottom layer of Spark must also rely on hdfs, because Spark only cares about computing, regardless of storage. In fact, it seems that the bottom layer of Hbase also relies on hdfs