What Is the Relationship Between Hive and Other Components?¶
Hive and HDFS
Hive is an Apache Hadoop project. Hive uses Hadoop Distributed File System (HDFS) as its file storage system. Hive parses and processes structured data stored on HDFS. All data files in the Hive database are stored in HDFS, and all data operations on Hive are also performed using HDFS APIs.
Hive and MapReduce
All data computing of Hive depends on MapReduce. MapReduce, also an Apache Hadoop project, is a parallel computing framework based on HDFS. During data analysis, Hive parses HiveQL statements submitted by users into MapReduce tasks and submits the tasks for MapReduce to execute.
Hive and DBService
MetaStore (metadata service) of Hive processes the structure and attribute information about Hive databases, tables, and partitions that are stored in a relational database. In MRS, the relational database is maintained by DBService.
Hive and Spark
Hive data computing can also be implemented on Spark. Spark, also an Apache project, is an in-memory distributed computing framework. During data analysis, Hive parses HiveQL statements submitted by users into Spark tasks and submits the tasks for Spark to execute.