Relationships Between Hive and Other Components
HDFS
Hive is a sub-project of Apache Hadoop, which uses HDFS as the file storage system. It parses and processes structured data with highly reliable underlying storage supported by HDFS. All data files in the Hive database are stored in HDFS, and all data operations on Hive are also performed using HDFS APIs.
MapReduce
Hive data computing depends on MapReduce. MapReduce is also a sub-project of Apache Hadoop and is a parallel computing framework based on HDFS. During data analysis, Hive parses HQL statements submitted by users into MapReduce tasks and submits the tasks for MapReduce to execute.
Tez
Tez, an open-source project of Apache, is a distributed computing framework that supports directed acyclic graphs (DAGs). When Hive uses the Tez engine to analyze data, it parses HQL statements submitted by users into Tez tasks and submits the tasks to Tez for execution.
DBService
MetaStore (metadata service) of Hive processes the structure and attribute information of Hive metadata, such as Hive databases, tables, and partitions. The information needs to be stored in a relational database and is managed and processed by MetaStore. In the product, the metadata of Hive is stored and maintained by the DBService component, and the metadata service is provided by the Metadata component.
Spark
Spark can be used as the execution engine of Hive. Hive SQL statements delivered by the client are processed at the logical layer on Hive, and physical execution plans are generated and converted into a directed acyclic graph (DAG) of a resilient distributed dataset (RDD), and then submitted to a Spark cluster as a task. This way, Hive query efficiency is improved thanks to the distributed memory computing capability of Spark.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.
