Differences and scenarios for Hadoop, Hive, HBase

It's inevitable that Hadoop, Hive, HBase will inevitably be the only thing that has started learning big data recently.

Here is a record of your own understanding of these 3:

1, Hadoop: It is a distributed computing + Distributed File system, the former is actually MapReduce, the latter is HDFs. The latter can be operated independently, the former can be used selectively, or it can be used without

2, Hive: Popular is a data warehouse, the data in the warehouse is managed by HDFs data file, it supports similar functions of SQL statement, you can use this statement to complete the Distributed Environment computing function, hive will convert the statement to MapReduce, and then to Hadoop execution. The calculations here are limited to lookup and analysis, not updates, additions, and deletions. Its advantage is the processing of historical data, using the popular theory is off-line calculation, because its bottom is mapreduce,mapreduce in real-time computing performance is poor. It does this by loading the data file in as a hive table (or external table) so that you feel that your SQL operation is a traditional table.

3, HBase: In layman's words, hbase functions like a database, traditional database management is centralized local data files, and HBase is based on HDFS implementation of distributed data file management, such as adding and deleting. In other words, hbase is just a persistent file (hfile) that leverages Hadoop's HDFs to help it manage data, and it has nothing to do with MapReduce. HBase's advantage lies in real-time computing, where all real-time data is directly stored in HBase, and the client accesses hbase directly through the API for real-time computing. Because it uses NoSQL, or a column structure, it improves the search performance and makes it possible to use the big data scenario, which is the difference between it and MapReduce.

Summarize:
Hadoop is the basis for hive and hbase, and Hive relies on Hadoop, and HBase relies only on Hadoop's HDFS modules.
Hive is suitable for analysis of offline data, operates in a common format (such as a common log file), data files managed by Hadoop, supports class-SQL, is more convenient than writing MapReduce Java code, is positioned as a data warehouse, stores and analyzes historical data
HBase is suitable for real-time computing, using a column-structured NoSQL, which operates on its own generated special-format hfile, Hadoop-managed data files, which are positioned as databases or DBMS

Finally, hive can manipulate the files in HDFs directly as data from its tables, or it can use the HBase database as its table