Course contentsShow
← System Design
Lesson 433 of 1,91912. NoSQL: Wide-Column StoresFree lesson

What is HBase?

HBase is an open-source, distributed, column-oriented store modeled after Google's BigTable, running on HDFS.

What is HBase?

You'll learn: What HBase is and how it fits into the wide-column store ecosystem as a distributed, column-oriented database.

The Big Picture

HBase is an open-source, distributed, column-oriented NoSQL database modeled after Google's BigTable. It runs on top of the Hadoop Distributed File System (HDFS), making it a natural fit for storing massive amounts of sparse data in Hadoop environments.

Think of HBase as a massive spreadsheet where rows can have billions of columns, but most cells are empty. Unlike traditional databases that waste space on null values, HBase only stores the columns that actually contain data—making it extremely efficient for sparse datasets.

Key Characteristics

HBase inherits the wide-column store principles you've already learned: column families, row keys for clustering, and LSM tree-based storage. What makes HBase unique is its tight integration with the Hadoop ecosystem—it stores data in HDFS, leveraging Hadoop's durability and fault tolerance.

Like Cassandra, HBase uses a write-ahead log (called HLog) and memstores that flush to immutable files (called HFiles, similar to SSTables). It also employs compaction strategies and Bloom filters for efficient reads.

However, HBase differs from Cassandra in one major way: it follows a master-slave architecture with a central coordinator (HMaster) and region servers that handle partitions of data. This gives HBase strong consistency guarantees within a single row, unlike Cassandra's eventual consistency default.

When to Use HBase

HBase excels when you need to store and scan massive volumes of time-series data, log files, or sensor readings in a Hadoop environment—especially when your data is sparse and you need random read/write access.

Key Takeaway: HBase is a BigTable-inspired, wide-column store built on HDFS that provides strong consistency and efficient sparse data storage for Hadoop-based big data applications.