# Intro
- No-SQL database that can support billions of rows and millions of columns
- Part of Hadoop ecosystem, typically built on top of HDFS
- Supports real-time read/write random access and scales to very large datasets across many machines (serves Facebook's messaging platform)
- Column-oriented, data is stored by columns physically
Row key
uniquely locates a row, like anindex
, only one index per table. Row keys are alphabetically sorted.11
comes before2
← Spark