YugaByte DB is a transactional database management system that can scale up and down across multiple regions for planet-scale and geo-distributed applications. According to the [CAP theorem][cap], YugaByte DB is consistent and partition tolerant. Combining SQL and NoSQL in one platform, it supports distributed ACID transactions, auto-sharding, and auto-balancing. Besides PostgreSQL-compatible SQL API, it provides another two APIs extended from Redis commands and Apache Cassandra Query Language (CQL), respectively. Built on a customized version of RocksDB, YugaByte DB's storage engine, DocDB, is a log-structured merge-tree (LSM) based "key to object/document" store. [cap]: https://en.wikipedia.org/wiki/CAP_theorem
YugaByte DB's first public beta release came out in November 2017. It was initially developed by the former team that built and ran Facebook's NoSQL platform that supported a number of Facebook's real-time applications. They left Facebook and found their own company, YugaByte Inc, aiming to build a database management system to unify the data layer for these mission-critical applications. Companies with lots of experts are able to offer complex DBaaS platforms which hide internal details of the data layer and benefit their app developers. However, for traditional enterprises and small startups, the data layers are mostly coupled within the applications. This is where YugaByte DB targets to rescue.
Relying on RocksDB, YugaByte DB's storage engine is responsible for converting every supported data formats (i.e., documents, CQL rows, and Redis data) to key-value pairs and store them in RocksDB. How data compression is accomplished in YugaByte DB depends on how it is done in RocksDB, which uses Dictionary Compression.
YugaByte DB uses MVCC and a variant of OCC for concurrency control. Under a distributed environment, it uses Two-Phase Commit with Early Acknowledgement. When a transaction wants to modify a number of rows, it first writes "provisional" records of each modified row into the target tablet storing the row. These records cannot be seen by the client unless the transaction commits. If conflicts occur when writing these records, the transaction will restart and abort. Otherwise, the transaction commits and notifies success to client. After that, the "provisional" records are applied and cleaned asynchronously.
YugaByte DB's storage engine, DocDB, is based on RocksDB. Unlike RocksDB, DocDB is a "key to object/document" store instead of a "key to value" store. Values in DocDB can be primitive types as well as object types (e.g., lists, sorted sets, and sorted maps) with arbitrary nesting.
YugaByte DB does not support foreign keys as none of its `CREATE TABLE` syntax offers keywords to set foreign key constraints.
Currently, YugaByte DB only supports Snapshot Isolation and is still working on supporting Serializable Isolation.
Only YugaByte DB's PostgreSQL-compatible API YSQL supports join operations (i.e., inner, outer, left, and right join). The other two APIs, YEDIS and YCQL, do not support this operation. Currently YSQL is still in beta and is based on part of the open source PostgreSQL 10.4 codebase, so the join algorithms it supports are the same as Postgres, namely Nested Loop Join, Hash Join, and Sort-Merge Join.
YugaByte DB uses the [Raft distributed consensus algorithm][raft] for replication, so all the changes to the database will be recorded in Raft logs, which can be used during recovery. [raft]: https://raft.github.io/raft.pdf
YugaByte DB offers the following three query APIs: - **YCQL:** Cassandra-compatible API with extended features including distributed ACID transactions, strongly consistent secondary indexes, and a native JSON data type. - **YEDIS:** Redis-compatible API that supports Redis commands and data types, with a new native Time Series data type provided. - **YSQL (beta):** PostgreSQL-compatible API.
YugaByte DB is a disk-oriented database management system. However, as its storage engine is implemented as a log-structured merge-tree (LSM), some of the data will be stored in memory before flushed out to disk.
YugaByte DB's storage model depends on RocksDB, which uses Static Sorted Table (SST) format.
YugaByte DB's storage engine relies on RocksDB, which is implemented as a log-structured merge-tree (LSM).
YugaByte DB uses shared-nothing system architecture. A table will be split into multiple tablets. Depending on the replication factor, each tablet has its corresponding number of replicas (tablet peers) across different nodes.
YugaByte DB supports non-materialized views in its YSQL API.
Commercial, Open Source
Linux, OS X