What is ACID?

ACID is an acronym for Atomicity, Consistency, Isolation, and Durability. These four properties are a set of guarantees that ensure database transactions are processed reliably. Atomicity means that a transaction is an all-or-nothing operation: either all of its changes are applied, or none are. Consistency ensures that a transaction transforms the database from one valid state to another, preserving all defined rules and constraints. Isolation makes concurrent transactions appear to execute serially, as if no other transactions are running, preventing data anomalies like dirty reads. Durability guarantees that once a transaction is committed, its changes survive system crashes or power losses, typically through write-ahead logging or persistent storage.

In practice, database systems enforce these properties through mechanisms like transaction logs, locking, and timestamps. For example, before any changes are written to the main data, they are first recorded in a transaction log. If a crash occurs before commit, the system can undo the partial changes (rollback). If the commit completes, the log ensures the changes can be reapplied after a restart. Isolation is typically implemented via pessimistic locking or optimistic concurrency control, with different isolation levels (such as Serializable, Repeatable Read, Read Committed) offering trade-offs between consistency and performance.

ACID is foundational to relational database management systems (RDBMS) like PostgreSQL, MySQL (with InnoDB), and Oracle. In distributed systems, achieving ACID properties across multiple nodes is more challenging and often requires expensive coordination protocols like two-phase commit (2PC). This has led to alternative models, such as BASE (Basically Available, Soft state, Eventual consistency), which relax consistency for better availability and scalability, as seen in many NoSQL databases.