sequencenumber

A sequence number is an identifier assigned to elements of a sequence to preserve order, detect missing items, and help reconcile data after communication or storage.

In computer networks, sequence numbers are used by transport protocols to order bytes within a stream. The most well-known example is TCP, which uses a 32-bit sequence number for bytes and an acknowledgment number to confirm receipt. Sequencing enables reassembly of data segments and supports flow control and loss recovery.

In databases and software, sequence numbers are often implemented as auto-incrementing identifiers (identity columns). They provide unique keys for records and can also serve as monotonically increasing counters for events or transactions. Depending on design, sequences may be global or scoped to a table or shard and may cache values to improve performance.

In messaging and event processing, sequence numbers or offsets track progress through a log or queue. They help detect out-of-order delivery, enable idempotent processing, and support replay or rewind of streams. Systems such as distributed log services often expose sequence numbers as part of their APIs.

Common considerations include wraparound of finite bit-width sequence spaces, handling of missing or duplicate items, and maintaining monotonicity in the presence of replication, restarts, or clock changes. Correct use often requires application-level logic to validate sequences and implement idempotent operations.

See also: TCP sequence number, offset, autoincrement, identity column, event sourcing.