netstrings

Netstrings are a simple encoding for transmitting strings over a network. Each netstring consists of a length prefix, a colon, the data, and a trailing comma. The length prefix is a base-10 integer indicating the number of bytes in the data. Netstrings are self-delimiting and unambiguous when read from a byte stream.

Format: <length>:<data>, for example 4:spam, encodes the ASCII string spam (4 bytes). 0:, encodes the empty string. Data may contain any binary bytes; the length prefix and trailing comma delimit the string regardless of its content.

Parsing and streaming: Parsing proceeds by reading digits until a colon to obtain length, then reading exactly that many bytes, and finally consuming the trailing comma. If the input arrives in chunks, the parser accumulates data until a complete netstring can be read. Multiple netstrings can be sent in sequence without additional separators.

Usage and advantages: Netstrings are simple to implement in most languages and require no escaping. They provide clear boundaries between messages, making them a convenient framing mechanism for streaming protocols and lightweight RPC where messages are discrete strings or binary blobs.

Limitations: The length prefix is encoded in decimal, so there is an overhead that grows with message size, particularly for very small messages. They are primarily a framing format for strings or binary blocks and are less suited to complex, nested data without a separate encoding layer.

History: Netstrings were introduced by Daniel J. Bernstein in the early 2000s as a compact, robust data-framing method for streaming protocols. The concept has appeared in various educational and practical examples since then.