Fast FIX Parser (FFP) is a library for parsing Financial Information eXchange (FIX) messages. It takes input bytes as they arrive from, for example, a socket, and converts them into a representation of FIX messages which can be further analysed for semantic checks, converted into “business” structures, etc. It also provides a way to specify which tags are allowed for a particular message and verifies this specification at runtime.
- Speed. On my rather old Core i5-430M 2.26GHz laptop, in a single thread, this parser can process about 260,000 messages with groups per second and about 520,000 simple messages per second. The processing time is more or less a linear function of the message length.
- It does not impose any particular I/O or threading model. In fact, it does no I/O at all, and there are no threads running in the background. This greatly simplifies integration of the library into an existing code base.
- The parser does not expect every chunk of its input data to be a complete FIX message. The input bytes can be fed into the parser as they become available, and the parser splits or combines the input into complete messages.