The Serial Line Internet Protocol (SLIP)[1] [2] is an encapsulation of the Internet Protocol designed to work over serial ports and router connections. It is documented in . On personal computers, SLIP has largely been replaced by the Point-to-Point Protocol (PPP), which is better engineered, has more features, and does not require its IP address configuration to be set before it is established. On microcontrollers, however, SLIP is still the preferred way of encapsulating IP packets, due to its very small overhead.
Some people refer to the successful and widely used RFC 1055 Serial Line Internet Protocol as "Rick Adams' SLIP",[3] to avoid confusion with other proposed protocols named "SLIP". Those other protocols include the much more complicated appendix D Serial Line Interface Protocol.
| Hex value | Dec Value | Oct Value | Abbreviation | Description | |
|---|---|---|---|---|---|
| 0xC0 | 192 | 300 | END | Frame End | |
| 0xDB | 219 | 333 | ESC | Frame Escape | |
| 0xDC | 220 | 334 | ESC_END | Transposed Frame End | |
| 0xDD | 221 | 335 | ESC_ESC | Transposed Frame Escape |
SLIP modifies a standard TCP/IP datagram by:
SLIP requires a serial port configuration of 8 data bits, no parity, and either EIA hardware flow control, or CLOCAL mode (3-wire null-modem) UART operation settings.
SLIP does not provide error detection, being reliant on upper layer protocols for this. Therefore, SLIP on its own is not satisfactory over a link which is error-prone, such as a poor quality dial-up connection.
SLIP escape characters were also required on some modem connections to escape Hayes command set, allowing therefore to pass binary data through those modems that would recognize some characters as commands.
A version of SLIP with header compression is called Compressed SLIP (CSLIP).[4] The compression algorithm used in CSLIP is known as Van Jacobson TCP/IP Header Compression.[5] CSLIP has no effect on the data payload of a packet and is independent of any compression by the serial line modem used for transmission. It reduces the Transmission Control Protocol (TCP) header from twenty bytes to seven bytes. CSLIP has no effect on User Datagram Protocol (UDP) datagrams.
See also: History of the Internet. RFC 1055, a "non-standard" for SLIP, traces its origins to the 3COM UNET TCP/IP implementation from the 1980s. Rick Adams added SLIP to the popular 4.2BSD in 1984 and it "quickly caught on". By the time of the RFC (1988), it is described as "commonly used on dedicated serial links and sometimes for dialup purposes".[6]
The last version of FreeBSD to include "slattach" (a command for connecting to slip) in the manual database is FreeBSD 7.4, released 2011. The manual claims that auto-negotiation exists for CSLIP. The FreeBSD version is inherited from 4.3BSD.[7]
Linux formerly used the same code base for SLIP and KISS (TNC). The split occurred before the start of kernel git history (Linux-2.6.12-rc2, 2005).[8] The SLIP driver offers a special "6-bit" escaped mode to accommodate modems incapable of handling non-ASCII characters.[9] The Linux slattach command (written independently) also has the ability to auto-detect CSLIP support.[10]
"Requirements for an Internet Standard Point-to-Point Protocol"