|
|
|
Network Working Group J. Oikarinen
Request for Comments: 1459 D. Reed
May 1993
Internet Relay Chat Protocol
Status of This Memo
This memo defines an Experimental Protocol for the Internet
community. Discussion and suggestions for improvement are requested.
Please refer to the current edition of the "IAB Official Protocol
Standards" for the standardization state and status of this protocol.
Distribution of this memo is unlimited.
Abstract
The IRC protocol was developed over the last 4 years since it was
first implemented as a means for users on a BBS to chat amongst
themselves. Now it supports a world-wide network of servers and
clients, and is stringing to cope with growth. Over the past 2 years,
the average number of users connected to the main IRC network has
grown by a factor of 10.
The IRC protocol is a text-based protocol, with the simplest client
being any socket program capable of connecting to the server.
|
|
The IRC (Internet Relay Chat) protocol has been designed over a
number of years for use with text based conferencing. This document
describes the current IRC protocol.
The IRC protocol has been developed on systems using the TCP/IP
network protocol, although there is no requirement that this remain
the only sphere in which it operates.
IRC itself is a teleconferencing system, which (through the use of
the client-server model) is well-suited to running on many machines
in a distributed fashion. A typical setup involves a single process
(the server) forming a central point for clients (or other servers)
to connect to, performing the required message delivery/multiplexing
and other functions.
|
|
The server forms the backbone of IRC, providing a point to which
clients may connect to to talk to each other, and a point for other
servers to connect to, forming an IRC network. The only network
configuration allowed for IRC servers is that of a spanning tree [see
Fig. 1] where each server acts as a central node for the rest of the
net it sees.
[ Server 15 ] [ Server 13 ] [ Server 14]
/ \ /
/ \ /
[ Server 11 ] ------ [ Server 1 ] [ Server 12]
/ \ /
/ \ /
[ Server 2 ] [ Server 3 ]
/ \ \
/ \ \
[ Server 4 ] [ Server 5 ] [ Server 6 ]
/ | \ /
/ | \ /
/ | \____ /
/ | \ /
[ Server 7 ] [ Server 8 ] [ Server 9 ] [ Server 10 ]
:
[ etc. ]
:
[ Fig. 1. Format of IRC server network ]
|
|
A client is anything connecting to a server that is not another
server. Each client is distinguished from other clients by a unique
nickname having a maximum length of nine (9) characters. See the
protocol grammar rules for what may and may not be used in a
nickname. In addition to the nickname, all servers must have the
following information about all clients: the real name of the host
that the client is running on, the username of the client on that
host, and the server to which the client is connected.
|
|
To allow a reasonable amount of order to be kept within the IRC
network, a special class of clients (operators) is allowed to perform
general maintenance functions on the network. Although the powers
granted to an operator can be considered as 'dangerous', they are
nonetheless required. Operators should be able to perform basic
network tasks such as disconnecting and reconnecting servers as
needed to prevent long-term use of bad network routing. In
recognition of this need, the protocol discussed herein provides for
operators only to be able to perform such functions. See sections
4.1.7 (SQUIT) and 4.3.5 (CONNECT).
A more controversial power of operators is the ability to remove a
user from the connected network by 'force', i.e. operators are able
to close the connection between any client and server. The
justification for this is delicate since its abuse is both
destructive and annoying. For further details on this type of
action, see section 4.6.1 (KILL).
|
|
A channel is a named group of one or more clients which will all
receive messages addressed to that channel. The channel is created
implicitly when the first client joins it, and the channel ceases to
exist when the last client leaves it. While channel exists, any
client can reference the channel using the name of the channel.
Channels names are strings (beginning with a '&' or '#' character) of
length up to 200 characters. Apart from the the requirement that the
first character being either '&' or '#'; the only restriction on a
channel name is that it may not contain any spaces (' '), a control G
(^G or ASCII 7), or a comma (',' which is used as a list item
separator by the protocol).
There are two types of channels allowed by this protocol. One is a
distributed channel which is known to all the servers that are
connected to the network. These channels are marked by the first
character being a only clients on the server where it exists may join
it. These are distinguished by a leading '&' character. On top of
these two types, there are the various channel modes available to
alter the characteristics of individual channels. See section 4.2.3
(MODE command) for more details on this.
To create a new channel or become part of an existing channel, a user
is required to JOIN the channel. If the channel doesn't exist prior
to joining, the channel is created and the creating user becomes a
channel operator. If the channel already exists, whether or not your
request to JOIN that channel is honoured depends on the current modes
of the channel. For example, if the channel is invite-only, (+i),
then you may only join if invited. As part of the protocol, a user
may be a part of several channels at once, but a limit of ten (10)
channels is recommended as being ample for both experienced and
novice users. See section 8.13 for more information on this.
If the IRC network becomes disjoint because of a split between two
servers, the channel on each side is only composed of those clients
which are connected to servers on the respective sides of the split,
possibly ceasing to exist on one side of the split. When the split
is healed, the connecting servers announce to each other who they
think is in each channel and the mode of that channel. If the
channel exists on both sides, the JOINs and MODEs are interpreted in
an inclusive manner so that both sides of the new connection will
agree about which clients are in the channel and what modes the
channel has.
|
|
The channel operator (also referred to as a "chop" or "chanop") on a
given channel is considered to 'own' that channel. In recognition of
this status, channel operators are endowed with certain powers which
enable them to keep control and some sort of sanity in their channel.
As an owner of a channel, a channel operator is not required to have
reasons for their actions, although if their actions are generally
antisocial or otherwise abusive, it might be reasonable to ask an IRC
operator to intervene, or for the usersjust leave and go elsewhere
and form their own channel.
The commands which may only be used by channel operators are:
KICK - Eject a client from the channel
MODE - Change the channel's mode
INVITE - Invite a client to an invite-only channel (mode +i)
TOPIC - Change the channel topic in a mode +t channel
A channel operator is identified by the '@' symbol next to their
nickname whenever it is associated with a channel (ie replies to the
NAMES, WHO and WHOIS commands).
|
|
The protocol as described herein is for use both with server to
server and client to server connections. There are, however, more
restrictions on client connections (which are considered to be
untrustworthy) than on server connections.
|
|
No specific character set is specified. The protocol is based on a a
set of codes which are composed of eight (8) bits, making up an
octet. Each message may be composed of any number of these octets;
however, some octet values are used for control codes which act as
message delimiters.
Regardless of being an 8-bit protocol, the delimiters and keywords
are such that protocol is mostly usable from USASCII terminal and a
telnet connection.
Because of IRC's scandanavian origin, the characters {}| are
considered to be the lower case equivalents of the characters []\,
respectively. This is a critical issue when determining the
equivalence of two nicknames.
|
|
Servers and clients send eachother messages which may or may not
generate a reply. If the message contains a valid command, as
described in later sections, the client should expect a reply as
specified but it is not advised to wait forever for the reply; client
to server and server to server communication is essentially
asynchronous in nature.
Each IRC message may consist of up to three main parts: the prefix
(optional), the command, and the command parameters (of which there
may be up to 15). The prefix, command, and all parameters are
separated by one (or more) ASCII space character(s) (0x20).
The presence of a prefix is indicated with a single leading ASCII
colon character (':', 0x3b), which must be the first character of the
message itself. There must be no gap (whitespace) between the colon
and the prefix. The prefix is used by servers to indicate the true
origin of the message. If the prefix is missing from the message, it
is assumed to have originated from the connection from which it was
received. Clients should not use prefix when sending a message from
themselves; if they use a prefix, the only valid prefix is the
registered nickname associated with the client. If the source
identified by the prefix cannot be found from the server's internal
database, or if the source is registered from a different link than
from which the message arrived, the server must ignore the message
silently.
The command must either be a valid IRC command or a three (3) digit
number represented in ASCII text.
IRC messages are always lines of characters terminated with a CR-LF
(Carriage Return - Line Feed) pair, and these messages shall not
exceed 512 characters in length, counting all characters including
the trailing CR-LF. Thus, there are 510 characters maximum allowed
for the command and its parameters. There is no provision for
continuation message lines. See section 7 for more details about
current implementations.
|
|
The protocol messages must be extracted from the contiguous stream of
octets. The current solution is to designate two characters, CR and
LF, as message separators. Empty messages are silently ignored,
which permits use of the sequence CR-LF between messages
without extra problems.
The extracted message is parsed into the components ,
and list of parameters matched either by or
components.
The BNF representation for this is:
::= [':' ]
::= | [ '!' ] [ '@' ]
::= { } |
::= ' ' { ' ' }
::= [ ':' | ]
::=
or NUL or CR or LF, the first of which may not be ':'>
::=
NUL or CR or LF>
::= CR LF
NOTES:
1) is consists only of SPACE character(s) (0x20).
Specially notice that TABULATION, and all other control
characters are considered NON-WHITE-SPACE.
2) After extracting the parameter list, all parameters are equal,
whether matched by or . is just
a syntactic trick to allow SPACE within parameter.
3) The fact that CR and LF cannot appear in parameter strings is
just artifact of the message framing. This might change later.
4) The NUL character is not special in message framing, and
basically could end up inside a parameter, but as it would
cause extra complexities in normal C string handling. Therefore
NUL is not allowed within messages.
5) The last parameter may be an empty string.
6) Use of the extended prefix (['!' ] ['@' ]) must
not be used in server to server communications and is only
intended for server to client messages in order to provide
clients with more useful information about who a message is
from without the need for additional queries.
Most protocol messages specify additional semantics and syntax for
the extracted parameter strings dictated by their position in the
list. For example, many server commands will assume that the first
parameter after the command is the list of targets, which can be
described with:
::= [ "," ]
::= | '@' | |
::= ('#' | '&')
::=
::= see RFC 952 [DNS:4] for details on allowed hostnames
::= { | | }
::= ('#' | '$')
::=
comma (',')>
Other parameter syntaxes are:
::= { }
::= 'a' ... 'z' | 'A' ... 'Z'
::= '0' ... '9'
::= '-' | '[' | ']' | '\' | '`' | '^' | '{' | '}'
::=
(0xd), and LF (0xa)>
|
|
Most of the messages sent to the server generate a reply of some
sort. The most common reply is the numeric reply, used for both
errors and normal replies. The numeric reply must be sent as one
message consisting of the sender prefix, the three digit numeric, and
the target of the reply. A numeric reply is not allowed to originate
from a client; any such messages received by a server are silently
dropped. In all other respects, a numeric reply is just like a normal
message, except that the keyword is made up of 3 numeric digits
rather than a string of letters. A list of different replies is
supplied in section 6.
|
|
This section is devoted to describing the actual concepts behind the
organization of the IRC protocol and how the current
implementations deliver different classes of messages.
1--\
A D---4
2--/ \ /
B----C
/ \
3 E
Servers: A, B, C, D, E Clients: 1, 2, 3, 4
[ Fig. 2. Sample small IRC network ]
|
|
Communication on a one-to-one basis is usually only performed by
clients, since most server-server traffic is not a result of servers
talking only to each other. To provide a secure means for clients to
talk to each other, it is required that all servers be able to send a
message in exactly one direction along the spanning tree in order to
reach any client. The path of a message being delivered is the
shortest path between any two points on the spanning tree.
The following examples all refer to Figure 2 above.
Example 1:
A message between clients 1 and 2 is only seen by server A, which
sends it straight to client 2.
Example 2:
A message between clients 1 and 3 is seen by servers A & B, and
client 3. No other clients or servers are allowed see the message.
Example 3:
A message between clients 2 and 4 is seen by servers A, B, C & D
and client 4 only.
|
|
The main goal of IRC is to provide a forum which allows easy and
efficient conferencing (one to many conversations). IRC offers
several means to achieve this, each serving its own purpose.
|
|
The least efficient style of one-to-many conversation is through
clients talking to a 'list' of users. How this is done is almost
self explanatory: the client gives a list of destinations to which
the message is to be delivered and the server breaks it up and
dispatches a separate copy of the message to each given destination.
This isn't as efficient as using a group since the destination list
is broken up and the dispatch sent without checking to make sure
duplicates aren't sent down each path.
|
|
In IRC the channel has a role equivalent to that of the multicast
group; their existence is dynamic (coming and going as people join
and leave channels) and the actual conversation carried out on a
channel is only sent to servers which are supporting users on a given
channel. If there are multiple users on a server in the same
channel, the message text is sent only once to that server and then
sent to each client on the channel. This action is then repeated for
each client-server combination until the original message has fanned
out and reached each member of the channel.
The following examples all refer to Figure 2.
Example 4:
Any channel with 1 client in it. Messages to the channel go to the
server and then nowhere else.
Example 5:
2 clients in a channel. All messages traverse a path as if they
were private messages between the two clients outside a channel.
Example 6:
Clients 1, 2 and 3 in a channel. All messages to the channel are
sent to all clients and only those servers which must be traversed
by the message if it were a private message to a single client. If
client 1 sends a message, it goes back to client 2 and then via
server B to client 3.
|
|
To provide IRC operators with some mechanism to send messages to a
large body of related users, host and server mask messages are
provided. These messages are sent to users whose host or server
information match that of the mask. The messages are only sent to
locations where users are, in a fashion similar to that of channels.
|
|
The one-to-all type of message is better described as a broadcast
message, sent to all clients or servers or both. On a large network
of users and servers, a single message can result in a lot of traffic
being sent over the network in an effort to reach all of the desired
destinations.
For some messages, there is no option but to broadcast it to all
servers so that the state information held by each server is
reasonably consistent between servers.
|
|
There is no class of message which, from a single message, results in
a message being sent to every other client.
|
|
Most of the commands which result in a change of state information
(such as channel membership, channel mode, user status, etc) must be
sent to all servers by default, and this distribution may not be
changed by the client.
|
|
While most messages between servers are distributed to all 'other'
servers, this is only required for any message that affects either a
user, channel or server. Since these are the basic items found in
IRC, nearly all messages originating from a server are broadcast to
all other connected servers.
|
|