The description here is far from complete and will therefore be further refined in upcoming releases. The protocols both from Erlang nodes towards EPMD (Erlang Port Mapper Daemon) and between Erlang nodes, however, are stable since many years.
The distribution protocol can be divided into four (4) parts:
1. Low level socket connection.
A node fetches the Port number of another node through the EPMD (at the other host) in order to initiate a connection request.
For each host where a distributed Erlang node is running there should also be an EPMD running. The EPMD can be started explicitly or automatically as a result of the Erlang node startup.
By default EPMD listens on port 4369.
3 and 4 are performed at the same level but the net_kernel disconnects the other node if it communicates using an invalid cookie (after one (1) second).
The integers in all multi-byte fields are in big-endian order.
The requests served by the EPMD (Erlang Port Mapper Daemon) are summarized in the figure below.
Each request
When a distributed node is started it registers itself in EPMD. The message ALIVE2_REQ described below is sent from the node towards EPMD. The response from EPMD is ALIVE2_RESP.
The connection created to the EPMD must be kept as long as the node is a distributed node. When the connection is closed the node is automatically unregistered from the EPMD.
The response message ALIVE2_RESP is described below.
Result = 0 -> ok, Result > 0 -> error
A node unregisters itself from the EPMD by simply closing the TCP connection towards EPMD established when the node was registered.
When one node wants to connect to another node it starts with a PORT_PLEASE2_REQ request towards EPMD on the host where the node resides in order to get the distribution port that the node listens to.
where N = Length - 1
Or
If Result > 0, the packet only consists of [119, Result].
EPMD will close the socket as soon as it has sent the information.
This request is used via the Erlang function
The response for a
NodeInfo is a string written for each active node. When all NodeInfo has been written the connection is closed by EPMD.
NodeInfo is, as expressed in Erlang:
io:format("name ~ts at port ~p~n", [NodeName, Port]).
This request is not really used, it should be regarded as a debug feature.
The response for a
NodeInfo is a string written for each node kept in EPMD. When all NodeInfo has been written the connection is closed by EPMD.
NodeInfo is, as expressed in Erlang:
io:format("active name ~ts at port ~p, fd = ~p ~n",
[NodeName, Port, Fd]).
or
io:format("old/unused name ~ts at port ~p, fd = ~p~n",
[NodeName, Port, Fd]).
This request will kill the running EPMD. It is almost never used.
The response fo a
where
where n = Length - 1
The current implementation of Erlang does not care if the connection to the EPMD is broken.
The response for a
where OKString is "STOPPED".
A negative response can look like this.
where NOKString is "NOEXIST".
This section describes the distribution handshake protocol
introduced in the OTP-R6 release of Erlang/OTP. This
description was previously located in
The TCP/IP distribution uses a handshake which expects a connection based protocol, i.e. the protocol does not include any authentication after the handshake procedure.
This is not entirely safe, as it is vulnerable against takeover attacks, but it is a tradeoff between fair safety and performance.
The cookies are never sent in cleartext and the handshake procedure expects the client (called A) to be the first one to prove that it can generate a sufficient digest. The digest is generated with the MD5 message digest algorithm and the challenges are expected to be very random numbers.
A challenge is a 32 bit integer number in big endian order. Below the function
A digest is a (16 bytes) MD5 hash of the Challenge (as text) concatenated
with the cookie (as text). Below, the function
An out_cookie is the cookie used in outgoing communication to a certain node,
so that A's out_cookie for B should correspond with B's in_cookie for A and
the other way around. A's out_cookie for B and A's in_cookie for B need NOT
be the same. Below the function
An in_cookie is the cookie expected to be used by another node when
communicating with us, so that A's in_cookie for B corresponds with B's
out_cookie for A. Below the function
The cookies are text strings that can be viewed as passwords.
Every message in the handshake starts with a 16 bit big endian integer
which contains the length of the message (not counting the two initial bytes).
In erlang this corresponds to the
Imagine two nodes, node A, which initiates the handshake and node B, which accepts the connection.
A connects to B via TCP/IP and B accepts the connection.
A sends an initial identification to B. B receives the message. The message looks like this (every "square" being one byte and the packet header removed):
+---+--------+--------+-----+-----+-----+-----+-----+-----+-...-+-----+ |'n'|Version0|Version1|Flag0|Flag1|Flag2|Flag3|Name0|Name1| ... |NameN| +---+--------+--------+-----+-----+-----+-----+-----+-----+-... +-----+
The 'n' is just a message tag.
Version0 and Version1 is the distribution version selected by node A,
based on information from EPMD. (16 bit big endian)
Flag0 ... Flag3 are capability flags, the capabilities defined in
B sends a status message to A, which indicates if the connection is allowed. The following status codes are defined:
This is the format of the status message:
+---+-------+-------+-...-+-------+ |'s'|Status0|Status1| ... |StatusN| +---+-------+-------+-...-+-------+
's' is the message tag Status0 ... StatusN is the status as a string (not terminated)
If status was 'alive', node A will answer with
another status message containing either 'true' which means that the
connection should continue (The old connection from this node is broken), or
If the status was
+---+--------+--------+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-...-+-----+ |'n'|Version0|Version1|Flag0|Flag1|Flag2|Flag3|Chal0|Chal1|Chal2|Chal3|Name0|Name1| ... |NameN| +---+--------+--------+-----+-----+-----+-----+-----+-----+-----+-----+-----+-----+-... +-----+
Where Chal0 ... Chal3 is the challenge as a 32 bit big endian integer and the other fields are B's version, flags and full nodename.
Now A has generated a digest and its own challenge. Those are sent together in a package to B:
+---+-----+-----+-----+-----+-----+-----+-----+-----+-...-+------+ |'r'|Chal0|Chal1|Chal2|Chal3|Dige0|Dige1|Dige2|Dige3| ... |Dige15| +---+-----+-----+-----+-----+-----+-----+-----+-----+-...-+------+
Where 'r' is the tag, Chal0 ... Chal3 is A's challenge for B to handle and Dige0 ... Dige15 is the digest that A constructed from the challenge B sent in the previous step.
B checks that the digest received from A is correct and generates a digest from the challenge received from A. The digest is then sent to A. The message looks like this:
+---+-----+-----+-----+-----+-...-+------+ |'a'|Dige0|Dige1|Dige2|Dige3| ... |Dige15| +---+-----+-----+-----+-----+-...-+------+
Where 'a' is the tag and Dige0 ... Dige15 is the digest calculated by B for A's challenge.
A checks the digest from B and the connection is up.
A (initiator) B (acceptor) TCP connect -----------------------------------------> TCP accept send_name -----------------------------------------> recv_name <---------------------------------------- send_status recv_status (if status was 'alive' send_status - - - - - - - - - - - - - - - - - - - -> recv_status) ChB = gen_challenge() (ChB) <---------------------------------------- send_challenge recv_challenge ChA = gen_challenge(), OCA = out_cookie(B), DiA = gen_digest(ChB,OCA) (ChA, DiA) send_challenge_reply --------------------------------> recv_challenge_reply ICB = in_cookie(A), check: DiA == gen_digest (ChB, ICB) ? - if OK: OCB = out_cookie(A), DiB = gen_digest (DiB) (ChA, OCB) <----------------------------------------- send_challenge_ack recv_challenge_ack DONE ICA = in_cookie(B), - else check: CLOSE DiB == gen_digest(ChA,ICA) ? - if OK DONE - else CLOSE
Currently (OTP-R16) the following capability flags are defined:
%% The node should be published and part of the global namespace -define(DFLAG_PUBLISHED,1). %% The node implements an atom cache (obsolete) -define(DFLAG_ATOM_CACHE,2). %% The node implements extended (3 * 32 bits) references. This is %% required today. If not present connection will be refused. -define(DFLAG_EXTENDED_REFERENCES,4). %% The node implements distributed process monitoring. -define(DFLAG_DIST_MONITOR,8). %% The node uses separate tag for fun's (lambdas) in the distribution protocol. -define(DFLAG_FUN_TAGS,16#10). %% The node implements distributed named process monitoring. -define(DFLAG_DIST_MONITOR_NAME,16#20). %% The (hidden) node implements atom cache (obsolete) -define(DFLAG_HIDDEN_ATOM_CACHE,16#40). %% The node understand new fun-tags -define(DFLAG_NEW_FUN_TAGS,16#80). %% The node is capable of handling extended pids and ports. This is %% required today. If not present connection will be refused. -define(DFLAG_EXTENDED_PIDS_PORTS,16#100). %% -define(DFLAG_EXPORT_PTR_TAG,16#200). %% -define(DFLAG_BIT_BINARIES,16#400). %% The node understands new float format -define(DFLAG_NEW_FLOATS,16#800). %% -define(DFLAG_UNICODE_IO,16#1000). %% The node implements atom cache in distribution header. -define(DFLAG_DIST_HDR_ATOM_CACHE,16#2000). %% The node understand the SMALL_ATOM_EXT tag -define(DFLAG_SMALL_ATOM_TAGS, 16#4000). %% The node understand UTF-8 encoded atoms -define(DFLAG_UTF8_ATOMS, 16#10000).
As of erts version 5.7.2 the runtime system passes a distribution
flag in the handshake stage that enables the use of a
where:
Also note that
Nodes with an erts version less than 5.7.2 does not pass the distribution flag that enables the distribution header. Messages passed between nodes are in this case on the following format:
where:
Type is: 112 (pass through)
The
Note followed by
Note followed by
Note followed by
Note followed by
distrvsn 2 was never used.
None, but the version number was increased anyway.
These are only recognized by Erlang nodes, not by hidden nodes.