path: root/erts/doc/src/erl_dist_protocol.xml

<?xml version="1.0" encoding="iso-8859-1" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">

<chapter>
  <header>
    <copyright>
      <year>2007</year>
      <year>2011</year>
      <holder>Ericsson AB, All Rights Reserved</holder>
    </copyright>
    <legalnotice>
  The contents of this file are subject to the Erlang Public License,
  Version 1.1, (the "License"); you may not use this file except in
  compliance with the License. You should have received a copy of the
  Erlang Public License along with this software. If not, it can be
  retrieved online at http://www.erlang.org/.

  Software distributed under the License is distributed on an "AS IS"
  basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
  the License for the specific language governing rights and limitations
  under the License.

  The Initial Developer of the Original Code is Ericsson AB.
    </legalnotice>

    <title>Distribution Protocol</title>
    <prepared></prepared>
    <docno></docno>
    <date>2007-09-21</date>
    <rev>PA1</rev>
    <file>erl_dist_protocol.xml</file>
  </header>

<p>
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.
</p>

<p>The distribution protocol can be divided into four (4) parts:</p>
<list>
  <item>
    <p>
      1. Low level socket connection.
    </p>
  </item>
  <item>
    2. Handshake, interchange node name and authenticate.
  </item>
  <item>
    3. Authentication (done by net_kernel).
  </item>
  <item>
    4. Connected. 
  </item>
</list>
<p>
  A node fetches the Port number of another node through the EPMD (at the
  other host) in order to initiate a connection request.
</p>
<p>
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.
</p>
<p>
By default EPMD listens on port 4369.
</p>
<p>
  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).
</p>

<p>The integers in all multi-byte fields are in big-endian order.</p>

  <section>
    <title>EPMD Protocol</title>
    <p>
      The requests served by the EPMD (Erlang Port Mapper Daemon) are 
      summarized in the figure below.
    </p>

      <image file="erl_ext_fig.gif">
	<icaption>
	  Summary of EPMD requests. 
	</icaption>
      </image>
      <p>
	Each request <c>*_REQ</c> is preceded by a two-byte length field. 
	Thus, the overall request format is:
      </p>

      <table align="left">
	<row>
	  <cell align="center">2</cell>
	  <cell align="center">n</cell>
	</row>
	<row>
	  <cell align="center">Length</cell>
	  <cell align="center">Request</cell>
	</row>
      <tcaption></tcaption></table>

      <section>
	<title>Register a node in the EPMD</title>
	<p>
	  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.
	</p>
	<table align="left">
	  <row>
	    <cell align="center">1</cell>
	    <cell align="center">2</cell>
	    <cell align="center">1</cell>
	    <cell align="center">1</cell>
	    <cell align="center">2</cell>
	    <cell align="center">2</cell>
	    <cell align="center">2</cell>
	    <cell align="center">Nlen</cell>
	    <cell align="center">2</cell>
	    <cell align="center">Elen</cell>
	  </row>
	  <row>
	    <cell align="center">120</cell>
	    <cell align="center">PortNo</cell>
	    <cell align="center">NodeType</cell>
	    <cell align="center">Protocol</cell>
	    <cell align="center">HighestVersion</cell>
	    <cell align="center">LowestVersion</cell>
	    <cell align="center">Nlen</cell>
	    <cell align="center">NodeName</cell>
	    <cell align="center">Elen</cell>
	    <cell align="center">Extra</cell>
	  </row>
	<tcaption>ALIVE2_REQ (120)</tcaption></table>
	  <taglist>
	    <tag><c>PortNo</c></tag>
	    <item>
	      The port number on which the node accept connection requests.
	    </item>
	    <tag><c>NodeType</c></tag>
	    <item>
	      77 = normal Erlang node, 72 = hidden node (C-node),...
	    </item>
	    <tag><c>Protocol</c></tag>
	    <item>
	      0 = tcp/ip-v4, ...
	    </item>
	    <tag><c>HighestVersion</c></tag>
	    <item>
	      The highest distribution version that this node can handle.
	      The value in R6B and later is 5.
	    </item>
	    <tag><c>LowestVersion</c></tag>
	    <item>
	      The lowest distribution version that this node can handle.
	      The value in R6B and later is 5.
	    </item>
	    <tag><c>Nlen</c></tag>
	    <item>
	      The length of the <c>NodeName</c>.
	    </item>
	    <tag><c>NodeName</c></tag>
	    <item>
	      The NodeName as a string of length <c>Nlen</c>.
	    </item>
	    <tag><c>Elen</c></tag>
	    <item>
	      The length of the <c>Extra</c> field.
	    </item>
	    <tag><c>Extra</c></tag>
	    <item>
	      Extra field of <c>Elen</c> bytes.
	    </item>
	  </taglist>
	<p>
	  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.
	</p>
	<p>
	  The response message ALIVE2_RESP is described below.
	</p>

	<table align="left">
	  <row>
	    <cell align="center">1</cell>
	    <cell align="center">1</cell>
	    <cell align="center">2</cell>
	  </row>
	  <row>
	    <cell align="center">121</cell>
	    <cell align="center">Result</cell>
	    <cell align="center">Creation</cell>
	  </row>
	<tcaption>ALIVE2_RESP (121)</tcaption></table>
	<p>
	  Result = 0 -> ok, Result > 0 -> error
	</p>
      </section>

      <section>
	<title>Unregister a node from the EPMD</title>
	<p>
	  A node unregisters itself from the EPMD by simply closing the
	  TCP connection towards EPMD established when the node was registered.
	</p>
      </section>

      <section>
	<title>Get the distribution port of another node</title>
	<p>
	  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.
	</p>

	<table align="left">
	<row>
	  <cell align="center">1</cell>
	  <cell align="center">N</cell>
	</row>
	<row>
	  <cell align="center">122</cell>
	  <cell align="center">NodeName</cell>
	</row>
      <tcaption>PORT_PLEASE2_REQ (122)</tcaption></table>
      <p>
	where N = Length - 1
      </p>

      <p>
      </p>
      <table align="left">
	<row>
	  <cell align="center">1</cell>
	  <cell align="center">1</cell>
	</row>
	<row>
	  <cell align="center">119</cell>
	  <cell align="center">Result</cell>
	</row>
	<tcaption>
	  PORT2_RESP (119) response indicating error, Result > 0.
	</tcaption>
      </table>
      <p>Or</p>
      <table align="left">
	<row>
	  <cell align="center">1</cell>
	  <cell align="center">1</cell>
	  <cell align="center">2</cell>
	  <cell align="center">1</cell>
	  <cell align="center">1</cell>
	  <cell align="center">2</cell>
	  <cell align="center">2</cell>
	  <cell align="center">2</cell>
	  <cell align="center">Nlen</cell>
	  <cell align="center">2</cell>
	  <cell align="center">Elen</cell>
	</row>
	<row>
	  <cell align="center">119</cell>
	  <cell align="center">Result</cell>
	  <cell align="center">PortNo</cell>
	  <cell align="center">NodeType</cell>
	  <cell align="center">Protocol</cell>
	  <cell align="center">HighestVersion</cell>
	  <cell align="center">LowestVersion</cell>
	  <cell align="center">Nlen</cell>
	  <cell align="center">NodeName</cell>
	  <cell align="center">Elen</cell>
	  <cell align="center">Extra</cell>
	</row>
      <tcaption>PORT2_RESP when Result = 0.</tcaption></table>
<p>
If Result > 0, the packet only consists of [119, Result].
</p>

        <p>EPMD will close the socket as soon as it has sent the information.</p>
      </section>

      <section>
	<title>Get all registered names from EPMD</title>
	<p>
	  This request is used via the Erlang function 
	  <c>net_adm:names/1,2</c>. A TCP connection is opened
	  towards EPMD and this request is sent.
	</p>
	<table align="left">
	  <row>
	    <cell align="center">1</cell>
	  </row>
	  <row>
	    <cell align="center">110</cell>
	  </row>
	<tcaption>NAMES_REQ (110)</tcaption></table>
	

	<p>The response for a <c>NAMES_REQ</c> looks like this:</p>
	<table align="left">
	  <row>
	    <cell align="center">4</cell>
	    <cell align="center">&nbsp;</cell>
	  </row>
	  <row>
	    <cell align="center">EPMDPortNo</cell>
	    <cell align="center">NodeInfo*</cell>
	  </row>
	<tcaption>NAMES_RESP</tcaption></table>
	<p>
	  NodeInfo is a string written for each active node.
	  When all NodeInfo has been written the connection is
	  closed by EPMD.
	</p>
	<p>
	  NodeInfo is, as expressed in Erlang:
	</p>
	<code>
	io:format("name ~s at port ~p~n", [NodeName, Port]).
	</code>
      </section>


      <section>
	<title>Dump all data from EPMD</title>
	<p>
	  This request is not really used, it should be regarded as a debug
	  feature.
	</p>
	<table align="left">
	  <row>
	  <cell align="center">1</cell>
	  </row>
	  <row>
	    <cell align="center">100</cell>
	  </row>
	<tcaption>DUMP_REQ</tcaption></table>

	<p>The response for a <c>DUMP_REQ</c> looks like this:</p>
	<table align="left">
	  <row>
	    <cell align="center">4</cell>
	    <cell align="center">&nbsp;</cell>
	  </row>
	  <row>
	    <cell align="center">EPMDPortNo</cell>
	    <cell align="center">NodeInfo*</cell>
	  </row>
	<tcaption>DUMP_RESP</tcaption></table>
	<p>
	  NodeInfo is a string written for each node kept in EPMD.
	  When all NodeInfo has been written the connection is
	  closed by EPMD.
	</p>
	<p>
	  NodeInfo is, as expressed in Erlang:
	</p>
	<code>
	io:format("active name     ~s at port ~p, fd = ~p ~n",
	          [NodeName, Port, Fd]).
	</code>
	<p>
	  or
	</p>
	<code>
	io:format("old/unused name ~s at port ~p, fd = ~p~n",
	          [NodeName, Port, Fd]).
	</code>

      </section>

      <section>
	<title>Kill the EPMD</title>
	<p>
	  This request will kill the running EPMD. It is almost never used.
	</p>
	<table align="left">
	  <row>
	  <cell align="center">1</cell>
	  </row>
	  <row>
	    <cell align="center">107</cell>
	  </row>
	<tcaption>KILL_REQ</tcaption></table>

	<p>The response fo a <c>KILL_REQ</c> looks like this:</p>
	<table align="left">
	  <row>
	    <cell align="center">2</cell>
	  </row>
	  <row>
	    <cell align="center">OKString</cell>
	  </row>
	<tcaption>KILL_RESP</tcaption></table>
	<p>
	  where <c>OKString</c> is "OK".
	</p>
      </section>

      <section>
	<title>STOP_REQ  (Not Used)</title>
	<p></p>
	<table align="left">
	  <row>
	    <cell align="center">1</cell>
	    <cell align="center">n</cell>
	  </row>
	  <row>
	    <cell align="center">115</cell>
	    <cell align="center">NodeName</cell>
	  </row>
	<tcaption>STOP_REQ</tcaption></table>
	<p>
	  where n = Length - 1
	</p>
	<p>
	  The current implementation of Erlang does not care if the connection
	  to the EPMD is broken.
	</p>
	<p>The response for a <c>STOP_REQ</c> looks like this.</p>
	<table align="left">
	  <row>
	    <cell align="center">7</cell>
	  </row>
	  <row>
	    <cell align="center">OKString</cell>
	  </row>
	<tcaption>STOP_RESP</tcaption></table>
	<p>
	  where OKString is "STOPPED".
	</p>
	<p>A negative response can look like this.</p>
	<table align="left">
	  <row>
	    <cell align="center">7</cell>
	  </row>
	  <row>
	    <cell align="center">NOKString</cell>
	  </row>
	<tcaption>STOP_NOTOK_RESP</tcaption></table>
	<p>
	  where NOKString is "NOEXIST".
	</p>
      </section>
<!--
      <section>
	<title>ALIVE_REQ (97)</title>
	<p></p>

	<table align="left">
	  <row>
	    <cell align="center">1</cell>
	    <cell align="center">2</cell>
	    <cell align="center">n</cell>
	  </row>
	  <row>
	    <cell align="center">97</cell>
	    <cell align="center">PortNo</cell>
	    <cell align="center">NodeName</cell>
	  </row>
	<tcaption></tcaption></table>

	<p>
	  where n = Length - 3
	</p>
	<p>
	  The connection created to the EPMD must be kept until the node is
	  not a distributed node any longer.
	</p>
      </section>

      <section>
	<title>ALIVE_OK_RESP (89)</title>
	<p></p>
	<table align="left">
	  <row>
	    <cell align="center">1</cell>
	    <cell align="center">2</cell>
	  </row>
	  <row>
	    <cell align="center">89</cell>
	    <cell align="center">Creation</cell>
	  </row>
	<tcaption></tcaption></table>
      </section>


      <section>
	<title>ALIVE_NOTOK_RESP  ()</title>
	<p>
	  EPMD closed the connection.
	</p>
      </section>

      <section>
	<title>PORT_PLEASE_REQ   (112)</title>
	<p></p>
      <table align="left">
	<row>
	  <cell align="center">1</cell>
	  <cell align="center">n</cell>
	</row>
	<row>
	  <cell align="center">112</cell>
	  <cell align="center">NodeName</cell>
	</row>
      <tcaption></tcaption></table>
      <p>
	where n = Length - 1
      </p>
      </section>

      <section>
	<title>PORT_OK_RESP ()</title>
	<p></p>
	<table align="left">
	  <row>
	    <cell align="center">2</cell>
	  </row>
	  <row>
	    <cell align="center">PortNo</cell>
	  </row>
	<tcaption></tcaption></table>

      </section>

      <section>
	<title>PORT_NOTOK_RESP ()</title>
	<p>
	  EPMD closed the connection.
	</p>
      </section>


      <section>
	<title>PORT_NOTOK_RESP ()</title>
	<p>
	  EPMD closed the connection.
	</p>
      </section>
-->

    </section>

      <section>
	<title>Handshake</title>
	<p>
	  The handshake is discussed in detail in the internal documentation for 
	  the kernel (Erlang) application.
	</p>
      </section>

      <section>
        <marker id="connected_nodes"/>
	<title>Protocol between connected nodes</title>
	<p>
	  As of erts version 5.7.2 the runtime system passes a distribution
	  flag in the handshake stage that enables the use of a
	  <seealso marker="erl_ext_dist#distribution_header">distribution
	  header</seealso> on all messages passed. Messages passed between
	  nodes are in this case on the following format:
	</p>
	<table align="left">
	  <row>
	    <cell align="center">4</cell>
	    <cell align="center">d</cell>
	    <cell align="center">n</cell>
	    <cell align="center">m</cell>
	  </row>
	  <row>
	    <cell align="center"><c>Length</c></cell>
	    <cell align="center"><c>DistributionHeader</c></cell>
	    <cell align="center"><c>ControlMessage</c></cell>
	    <cell align="center"><c>Message</c></cell>
	  </row>
	<tcaption></tcaption></table>
	<p>
	  where:
	</p>
	<p>
	  <c>Length</c> is equal to d + n + m
	</p>
	<p>
	  <c>ControlMessage</c> is a tuple passed using the external format of
	  Erlang.
	</p>
	<p>
	  <c>Message</c> is the message sent to another node using the '!'
	  (in external format). Note that <c>Message</c> is only passed in
	  combination with a <c>ControlMessage</c> encoding a send ('!').
	</p>
	<p>
	  Also note that <seealso marker="erl_ext_dist#overall_format">the
	  version number is omitted from the terms that follow a
	  distribution header</seealso>.
	</p>
	<p>
	  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:
	</p>
	<table align="left">
	  <row>
	    <cell align="center">4</cell>
	    <cell align="center">1</cell>
	    <cell align="center">n</cell>
	    <cell align="center">m</cell>
	  </row>
	  <row>
	    <cell align="center"><c>Length</c></cell>
	    <cell align="center"><c>Type</c></cell>
	    <cell align="center"><c>ControlMessage</c></cell>
	    <cell align="center"><c>Message</c></cell>
	  </row>
	<tcaption></tcaption></table>
	<p>
	  where:
	</p>
	<p>
	  <c>Length</c> is equal to 1 + n + m
  	</p>
	<p>
	  Type is: 112 (pass through)
	</p>
	<p>
	  <c>ControlMessage</c> is a tuple passed using the external format of
	  Erlang.
	</p>
	<p>
	  <c>Message</c> is the message sent to another node using the '!'
	  (in external format).  Note that <c>Message</c> is only passed in
	  combination with a <c>ControlMessage</c> encoding a send ('!').
	</p>
	<p>
	  The <c>ControlMessage</c> is a tuple, where the first element
	  indicates which distributed operation it encodes.
	</p>
	<taglist>
	  <tag><c>LINK</c></tag>
	  <item>
	    <p>
	      <c>{1, FromPid, ToPid}</c>
	    </p>
	  </item>
	  
	  <tag><c>SEND</c></tag>
	  <item>
	    <p>
	      <c>{2, Cookie, ToPid}</c>
	    </p>
	    <p>
	    <em>Note</em> followed by <c>Message</c>
	    </p>
	  </item>

	  <tag><c>EXIT</c></tag>
	  <item>
	    <p>
	      <c>{3, FromPid, ToPid, Reason}</c>
	    </p>
	  </item>
	  
	  <tag><c>UNLINK</c></tag>
	  <item>
	    <p>
	      <c>{4, FromPid, ToPid}</c>
	    </p>
	  </item>
	  
	  <tag><c>NODE_LINK</c></tag>
	  <item>
	    <p>
	      <c>{5}</c>
	    </p>
	  </item>
	  
	  <tag><c>REG_SEND</c></tag>
	  <item>
	    <p>
	      <c>{6, FromPid, Cookie, ToName}</c>
	    </p>
	    <p>
	    <em>Note</em> followed by <c>Message</c>
	    </p>
	  </item>

	  <tag><c>GROUP_LEADER</c></tag>
	  <item>
	    <p>
	      <c>{7, FromPid, ToPid}</c>
	    </p>
	  </item>
	
	  <tag><c>EXIT2</c></tag>
	  <item>
	    <p>
	      <c>{8, FromPid, ToPid, Reason}</c>
	    </p>
	  </item>
	</taglist>
      </section>

  
    <section>
      <title>New Ctrlmessages for distrvsn = 1 (OTP R4)</title>
      <taglist>
	<tag><c>SEND_TT</c></tag>
	<item>
	<p>
	  <c>{12, Cookie, ToPid, TraceToken}</c>
	</p>
	<p>
	<em>Note</em> followed by <c>Message</c>
	</p>
	</item>

	<tag><c>EXIT_TT</c></tag>
	<item>
	<p>
	  <c>{13, FromPid, ToPid, TraceToken, Reason}</c>	  
	</p>
	</item>
      
	<tag><c>REG_SEND_TT</c></tag>
	<item>
	<p>
	  <c>{16, FromPid, Cookie, ToName, TraceToken}</c>
	</p>
	<p>
	<em>Note</em> followed by <c>Message</c>
	</p>
	</item>
      
	<tag><c>EXIT2_TT</c></tag>
	<item>
	<p>
	  <c>{18, FromPid, ToPid, TraceToken, Reason}</c>
	</p>
	</item>
      </taglist>
    </section>
    
    <section>
      <title>New Ctrlmessages for distrvsn = 2</title>
      <p>
	distrvsn 2 was never used.
      </p>
    </section>
    
    <section>
      <title>New Ctrlmessages for distrvsn = 3 (OTP R5C)</title>
      <p>
	None, but the version number was increased anyway.
      </p>
    </section>

    <section>
      <title>New Ctrlmessages for distrvsn = 4 (OTP R6)</title>
      <p>
	These are only recognized by Erlang nodes, not by hidden nodes.
      </p>
      <taglist>
	<tag><c>MONITOR_P</c></tag>
	<item>
	<p>
	  <c>{19, FromPid, ToProc, Ref}</c>
	  
	  <c>FromPid</c> = monitoring process
	  <c>ToProc</c> = monitored process pid or name (atom)
	</p>
	</item>
      
	<tag><c>DEMONITOR_P</c></tag>
	<item>
	<p>
	  <c>{20, FromPid, ToProc, Ref}</c>
	  We include the FromPid just in case we want to trace this.
	  
	  <c>FromPid</c> = monitoring process
	  <c>ToProc</c> = monitored process pid or name (atom)
	</p>
	</item>

	<tag><c>MONITOR_P_EXIT</c></tag>
	<item>
	<p>
	  <c>{21, FromProc, ToPid, Ref, Reason}</c>
	  
	  <c>FromProc</c> = monitored process pid or name (atom)
	  <c>ToPid</c> = monitoring process
	  <c>Reason</c> = exit reason for the monitored process
	</p>
	</item>
      </taglist>
    </section>
  </chapter>