Erlang OTP: Distributed & Fault-Tolerant Systems Programming · Lesson

Global Process Registration

Manage global process names and discover processes across a distributed Erlang cluster for coordinated operations.

Global Process Registration is a free Erlang OTP: Distributed & Fault-Tolerant Systems Programming lesson on CoddyKit — lesson 3 of 4. You can read the complete lesson below for free — then practise it hands-on in the browser with a built-in code editor and a 24/7 AI tutor. It is part of the Erlang OTP: Distributed & Fault-Tolerant Systems Programming learning path, one of 4 lessons in the course, and your progress syncs across the web and the CoddyKit app.

Global Names Across Nodes

In distributed Erlang, processes often need to find each other, even if they're on different machines. How do you find a process when you don't know which node it lives on?

Erlang's global module provides a solution: global process registration. It allows you to give a process a unique name that is known across all connected nodes in your cluster.

Local vs. Global Processes

Normally, when you register a process using register/2, its name is only known on the local node. If my_process is on nodeA@host, a process on nodeB@host can't find it just by its name.

Global registration solves this! It acts like a distributed directory service. Any process on any connected node can look up a globally registered name and get the corresponding Process ID (PID), regardless of where that process is running.

`global:register_name/2`

To make a process accessible globally, you use global:register_name(Name, Pid).

Name: An atom that will be the unique global name (e.g., my_service).
Pid: The Process ID of the process you want to register.

If the name is already taken, register_name returns {false, OldPid}, otherwise true. The global module ensures that a global name is unique across the entire distributed cluster.

`global:whereis_name/1`

Once a process is registered, any other process can find it using global:whereis_name(Name).

Name: The global name (atom) you're looking for.

This function returns the PID of the registered process if found, or the atom undefined if no process is registered under that name.

It's crucial to check for undefined before attempting to send messages to the returned PID.

Full Global Communication Demo

This runnable example brings it all together! A server process starts and registers itself globally. After a short delay, a client process then finds this global server and sends it a message, receiving a reply.

Observe how global:register_name ensures the server is known, and global:whereis_name allows the client to locate it, enabling seamless communication.

-module(global_demo).
-export([main/0]).

% Server Functions
server_init() ->
    case global:register_name(my_global_service, self()) of
        true ->
            io:format("Server (~p) registered as my_global_service~n", [self()]),
            server_loop();
        {false, OldPid} ->
            io:format("Error: my_global_service already registered by ~p~n", [OldPid])
    end.

server_loop() ->
    receive
        {From, hello} ->
            io:format("Server (~p) received 'hello' from ~p~n", [self(), From]),
            From ! {self(), "Hi from server!"},
            server_loop();
        _ ->
            io:format("Server (~p) received unknown message~n", [self()]),
            server_loop()
    end.

% Client Functions
client_task() ->
    io:format("Client (~p) trying to find my_global_service...~n", [self()]),
    case global:whereis_name(my_global_service) of
        undefined ->
            io:format("Client: my_global_service not found!~n");
        ServerPid ->
            io:format("Client: Found server ~p, sending 'hello'...~n", [ServerPid]),
            ServerPid ! {self(), hello},
            receive
                {ServerPid, Response} ->
                    io:format("Client: Received response from server: '~s'~n", [Response])
            after 5000 ->
                io:format("Client: No response from server within 5 seconds.~n")
            end
    end.

% Main entry point
main() ->
    % Start the server process
    spawn(?MODULE, server_init, []),
    timer:sleep(100), % Give server time to register

    % Start the client process
    spawn(?MODULE, client_task, []),

    timer:sleep(2000). % Allow processes to run and communicate

Removing Global Names

When a globally registered process terminates, the global module automatically unregisters its name. However, you can also explicitly unregister a name using global:unregister_name(Name).

This is useful if you want to replace a service, or if a process needs to temporarily relinquish its global name. Remember, once unregistered, other processes can no longer find it by that name.

Handling Name Collisions

What happens if two different processes (even on different nodes) try to register the same global name?

The global module ensures that a global name is unique across the entire distributed system. The first process to successfully register the name wins. Subsequent attempts to register the same name will fail and return {false, OldPid}, indicating who already holds the name.

This prevents ambiguity and ensures that whereis_name/1 always returns a single, correct PID.

Practical Applications

Global process registration is ideal for implementing:

Singleton Services: A single instance of a service (e.g., a configuration manager, a logger) accessible from anywhere.
Resource Managers: A process responsible for managing a shared resource that multiple parts of your distributed system need to access.
Entry Points: Providing a well-known name for the main entry point of a distributed application.

It simplifies process discovery in complex distributed architectures.

How Global Registration Works

The global module achieves its magic by maintaining a consistent view of registered names across all connected Erlang nodes.

When a name is registered, this information is broadcast to all other nodes. When a node connects or disconnects, the global module updates its internal state to reflect the current cluster topology and available global names.

This ensures that global:whereis_name/1 can quickly find the correct PID, no matter where it resides.

Global Registration Check

Consider a distributed Erlang system with two nodes, nodeA and nodeB, both connected. A process on nodeA successfully registers itself with global:register_name(my_service, self()).

What will happen if a process on nodeB then tries to call global:register_name(my_service, self())?

Global Names: A Distributed Directory

In this lesson, you learned about global process registration using Erlang's global module:

It provides a way to assign unique names to processes that are discoverable across an entire distributed Erlang cluster.
global:register_name(Name, Pid) makes a process globally accessible.
global:whereis_name(Name) allows any process on any connected node to find the PID associated with a global name.
The global module handles name uniqueness and automatically unregisters processes when they terminate, simplifying distributed process management.

Global names are fundamental for building robust, fault-tolerant, and discoverable services in distributed Erlang applications.

Frequently asked questions

Is the “Global Process Registration” lesson free?

Yes — the full text of “Global Process Registration” is free to read here on the web, and the Erlang OTP: Distributed & Fault-Tolerant Systems Programming course includes 4 lessons in total. To practise it interactively (a built-in code editor and a 24/7 AI tutor) and unlock the rest of the Erlang OTP: Distributed & Fault-Tolerant Systems Programming course, upgrade to CoddyKit PRO.

What will I learn in “Global Process Registration”?

Manage global process names and discover processes across a distributed Erlang cluster for coordinated operations. You practise Erlang OTP: Distributed & Fault-Tolerant Systems Programming with hands-on code you run directly in the browser, and a 24/7 AI tutor answers your questions as you work through the lesson.

Do I need any experience to start Erlang OTP: Distributed & Fault-Tolerant Systems Programming?

No prior experience is required. Erlang OTP: Distributed & Fault-Tolerant Systems Programming on CoddyKit is structured for beginners through advanced learners; this is — lesson 3 of 4, so you can start here or from the beginning and move at your own pace.

How long does the “Global Process Registration” lesson take?

Most CoddyKit lessons take about 5–10 minutes. Each one is bite-sized and interactive, so you make steady progress and pick up exactly where you left off across the web and the app.

Can I write and run code in this Erlang OTP: Distributed & Fault-Tolerant Systems Programming lesson?

Yes. Every Erlang OTP: Distributed & Fault-Tolerant Systems Programming lesson includes a built-in code editor, so you write and run real code right in your browser and get instant AI feedback — no local setup required.

All lessons in this course

← Back to Erlang OTP: Distributed & Fault-Tolerant Systems Programming