UDP Broadcast: Send to All Hosts on a Subnet (2026)

UDP broadcast lets a single machine send a packet to every other machine on its local network segment without knowing their individual IP addresses.

Here’s a UDP broadcast in action, sending a simple "hello" message to the broadcast address 192.168.1.255 on a subnet. The netcat utility is perfect for this:

# On the sender machine:
echo "Hello, everyone!" | nc -u 192.168.1.255 12345

# On a receiver machine (replace 192.168.1.255 with your subnet's broadcast address):
nc -ul 12345

When you run the nc -ul 12345 command on another machine on the same subnet, you’ll see "Hello, everyone!" appear in its terminal. The -u flag specifies UDP, -l means listen, and 12345 is the arbitrary port number we’re using for this example. The sender doesn’t need to know the receiver’s IP; it just blasts the packet to the broadcast address.

The core problem UDP broadcast solves is efficient one-to-many communication within a local subnet. Imagine a service that needs to announce its presence or a status update to all clients nearby. Without broadcast, the server would have to:

1. Maintain a list of all clients.
2. Send a separate packet to each client.
3. Handle clients joining and leaving dynamically, updating its list constantly.

This is inefficient and complex. UDP broadcast bypasses all of that. The sender just sends one packet to a special IP address that routers are configured not to forward. This address, the broadcast address, is typically the last address in a subnet (e.g., 192.168.1.255 for 192.168.1.0/24). Every device on that subnet is configured to listen to its broadcast address by default.

Internally, when a host receives a packet addressed to its subnet’s broadcast address, its network interface card (NIC) is designed to pass that packet up to the operating system’s network stack, regardless of the specific IP address it’s currently using. The OS then delivers it to any application that has bound to that UDP port.

The key levers you control are:

• The Broadcast Address: This is derived from your subnet’s network address and subnet mask. For a /24 subnet (like 192.168.1.0/255.255.255.0), the broadcast address is 192.168.1.255. For a /16 (like 172.16.0.0/255.255.0.0), it’s 172.16.255.255. You can find your subnet’s broadcast address using ip addr show on Linux and looking at the brd field, or by calculating it.
• The Port Number: This is the UDP port your application listens on. All devices wanting to receive the broadcast must listen on the exact same port.
• Application Logic: Your application needs to be written to send to the broadcast address and listen on the specific UDP port.

A common misconception is that broadcast packets can traverse routers. This is fundamentally untrue by design. Routers, by default, will discard broadcast packets because forwarding them could quickly overwhelm networks with redundant traffic. Broadcasts are strictly a Layer 3 (IP) concept confined to a single broadcast domain, which is typically a single subnet.

The next thing you’ll likely encounter is realizing that while broadcast is great for local subnets, you’ll need multicast or other discovery mechanisms for inter-subnet communication.