When dealing with multicast traffic across multiple network interfaces, traditional routing methods can present unexpected limitations. The scenario where you need to direct different multicast groups through specific NICs requires careful configuration, especially when transitioning from deprecated tools like route to modern alternatives like ip route.

The fundamental issue arises because multicast routing traditionally expects a single default route for the entire 224.0.0.0/4 range. When attempting to add identical multicast routes for multiple interfaces using:

ip route add 224.0.0.0/4 dev eth0
ip route add 224.0.0.0/4 dev eth1

The second command fails with "RTNETLINK answers: File exists" because the kernel doesn't support duplicate multicast routes by default.

Several solutions exist depending on your specific requirements:

1. Using Specific Multicast Group Ranges

If you know the specific multicast groups in advance, you can assign different ranges to different interfaces:

ip route add 233.0.0.0/8 dev eth0
ip route add 239.0.0.0/8 dev eth1

This works because 233.x.x.x and 239.x.x.x are different multicast address ranges within the larger 224.0.0.0/4 space.

2. Network Namespace Solution

For complete isolation, consider using network namespaces:

ip netns add netns0
ip link set eth0 netns netns0
ip netns exec netns0 ip route add 224.0.0.0/4 dev eth0

ip netns add netns1
ip link set eth1 netns netns1
ip netns exec netns1 ip route add 224.0.0.0/4 dev eth1

The most elegant solution is to properly configure your application sockets to specify the outgoing interface:

// For IPv4
struct in_addr localInterface;
localInterface.s_addr = inet_addr("192.168.0.2");
setsockopt(socket, IPPROTO_IP, IP_MULTICAST_IF, 
           (char *)&localInterface, 
           sizeof(localInterface));

// For IPv6
unsigned int ifindex = if_nametoindex("eth0");
setsockopt(socket, IPPROTO_IPV6, IPV6_MULTICAST_IF, 
           &ifindex, 
           sizeof(ifindex));

This approach is superior because:

• It doesn't require kernel routing table modifications
• The application explicitly controls which interface to use
• It's more portable across different systems

After implementation, verify your configuration with:

ip mroute show
netstat -gn

When dealing with high multicast traffic volumes:

• Consider using PIM (Protocol Independent Multicast) for more efficient routing
• Monitor network interface buffers to prevent packet drops
• Use ethtool -k to verify multicast optimizations are enabled

---

When working with multicast applications on Linux systems, proper routing configuration is crucial. Many developers face challenges when trying to route multicast traffic through multiple network interfaces. The traditional approach using the deprecated route command doesn't translate well to the modern ip route utility.

Consider a scenario where your application needs to handle two multicast groups on separate networks:

Socket 1: Bound to 192.168.0.2 → Multicast group 233.255.10.1
Socket 2: Bound to 10.57.31.2 → Multicast group 239.255.100.1

The old method using route worked fine:

route add -net 224.0.0.0 netmask 240.0.0.0 eth0
route add -net 224.0.0.0 netmask 240.0.0.0 eth1

But with ip route, the second command fails:

ip route add 224.0.0.0/4 dev eth0  # Works
ip route add 224.0.0.0/4 dev eth1  # Fails with "RTNETLINK answers: File exists"

The correct approach involves two key components:

1. Network Configuration

For multicast routing, you need to specify more specific routes:

ip route add 233.255.10.1/32 dev eth0
ip route add 239.255.100.1/32 dev eth1

Or if you need broader ranges:

ip route add 233.255.0.0/16 dev eth0
ip route add 239.255.0.0/16 dev eth1

2. Application-Level Configuration

The more robust solution is to properly configure your sockets using IP_MULTICAST_IF:

// C example for setting multicast interface
struct in_addr multicastInterface;
multicastInterface.s_addr = inet_addr("192.168.0.2");

if (setsockopt(socket_fd, IPPROTO_IP, IP_MULTICAST_IF, 
              (char*)&multicastInterface, sizeof(multicastInterface)) < 0) {
    perror("setsockopt IP_MULTICAST_IF");
    exit(EXIT_FAILURE);
}

Check your multicast routes with:

ip route show table all | grep 22

For socket-level verification, use netstat or ss:

ss -ulnp | grep multicast_group

• Always prefer specific multicast routes over broad ranges
• Implement proper socket configuration in your application
• Consider using network namespaces for complex multicast scenarios
• Document your multicast IP ranges and interface assignments

For systems that need persistent configuration, consider adding to /etc/network/interfaces:

post-up ip route add 233.255.10.1/32 dev eth0
post-up ip route add 239.255.100.1/32 dev eth1

Or create a dedicated routing table:

echo "100 multicast_routes" >> /etc/iproute2/rt_tables
ip route add 233.255.0.0/16 dev eth0 table multicast_routes
ip route add 239.255.0.0/16 dev eth1 table multicast_routes