The Multicast Application Ports

The Internet community has recognized the need to be judicious when assigning port numbers (see ). With unicast applications, the need for explicit port assignment has been reduced by techniques such as locally assigning a dynamic port, combined with some mechanism for advertising that port (see ). Dynamic assignment does not work with multicast applications because it is impossible to guarantee that the port remains unused by all hosts that may want to join a given multicast group. The result is that each multicast application-layer protocol has had to have its own dedicated port assignment. Even worse, each different use of that multicast application-layer protocol has had to have a different unique port assigned. In the TCP/IP model, the port number in the transport layer multiplexes applications within a host (see and ). With multicast, the use of a port number to multiplex applications is unnecessary because the destination multicast address already provides a unique identifier for the application. Because of the desire to conserve port numbers and the fact that a port is not necessary to multiplex multicast applications, this document assigns two UDP ports that may be used with multicast applications. Use of these ports is optional because there may be circumstances where assigning a port is preferred, such as when participants cannot meet the requirements in and .

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here.

This document assigns UDP port 49151 and gives it the service name "Multicast Application Port". It also assigns UDP port 49150 and give it the service name "Multicast Application Port (Secondary)". These two ports are at the end of the range of User ports to differentiate them from other port assignments. Two ports are assigned because RTP recommends using a sequential port pair: an even port for RTP and the next higher port for RTCP (see ). RTP is a popular protocol for multicast video streams and many implementations require this port scheme. The assigned ports may be used as a source port if the application exclusively uses multicast messages. If any application messages are unicast, then a dynamic port should be used as the source port. This allows receivers to know which port to send replies to.

Hosts shall regard the use of these ports as non-exclusive. In practice, this means hosts using POSIX socket APIs would consider the SO_REUSEADDR socket option as having been enabled, regardless of the actual state of that flag . Hosts shall discard all incoming, non-multicast packets that use these destination ports.

Applications running on non-conformant hosts can successfully use these ports if they meet the requirements in this section. Applications running on a non-conformant host shall not prevent other applications from using these ports. In practice, this means that applications using POSIX socket APIs must enable the SO_REUSEADDR and/or SO_REUSEPORT socket options . Applications running on a non-conformant host shall discard all datagrams that do not have the multicast address used by the application.

Applications running on non-conformant hosts are vulnerable to a denial of service attack if another application claims exclusive access to the ports.

IANA is requested to assign the following ports: Contact IETF Chair Description Multicast data where the application is identified by the multicast address Reference This document Port Number 49151]]> Contact IETF Chair Description A secondary port for multicast data identified by the multicast address. Used when the protocol requires consecutive ports. Reference This document Port Number 49150]]> IANA is requested to update its "Application for Service Names and User Port Numbers" to reference this document, ask if a Multicast Application Port may be used, and require an explanation if not.

Special thanks to the National Marine Electronics Association for their contributions in developing marine industry standards and their support for this research. Thanks also to the members of the PIM working group for their review of this draft.