PIM Null-Register packing

The DR periodically sends PIM Null-Registers to keep the state of existing multicast sources active on the RP. As the number of multicast sources increases, the number of PIM Null-Register messages that are sent also increases. This results in more PIM packet processing at the RP and the DR. This document specifies a method to efficiently pack the content of multiple PIM Null-Register and Register-Stop messages into a single message. The document also discusses interoperability between PIM routers that support PIM Packed Null-Registers and PIM Packed Register-Stops and PIM routers that do not.

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.

Rendezvous Point Designated Router

This section allocates a bit in the PIM Register-Stop message Flag Bits field for the RP to indicate its ability to receive PIM Packed Null-Register messages (Section 3), and send PIM Packed Register-Stop messages (Section 4).

0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |PIM Ver| Type |P|6 5 4 3 2 1 0| Checksum | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Group Address (Encoded-Group format) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Source Address (Encoded-Unicast format) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Figure 1: PIM Register-Stop message with Packing Capability option The fields in the PIM Register-Stop message are defined in Section 4.9.4 of , and the common header in . Packing Capability bit (P-bit / Flag Bit TBD1): When set, it indicates the ability of the RP to receive PIM Packed Null-Register messages, and send PIM Packed Register-Stop messages.

0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |PIM Ver| Type |Subtype| FB | Checksum | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Group Address[1] (Encoded-Group format) | | Source Address[1] (Encoded-Unicast format) | . . . . . . . . . Group Address[N] . | Source Address[N] | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Figure 2: PIM Packed Null-Register message format The fields in the PIM Packed Null-Register message are defined in Section 4.9.4 of , and the common header in Type, Subtype: The PIM Packed Null-Register Type value TBD2. N: The total number of records; A record consists of a Group Address and Source Address pair. After parsing the PIM common header, individual records are then parsed one by one until the length of the PIM Packed Null-Register message. This length is inferred from the IP layer. Sending or receiving a PIM Packed Null-Register message is the equivalent, for all purposes, of sending or receiving an individual Null-Register message for each record represented in the PIM Packed Null-Register message.

0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |PIM Ver| Type |Subtype| FB | Checksum | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Group Address[1] (Encoded-Group format) | | Source Address[1] (Encoded-Unicast format) | . . . . . . . . . Group Address[N] . | Source Address[N] | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Figure 3: PIM Packed Register-Stop message format The fields in the PIM Packed Register-Stop message are defined in Section 4.9.4 of , and the common header in Type, Subtype: The PIM Packed Register-Stop Type TBD3 N: The total number of records; A record consists of a Group Address and Source Address pair. After parsing the PIM common header, individual records are then parsed one by one until the length of the PIM Packed Register-Stop message. This length is inferred from the IP layer. Sending or receiving a PIM Packed Register-Stop message is the equivalent, for all purposes, of sending or receiving an individual Null-Register message for each record represented in the PIM Packed Register-Stop.

As specified in , the DR sends PIM Register messages towards the RP when a new source is detected. When this feature is enabled/configured, an RP supporting this specification MUST set the P-bit (Flag bit TBD1) in all Register-Stop messages. When a Register-Stop message with the P-bit set is received, the DR SHOULD send PIM Packed Null-Register messages (Section 3) to the RP instead of multiple Register messages with the N-bit set . The DR MAY use a mixture of PIM Packed Null-Register messages and Register messages. The decision is up to the implementation and out of the scope of this document. However, it is RECOMMENDED to stick to the PIM Packed Null-Register and PIM Packed Register-Stop formats as long as the RP and DR have the feature enabled. The RP, after receiving a PIM Packed Null-Register message, SHOULD start sending PIM Packed Register-Stop messages (Section 4) to the corresponding DR instead of individual Register-Stop messages. The RP MAY use a mixture of PIM Packed Register-Stop messages and individual Register-Stop messages. The decision is up to the implementation and out of the scope of this document. However, it is RECOMMENDED to stick to the PIM Packed Null-Register and PIM Packed Register-Stop formats as long as the RP and DR have the feature enabled.

The PIM Packed Null-Register packet format should be enabled only if it is supported by all the routers in the Anycast-RP set . This consideration applies to PIM Anycast RP with MSDP as well.

A router (DR) can decide to use the PIM Packed Null-Register message format based on the Packing Capability received from the RP as part of the PIM Register-Stop. This ensures compatibility with routers that do not support processing of the new packet format. The Packing Capability information MUST be indicated by the RP via the PIM Register-Stop message sent to the DR. Thus, a DR will switch to the new packet format only when it learns that the RP is capable of handling the PIM Packed Null-Register messages. Conversely, a DR that does not support the packed format can continue generating the PIM Null-Register as defined in (Section 4.4).

Consider a PIM RP router that supports PIM Packed Null-Registers and PIM Packed Register-Stops. In scenarios where this router now no longer supports this feature, for example, in case of a software downgrade, it will not send a PIM Register-Stop message to the DR in response to a PIM Packed Null-Register message. When the DR switches to Data Registers from Null-Registers, it MUST start a Packed_Register_Probe_Time timer. If no PIM Packed Register-Stop or Register-Stop with the P-bit set is received within Packed_Register_Probe_Time seconds, the DR can decide that the RP no longer supports PIM Packed Null-Registers. The Packed_Register_Probe_Time timer is configurable; its default value is 60 seconds. When Packed_Register_Probe_Time expires, The DR MAY also send an unpacked PIM Null-Register and check the PIM Register-Stop to see if the P-bit is set or not. If it is not set then the DR will continue sending unpacked PIM Null-Register messages. In case the network manager disables the Packing Capability at the RP, or in other words, disables the feature from the RP, the router MUST NOT advertise the Packing Capability. However, an implementation MAY choose to still parse any packed registers if they are received. This may be particularly useful in the transitional period after the network manager disables it.

As explained in Section 4.4.1 of , the DR may perform Path MTU Discovery to the RP before sending PIM Packed Null-Register messages. Similarly, the RP may perform Path MTU Discovery to the DR before sending PIM Packed Register-Stop messages. In both cases, the number of records in a message should be limited such that it can fit within the Path MTU.

The Security Considerations from apply to this document. In particular, the effect of forging a PIM Packed Null-Register or Register-Stop message would be amplified to all the records included instead of just one. By forging a PIM Register-Stop message and setting the P-bit, an attacker can trigger the use of PIM Packed Null-Register messages by a DR thus creating unnecessary churn in the network.

When this document is published, IANA is asked to assign a Packing Capability bit (TBD1) in the PIM Register-Stop Common Header from the PIM Message Types registry. When this document is published, IANA is asked to assign a PIM message type (TBD2) for the PIM Packed Null-Register from the PIM Message Types registry. The Flag Bits (0-3) for PIM message type (TBD2) are requested to be "Unassigned". When this document is published, IANA is asked to assign a PIM message type (TBD3) for the PIM Packed Register-Stop from the PIM Message Types registry. The Flag Bits (0-3) for PIM message type (TBD3) are requested to be "Unassigned".

The authors would like to thank Stig Venaas, Alvaro Retana, Anish Peter, Zheng Zhang and Umesh Dudani for their helpful comments on the document.