| Internet-Draft | PIM Forwarding Enhancements | June 2026 |
| Gopal, et al. | Expires 19 December 2026 | [Page] |
The Protocol Independent Multicast (PIM) Flooding Mechanism (PFM) is an experimental extension that provides a generic hop-by-hop message exchange framework for distributing multicast information among PIM routers. Existing PFM procedures enable efficient source discovery without reliance on Rendezvous Points, shared trees, or initial data registers.¶
This document specifies further experimental enhancements to PFM forwarding behavior to improve efficiency and scalability. In particular, it introduces mechanisms to reduce redundant message transmission over multiple parallel links and extends the encoding of multicast information through additional Type-Length-Value (TLV) structures and sub-TLVs to convey richer flow-related data. These enhancements optimize control-plane overhead while preserving interoperability with existing PFM procedures, enabling more efficient dissemination of multicast state in PIM networks.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 19 December 2026.¶
Copyright (c) 2026 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
PIM Flooding Mechanism [RFC8364] is an experimental extension that allows a PIM router in the network to originate a PFM message to distribute announcements of active sources to its PIM neighbors [RFC7761]. Each receiving PIM neighbor then processes the PFM message and forwards it further in accordance with RFC8364. To prevent loops, the originator address as defined in Section 3.1 [RFC8364] is used for Reverse Path Forwarding (RPF) checking at each router. This RPF check is defined in Section 3.4.1 [RFC8364]. Periodic PFM messages are exchanged to keep the multicast information updated across the PIM domain (Section 3.4.2 [RFC8364])¶
The TLV defined in [RFC8364] for source discovery conveys only source and group information. It does not provide a mechanism to include additional attributes describing a multicast flow.¶
In addition, PFM messages are flooded on all PIM-enabled links. When two routers maintain multiple PIM adjacencies, identical PFM messages are transmitted across each link. Receivers perform RPF checks and discard duplicates as needed. This behavior introduces unnecessary processing overhead, both periodically and upon source discovery.¶
This document defines two further independent experimental enhancements to PFM message exchange:¶
A new TLV that supports Sub-TLVs, enabling the inclusion of additional flow-related information. This enhancement is specified in Section 2.¶
An optimization for PFM message exchange across multiple PIM adjacencies between the same pair of routers. By leveraging PIM Router-IDs [RFC6395], routers can identify such adjacencies and limit message transmission to a subset of links, reducing redundant processing. This optimization applies to point-to-point links and does not alter behavior on shared LANs. This enhancement is specified in Section 3.¶
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 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
PFM-SD [RFC8364] defines a Group Source Holdtime (GSH) TLV for announcing active sources. The GSH TLV conveys only source and group information. This document defines an extension that allows PIM routers to exchange additional information associated with multicast sources.¶
This document defines a new GSI TLV (Type TBD1). The GSI TLV is functionally similar to the GSH TLV but applies to a single (S,G) entry and supports the inclusion of Sub-TLVs to convey additional flow-specific information. Support for the GSI TLV is advertised using a PIM Hello option (TBD2), as described in Section 2.2¶
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|T| Type = TBD1 | Length |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Group Address (Encoded-Group format) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Source Address (Encoded-Unicast format) |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Holdtime | Type Sub-TLV 1 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Length Sub-TLV 1 | Value Sub-TLV 1 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +
| . |
| . |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| . |
| . |
| Type Sub-TLV n | Length Sub-TLV n |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Value Sub-TLV n |
| . |
| . |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
¶
The format of the GSI TLV is as follows:¶
Zero or more Sub-TLVs MAY be included. Each Sub-TLV consists of:¶
A PIM router indicates support for the GSI TLV defined in this document by including the Group Source Info TLV Hello option in PIM Hello messages. The format of the Hello option is as follows:¶
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| OptionType = TBD2 | Length = 0 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
¶
The presence of this option signifies that the router supports the GSI TLV.¶
Routers that support the GSI TLV and have GSI TLV usage enabled MUST track which neighbors advertise support for the GSI TLV via the Hello option (Section 2.2). This tracking is beneficial in heterogeneous networks where only certain routers support the GSI TLV. Routers that advertise support for the GSI TLV via the Hello option MUST use GSI TLVs by default when all neighbors on the outgoing interface support the GSI TLV. An implementation MUST provide a configuration knob to disable GSI TLV usage; when disabled, the router MUST NOT advertise the Hello option and MUST NOT originate GSI TLVs.¶
Operationally, enabling GSI TLV usage is justified when routers need to signal flow-specific attributes using Sub-TLVs and when peers on the path support the GSI TLV. In mixed-capability or staged-deployment environments where such signaling is not required, operators can keep GSI TLV usage disabled and rely on GSH TLVs as specified in [RFC8364].¶
Both the GSH TLV (defined in [RFC8364]) and the GSI TLV defined in this document can coexist in the same PFM message. The GSH TLV remains available for backward compatibility and for cases where Sub-TLVs are not needed. The GSI TLV is used when routers need to convey additional flow-specific information via Sub-TLVs. When both GSH and GSI TLVs are present for the same (S,G), GSI takes precedence and GSH serves as fallback for neighbors that do not support the GSI TLV.¶
When GSI TLV usage is enabled, a router that supports the GSI TLV:¶
MUST advertise its capability by including the Hello option (OptionType TBD2) in PIM Hello messages.¶
MUST track, per PIM interface, whether all neighbors support the GSI TLV. The scope and persistence of this state are implementation-specific. An implementation MAY retain this state for operational visibility even when GSI TLV usage is disabled.¶
If acting as a First Hop Router (FHR), MUST originate a GSI TLV when all neighbors on the outgoing interface support the GSI TLV.¶
If acting as an FHR, MUST originate a GSH TLV [RFC8364] when any neighbor on the outgoing interface does not support the GSI TLV.¶
MUST forward the GSI TLV unchanged on interfaces where all neighbors support the GSI TLV.¶
MUST convert each GSI TLV to a GSH TLV [RFC8364] where there are interfaces with at least one neighbor that does not support the GSI TLV, and forward the converted GSH TLVs only on those interfaces. The conversion MUST preserve the group, source, and holdtime fields, and MUST ignore Sub-TLVs. Multiple (S,G) entries for the same group SHOULD be aggregated into a single GSH TLV. This aggregation is RECOMMENDED and is not a mandatory compliance requirement. However, it MUST still send GSI TLVs on all interfaces where the neighbors do support it.¶
To apply the forwarding optimization defined in this document, PIM routers MUST advertise a Router-ID as specified in [RFC6395]. Within a PIM VRF, a router MUST use the same 4-octet Router-ID in PIM Hello messages on all interfaces and MUST cache Router-IDs learned from neighbors. Within a PIM VRF, a router MUST identify interfaces with a single neighbor sharing the same Router-ID, indicating multiple adjacencies to that neighbor. This identification is necessary for applying the forwarding optimization defined in this document. Router-IDs are assumed to be unique within the PIM domain. If this assumption is violated, the optimization defined in this document cannot be applied.¶
A PIM router indicates support for the forwarding optimization by including the PFM Optimization Hello option (OptionType TBD3) in PIM Hello 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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| OptionType = TBD3 | Length = 0 |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
¶
A router that supports this optimization MUST track, per interface, whether all neighbors support the option. This tracking is beneficial in heterogeneous networks where only certain routers support the optimization.¶
For each learned Router-ID, the router MUST maintain a set of interfaces, denoted as PFM_OPT_IF, that satisfy both of the following conditions:¶
The neighbor with this Router-ID is the sole PIM neighbor on this interface.¶
The neighbor is advertising the PFM optimization option TBD3 on this interface.¶
PFM message exchange MAY be optimized on interfaces in the PFM_OPT_IF set. This is discussed in Section 3.4.¶
Router C
|
| LAN 1
|
Router A --------------------------------- Router B
| |
|----------------- Link L1 ----------------|
| |
|----------------- Link L2 ----------------|
| |
|----------------- Link L3 ----------------|
|__________________________________________|
|
| LAN 2
|
Router D
This figure is schematic and not intended to imply a specific physical layout. It depicts two routers (A and B) connected by three parallel point-to-point links (L1, L2, L3), with shared LAN segments (LAN 1 and LAN 2) attached to Routers C and D.¶
When two routers maintain multiple adjacencies and are the only neighbors on those links, PFM messages are typically transmitted on all links and filtered by RPF checks at the receiver. This results in redundant processing. If both routers advertise Router-IDs and support the optimization, each router forms a PFM_OPT_IF set containing eligible interfaces.¶
Implementations can experience transient states while link failures are being detected and propagated. During such intervals, a PFM message sent on the selected interface can be lost or delayed if that interface fails before the sender updates its PFM_OPT_IF state. This is not unique to the optimization in this document and can also occur in regular PFM operation when a message is sent on the selected RPF path and that path fails. Implementations SHOULD apply robust and timely link-failure detection and update forwarding/interface-selection state promptly. Even if one message is lost, subsequent periodic PFM messages refresh state.¶
When Relaxed-RPF is enabled:¶
A sender MUST select a single interface from its PFM_OPT_IF set for PFM transmission to that neighbor. The selection method is implementation-specific.¶
On shared LANs, the sender MUST send PFM messages as normal since optimization cannot be applied when there are more than two routers on the network segment.¶
A receiver that supports Relaxed-RPF MUST:¶
Referring to Figure 1, when Router A originates or forwards a PFM message, it transmits the message on one of links L1, L2, or L3. This behavior reduces processing overhead on point-to-point links. The selection of the interface from the PFM_OPT_IF set is implementation-specific. Router A also sends the message on both LAN 1 and LAN 2 so Routers C and D receive the message.¶
Routers MUST update the PFM_OPT_IF set upon neighbor or capability changes:¶
Neighbor Addition: If the new neighbor is the sole neighbor on the interface and advertises both a Router-ID and the optimization option, the interface MUST be added to the corresponding PFM_OPT_IF set. If no set exists, it MUST be created. If a second neighbor appears on the interface, the interface MUST be removed from the PFM_OPT_IF set.¶
Neighbor Removal: After a neighbor removal event, if exactly one neighbor remains and it advertises both a Router-ID and the optimization option, the interface MUST be added to the PFM_OPT_IF set for that Router-ID. If no set exists, it MUST be created.¶
Router-ID Changes: If a neighbor starts advertising a Router-ID and satisfies all conditions, the interface MUST be added to the PFM_OPT_IF set. If a neighbor stops advertising a Router-ID, the interface MUST be removed from the PFM_OPT_IF set for that Router-ID. If the set becomes empty, it MUST be deleted.¶
Optimization Capability Changes: If a neighbor starts advertising the optimization option and satisfies all conditions, the interface MUST be added to the PFM_OPT_IF set. If a neighbor stops advertising the optimization option, the interface MUST be removed from the PFM_OPT_IF set for that Router-ID. If the set becomes empty, it MUST be deleted.¶
These procedures apply during topology changes, configuration updates, and software upgrades or downgrades. Routers MUST maintain accurate PFM_OPT_IF state for each Router-ID.¶
When the TBD3 optimization is enabled on a PIM router, the router MUST NOT forward a PFM message on a link if both of the following conditions are true: (1) the link has only one neighbor, and (2) that neighbor's Router-ID matches the Router-ID of the router that originated the PFM message. It is sufficient for the neighbor to advertise only the Router-ID, without any additional optimization options, since this information alone ensures the message is not sent back to its original sender, thereby reducing unnecessary PFM message forwarding.¶
When it comes to general PIM message security, see [RFC7761]. For PFM security see [RFC8364]. This optimization relies on correct Router-ID and capability advertisement in PIM Hellos, as well as general PIM hello integrity. For the new PFM TLV, the security considerations are the same as for the existing PFM TLV defined in [RFC8364].¶
All registries referenced in this section are under the "Protocol Independent Multicast (PIM) Parameters" registry group.¶
This document requires the assignment of a new PFM TLV Type TBD1 in the "PIM Flooding Mechanism Message Types" registry.¶
PIM Flooding Mechanism Message Types
Type Name Reference
-----------------------------------------------
TBD1 Group Source Info [This document]
¶
Also, a new registry "PIM Flooding Mechanism Group Source Info Sub-TLV Types" registry needs to be created. Assignments for the new registry are to be made according to the policy "IETF Review" as defined in [RFC8126]. The initial content of the registry should be:¶
PIM Flooding Mechanism
Group Source Info Sub-TLV Types
Type Name Reference
-----------------------------------------------
0-32767 Unassigned
¶
This document requires the assignment of two new PIM Hello Options:¶
PIM Hello Options
Value Length Name Reference
--------------------------------------------------
TBD2 0 GSI TLV support [This document]
TBD3 0 PFM Optimization [This document]
¶