LSR Extensions for BIER over Ethernet

Bit Index Explicit Replication (BIER) is an architecture that provides multicast forwarding through a "BIER domain" without requiring intermediate routers to maintain multicast related per-flow state. BIER can be supported in MPLS and non-MPLS networks. specifies a common BIER header format for both MPLS and non-MPLS networks, though the first 20-bits of the BIER header (referred as BIFT-id) is a "MPLS Label" in case of MPLS networks and is a "domain-wide-unique-value" representing the combination of SD-BSL-SI in case of non-MPLS networks. specifies two optional ways of statically assigning domain-wide-unique mapping between BIFT-IDs and SD-BSL-SI combination. However, BIER architecture [RFC8279] does NOT require domain-wide-unique BIFT-IDs to be used (even for non-MPLS encapsulation). As discussed in , the BIFT-ID in case of non-MPLS encapsulation can also just be a local 20-bit opaque value and signaled just like in MPLS case. As an example, suppose a particular BIER domain contains a SD (SD 0), supports two BSLs (256 and 512), and contains 1024 BFRs. A BFR that is provisioned for above SD, and that supports both BSLs, could advertise the following set of BIFT-id's: BIFT-id 1: corresponding to SD 0, BSL 256, SI 0. BIFT-id 2: corresponding to SD 0, BSL 256, SI 1. BIFT-id 3: corresponding to SD 0, BSL 256, SI 2. BIFT-id 4: corresponding to SD 0, BSL 256, SI 3. BIFT-id 5: corresponding to SD 0, BSL 512, SI 0. BIFT-id 6: corresponding to SD 0, BSL 512, SI 1. Notice that the example uses ranges of continuous BIFT-ids: BIFT-id range [1 to 4] correspond to <SD 0, BSL 256>. The first BIFT-id in the range correspond to SI=0, the second correspond to SI=1, and so on. BIFT-id range [5 to 6] correspond to <SD 0, BSL 512>. The first BIFT-id in the range correspond to SI=0, the second correspond to SI=1. Strictly speaking, using contiguous range is not required, but it is done for the purpose of simplified signaling similar to MPLS label blocks (notice that locally assigning BIFT-ID ranges requires no manual processing just like in the case of MPLS label block allocation). Processing and forwarding of BIER packets requires special software and hardware capabilities. The BFRs supporting a BIER encapsulation type MUST advertise this capability (along with the other required parameters specific to the encapsulation) to the other routers in BIER domain. This advertisement, for example, will enable the other BFRs in the BIER domain in deciding, whether to include or exclude the advertising router from the BAR and/or IPA algorithm while computing the multicast path for a specific encapsulation type. and specifies the required extensions to the IS-IS and OSPFv2 protocol respectively for the distribution of BIER sub-domain information including the Sub-sub-TLV required to support BIER in MPLS encapsulation for MPLS networks. This document specifies the required extensions to the IS-IS and OSPFv2 protocol for supporting BIER using BIER in Ethernet encapsulation with dynamically and locally assigned BIFT-IDs. Support for other encapsulation types are outside the scope of this document.

Some of the terminology specified in is replicated here and extended by necessary definitions: Bit Index Explicit Replication (The overall architecture of forwarding multicast using a Bit Position). BIER in MPLS encapsulation. (Encapsulation of BIER header inside MPLS header in MPLS networks). BIER in Ethernet encapsulation. (Encapsulation of BIER header inside Ethernet header (EtherType=0xAB37) in non-MPLS networks). Bit Forwarding Router (A router that participates in Bit Index Multipoint Forwarding). A BFR is identified by a unique BFR- prefix in a BIER domain. Bit Index Forwarding Table used to forward the BIER packets in a domain. BIER Algorithm. Used to calculate underlay nexthops as defined by the BAR value. IGP Algorithm. May be used to modify, enhance or replace the calculation of underlay paths as defined by the BAR value

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.

A BIER sub-domain MAY support multiple BIER encapsulation types like BIER-MPLS, BIER-ETH. The different encapsulation types supported by a BFR in a sub-domain MUST share the same BFR-id. This would allow the BFR's in transit to translate the encapsulation from one type to the other while forwarding the packet in the BIER sub-domain. When a BFIR/BFR supports multiple BIER encapsulation types, when sending to a BIER neighbor it MUST use a type that the neighbor also supports. If the neighbor also supports more than one encapsulation type that this BFIR/BFR supports, the type selection could be a matter of local policy and is outside the scope of this document.

BIER Info sub-TLV defined in is used to advertise the sub-domain id, and other associated parameters of the sub-domain like BFR-id, MT, BAR, IPA. This document introduces new sub-sub-TLV under BIER Info sub-TLV to advertise the ethernet encapsulation capability and other associated parameters of the encapsulation. This sub-sub-TLV carries the information for the BIER Ethernet encapsulation including the BitString length supported for a certain <MT,SD> pair. It is advertised within the BIER Info sub-TLV defined in which in-turn is carried within the TLVs 235, 237 or TLVs 135 , or TLV 236 . This sub-sub-TLV MAY appear multiple times within a single BIER Info sub-TLV. If the same BitString length is repeated in multiple BIER Ethernet encapsulation sub-sub-TLVs inside the same BIER Info sub-TLV, the BIER Info sub-TLV MUST be ignored.

2 (suggested value - To be assigned by IANA). 4 A 1 octet field encoding the Maximum Set Identifier (Section 1 of [RFC8279]) used in the encapsulation for this BIER subdomain for this BitString length. The first BIFT-id is for SI=0, the second BIFT-id is for SI=1, etc. If the BIFT-id associated with the Maximum Set Identifier exceeds the 20-bit range, the sub-sub-TLV MUST be ignored. A 4 bit field encoding the bitstring length (as per ) supported for the encapsulation. A 20 bit field encoding the first BIFT-id of the BIFT-id range. The "BIFT-id range" is the set of 20-bit values beginning with the BIFT-id and ending with (BIFT-id + (Max SI)). A unique BIFT-id range is allocated for each BitString length and sub-domain-id. These BIFT-id's are used for BIER forwarding as described in and . The size of the BIFT-id range is determined by the number of SI's (Section 1 of ) that are used in the network. Each SI maps to a single BIFT-id in the BIFT-id range: the first BIFT-id is for SI=0, the second BIFT-id is for SI=1, etc. If the BIFT-id associated with the Maximum Set Identifier exceeds the 20-bit range, the BIER Ethernet Encapsulation Sub-sub-TLV containing the error MUST be ignored.

BIER Sub-TLV defined in is used to advertise the sub-domain id, and other associated parameters of the sub-domain like BFR-id, MT, BAR, IPA. This document introduces new Sub-TLV under BIER Sub-TLV to advertise the ethernet encapsulation capability and other associated parameters of the encapsulation. This Sub-TLV carries the information for the BIER Ethernet encapsulation including the BitString length supported for a certain <MT,SD> pair. It is advertised within the BIER Sub-TLV defined in which in-turn is carried within the OSPFv2 Extended Prefix TLV defined in . This Sub-TLV MAY appear multiple times within a single BIER Sub-TLV. If the same BitString length is repeated in multiple BIER Ethernet encapsulation Sub-TLVs inside the same BIER Sub-TLV, the BIER Sub-TLV MUST be ignored.

11 (suggested value - To be assigned by IANA). 8 A 1 octet field encoding the Maximum Set Identifier (Section 1 of [RFC8279]) used in the encapsulation for this BIER subdomain for this BitString length. The first BIFT-id is for SI=0, the second BIFT-id is for SI=1, etc. If the BIFT-id associated with the Maximum Set Identifier exceeds the 20-bit range, the sub-sub-TLV MUST be ignored. A 3-octet field, where the 20 rightmost bits represent the first BIFT-id in the BIFT-id range. The 4 leftmost bits MUST be ignored. The "BIFT-id range" is the set of 20-bit values beginning with the BIFT-id and ending with (BIFT-id + (Max SI)). A unique BIFT-id range is allocated for each BitString length and sub-domain-id. These BIFT-id's are used for BIER forwarding as described in and . The size of the BIFT-id range is determined by the number of SI's (Section 1 of ) that are used in the network. Each SI maps to a single BIFT-id in the BIFT-id range: the first BIFT-id is for SI=0, the second BIFT-id is for SI=1, etc. If the BIFT-id associated with the Maximum Set Identifier exceeds the 20-bit range, the BIER Ethernet Encapsulation Sub-sub-TLV containing the error MUST be ignored. A 4 bit field encoding the bitstring length (as per ) supported for the encapsulation. SHOULD be set to 0 on transmission and MUST be ignored on reception.

Security concerns for IS-IS are addressed in and and the security concerns for IS-IS extensions for BIER are addressed in . This document introduces new sub-sub-TLV for the already existing IS-IS TLVs defined for distributing the BIER sub-domain information in . It does not introduce any new security risks to IS-IS. Security concerns and required extensions for OSPFv2 are addressed in and and the security concerns for OSPFv2 extensions for BIER are addressed in . This document introduces new Sub-TLV for the already existing OSPFv2 TLV defined for distributing the BIER sub-domain information in . It does not introduce any new security risks to OSPFv2.

The document requests new allocations from the IANA registries as follows

BIER Ethernet Encapsulation sub-sub-TLV: 2 (suggested)

BIER Ethernet Encapsulation Sub-TLV: 11 (suggested)

The author wants to thank Antonie Przygienda for his comments and suggestions.