rfc-5543

BGP Traffic Engineering Attribute

H. Ould-Brahim, D. Fedyk, Y. Rekhter
date2009-05 streamIETF areaint wgsoftwire statusPROPOSED STANDARD pages6 canonicalhttps://www.rfc-editor.org/rfc/rfc5543 doi10.17487/RFC5543
This document defines a new BGP attribute, the Traffic Engineering attribute, that enables BGP to carry Traffic Engineering information. The scope and applicability of this attribute currently excludes its use for non-VPN reachability information. [STANDARDS-TRACK]

updated by

Extracted elements (14)

design-rationale §3

The attribute reuses the Interface Switching Capability Descriptor format from OSPF-GMPLS (RFC 4203) and IS-IS-GMPLS (RFC 5307) extensions rather than defining a new encoding, enabling consistent TE information representation across IGP and BGP signaling.

bgp, mpls, routing

design-rationale §1

The scope of the TE attribute is explicitly limited to VPN reachability information, and procedures for modifying the attribute when re-advertising a route are explicitly declared out of scope, deferring those problems to future work.

bgp, routing

interoperability-note §5

Use of the TE attribute may increase the number of BGP Update messages required: when routes share all other attributes but differ in TE attributes, they must be distributed in per-route BGP Update messages rather than batched in a single message.

bgp, routing

interoperability-note §4

When aggregating routes with identical Traffic Engineering attributes, the procedure of RFC 4201 (Link Bundling in MPLS TE) MUST be followed to construct the resulting Traffic Engineering attribute.

bgp, mpls, routing

normative-requirement §4 MUST NOT

Routes carrying the Traffic Engineering attribute SHALL NOT be aggregated unless they share identical Traffic Engineering attributes, because these routes have additional semantics that could affect traffic-forwarding behavior.

bgp, routing

normative-requirement §3 MUST

The Reserved field SHOULD be set to 0 on transmit and MUST be ignored on receive.

bgp, routing

protocol-element §3

For L2SC (Layer-2 Switch Capable) and LSC (Lambda Switch Capable) switching capability values, no Switching Capability specific information field is present in the descriptor.

bgp, mpls, routing

protocol-element §3

Maximum LSP Bandwidth is encoded as eight 4-octet IEEE 754 floating-point fields, one per priority level (0 through 7), with priority 0 first. Units are bytes per second (not bits).

bgp, mpls, routing

protocol-element §3

The Traffic Engineering attribute is an optional, non-transitive BGP attribute that carries Traffic Engineering information identical to the Interface Switching Capability Descriptor defined in RFC 4203 and RFC 5307. Its current scope is limited to VPN reachability information only.

bgp, mpls, routing

registry §6

This document allocates a new BGP Path Attribute code point for the Traffic Engineering attribute (optional, non-transitive) from the IANA BGP Path Attributes registry.

bgp, registry

security-consideration §7

This BGP extension does not change the underlying security issues inherent in BGP; BGP security considerations from RFC 4271 apply without modification.

bgp, security

wire-format §3

For PSC-1/2/3/4 switching capability: the Switching Capability specific information includes a 4-octet IEEE floating-point Minimum LSP Bandwidth (bytes/sec) followed by a 2-octet Interface MTU integer.

bgp, mpls, routing

wire-format §3

For TDM switching capability: the Switching Capability specific information includes a 4-octet IEEE floating-point Minimum LSP Bandwidth (bytes/sec) followed by a 1-octet Indication field (0 = Standard SONET/SDH, 1 = Arbitrary SONET/SDH).

bgp, mpls, routing

wire-format §3

The TE attribute contains one or more descriptors with: 1-octet Switching Cap, 1-octet Encoding, 2-octet Reserved, eight 4-octet IEEE floating-point Max LSP Bandwidth fields (priorities 0-7, in bytes/sec), followed by variable-length Switching Capability specific information.

bgp, mpls, routing