| Internet-Draft | BGP BFD Strict-Mode | January 2024 |
| Zheng, et al. | Expires 6 July 2024 | [Page] |
This document specifies extensions to RFC4271 BGP-4 that enable a BGP speaker to negotiate additional Bidirectional Forwarding Detection (BFD) extensions using a BGP capability. This BFD Strict-Mode Capability enables a BGP speaker to prevent a BGP session from being established until a BFD session is established. It is referred to as BGP BFD "strict-mode". BGP BFD strict-mode will be supported when both the local speaker and its remote peer are BFD strict-mode capable.¶
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 6 July 2024.¶
Copyright (c) 2024 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.¶
Bidirectional Forwarding Detection BFD [RFC5882] enables routers to monitor data plane connectivity and to detect faults in the bidirectional forwarding path between them. This capability is leveraged by routing protocols such as BGP [RFC4271] to rapidly react to topology changes in the face of path failures.¶
The BFD interaction with BGP is specified in Section 10.2 of [RFC5882]. When BFD is enabled for a BGP neighbor, faults in the bidirectional forwarding detected by BFD result in session termination. It is possible in some failure scenarios for the network to be in a state such that a BGP session may be established but a BFD session cannot be established. In some other scenarios, it may be possible to establish a BGP session, but a degraded or poor-quality link may result in the corresponding BFD session going up and down frequently.¶
To avoid situations which result in routing churn and to minimize the impact of network interruptions, it will be beneficial to disallow BGP to establish a session until BFD session is successfully established and has stabilized. We refer to this mode of operation as BGP BFD "strict-mode". However, always using "strict-mode" would preclude BGP operation in an environment where not all routers support BFD strict-mode or have BFD enabled. This document defines BGP "strict-mode" operation as preventing BGP session establishment until both the local and remove speakers have a stable BFD session. The document also specifies the BGP protocol extensions for BGP capability [RFC5492] for announcing BFD parameters including a BGP speaker's support for "strict-mode", i.e., requiring a BFD session for BGP session establishment.¶
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.¶
The BGP Strict-Mode Capability [RFC5492] will allow a BGP speaker's to advertise this capability. The capability is defined as follows:¶
Capability code: 74¶
Capability length: 0 octets¶
A BGP speaker which supports capabilities advertisement and has BFD strict-mode enabled MUST include the BFD Strict-Mode Capability.¶
A BGP speaker which supports the BFD Strict-Mode Capability, examines the list of capabilities present in the capabilities that the speaker receives from its peer. If both the local and remote BGP speakers include the BFD Strict-Mode Capability, the BGP finite state machine does not transition to the Established state from OpenConfirm state [RFC4271] until the BFD session is in the Up state (see below for AdminDown state). This means that a KEEPALIVE message is not sent nor is the KeepaliveTimer set.¶
If the BFD session does not transition to the Up state, and the HoldTimer has been negotiated to a non-zero value, the BGP FSM will close the session appropriately. If the HoldTimer has been negotiated to a zero value, the session should be closed after a time of X. This time X is referred as "BGP BFD Hold time". The proposed default BGP BFD Hold time value is 30 seconds. The BGP BFD Hold time value is configurable.¶
If BFD session is in the AdminDown state, then the BGP finite state machine will proceed normally without input from BFD. This means that BFD session "AdminDown" state WILL NOT prevent the BGP state transition from the OpenConfirm state to the Established state.¶
Once the BFD session has transitioned to the Up state, the BGP FSM may proceed to transition from the OpenConfirm state to state Established state. Once in the Established state, a KEEPALIVE message is sent and a KeepaliveTimer for the BGP peer is started.¶
BGP strict-mode cannot be enabled unless BFD is configured for the BGP peer. If BFD is removed for the BGP peer, then BGP strict-mode will also be disabled.¶
Note that it is fully possible to have BFD enabled between the peers without BGP strict-mode.¶
If either BGP peer has not advertised the BFD Strict-Mode Capability, then a BFD session WILL NOT be required for the BGP session to reach Established state. This does not preclude usage of BFD after BGP session establishment [RFC5882].¶
If BFD strict-mode is enabled or disabled for a BGP peer and the BGP session state is not Established state, then the BGP will close the session.¶
If BFD strict-mode is enabled or disabled for a BGP peer and the BGP session state is in the Established state, the local BFD strict-mode configuration will be modified but the session will remain in Established state.¶
Since BFD strict-mode is only applicable during BGP session establishment, this inconsistency would not have an impact on the Established session unless the remote BGP peer is waiting in OpenConfirm state. To avoid this situation, BFD strict-mode SHOULD be modified consistently on both the local and remote BGP peers.¶
When BGP sessions are closed according to the procedures in this document, the session SHOULD be terminated with a NOTIFICATION message with the Cease Code and the "BFD Down" Subcode. ([I-D.ietf-idr-bfd-subcode]) This informs the operator that interaction with BFD is the root cause of the BGP session being unable to move to the Established state.¶
The use of BGP BFD strict-mode along with mechanisms such as hold-down (a delay in the initial BGP Establishment state following BFD session establishment) and/or dampening (a delay in the BGP Establishment state following failure detected by BFD) may help reduce the frequency of BGP session flaps and therefore reduce the associated routing churn.¶
To avoid deadlock when utilizing both BFD hold-down and BGP BFD strict-mode, when strict-mode is enabled for a peer, the BGP FSM MUST be enabled. That is, BFD hold-down procedures MUST NOT prevent BGP from establishing a connection with the remote BGP speaker.¶
If both the local and remote BGP speakers include the BFD Strict-Mode Capability, the BGP state machine is permitted to transition to the Established state from the OpenConfirm state after the locally configured BFD hold-down interval is observed. That is, the BFD session has been Up for the desired amount of time. Implementations MAY start the BFD session associated with the BGP bfd-strict session prior to the BGP FSM starting.¶
It is RECOMMENDED that the BFD hold-down intervals used with BFD strict-mode, when configured, use similar values. Similarly, the negotiated BGP holdtime SHOULD be long enough to account for the time between the BGP FSM reaching the OpenConfirm state, the BFD hold-down interval, and any delay for the BFD session being initiated. Failure to do so can result in the BGP speaker that has transitioned to the Established state expiring its BGP holdtime and closing the connection. This is because the remote BGP speaker hasn't transitioned to Established and begun sending KEEPALIVE messages.¶
A BGP speaker SHOULD log a message if it closes its session due to hold timer expiration while waiting for the BFD hold-down interval.¶
The behavior of BGP speakers implementing BFD hold-down without negotiating the BFD strict-mode feature is out of scope of this document. However, the authors are aware that inconsistent behaviors in BGP implementations for BFD hold-down without BGP BFD strict-mode may result in BGP session deadlock.¶
Auto-configuration is possible for the enabling BGP BFD strict-mode. However, the configuration automation is out of the scope of this document.¶
To simplify troubleshoorting and avoid inconsistencies, it is RECOMMENDED that BFD strict-mode configuration be consistent for both BGP peers.¶
The mechanism defined in this document interacts with the BGP finite state machine when so configured. The security considerations of BFD thus, become considerations for BGP-4 [RFC4271] so used. Given that a BFD session is required for a BGP session, a Denial-of-Service (DoS) attack on BGP can now be mounted by preventing a BFD session between the BGP peers from being established or interrupting an existing BFD session. The use of the BFD Authentication mechanism defined in [RFC5880] is thus RECOMMENDED when used to protect BGP-4 [RFC4271].¶
This document defines the BGP BFD Strict-Mode Capability. The Capability Code 74 has been assigned from the First-Come-First-Served range (64-238) of the Capability Codes registry.¶
The authors would like to acknowledge the review and inputs from Shyam Sethuram, Mohammed Mirza, Bruno Decraene, Carlos Pignataro, Enke Chen, and Anup Kumar.¶