MC-LAG BFD Sub-Second Failure Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current approaches for link failure detection in Multi-Chassis Link Aggregation Groups (MC-LAG) using Link Aggregation Control Protocol (LACP) have a minimal timer of 1 second, resulting in a convergence time of at least 3 seconds, which is inadequate for fast failure detection.
Innovation Solution
The method involves transmitting Multiprotocol Label Switching (MPLS) encapsulated packets with a generic associated channel header and a bidirectional forwarding detection (BFD) control packet over MC-LAG links, enabling sub-second link failure detection by establishing micro-BFD sessions and using MPLS encapsulation to communicate BFD control packets as a single aggregation system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If LACP is used for link failure detection in MC-LAG, then the system can detect link failures, but the convergence time is at least 3 seconds due to the minimal timer of 1 second
Solution Approach 1:
The patent changes the detection parameter from LACP's minimal 1-second timer to BFD's configurable sub-second timers (e.g., 50ms, 100ms). By implementing micro-BFD sessions with shorter transmission intervals, the system achieves faster link failure detection while maintaining reliability, directly resolving the contradiction between detection capability and convergence time.
Solution Approach 2:
The patent introduces MPLS encapsulation as an intermediary mechanism to transport BFD control packets over MC-LAG links. This intermediary approach allows BFD packets to be properly routed through the MC-LAG infrastructure without disrupting the existing LAG configuration, enabling fast failure detection while maintaining system stability.
2Speed
If micro-BFD sessions are established over each LAG member link, then fast failure detection is achieved, but the complexity of implementing BFD over MC-LAG increases
Solution Approach 1:
The patent makes the MC-LAG system universally compatible with both LACP and BFD protocols. By implementing micro-BFD sessions that operate alongside existing LAG member links, the system gains fast failure detection capability without requiring separate dedicated infrastructure, thus reducing implementation complexity while maintaining multi-protocol functionality.
Solution Approach 2:
The patent segments the failure detection function into independent micro-BFD sessions for each LAG member link. This segmentation allows each session to operate autonomously with its own timer and state machine, simplifying the overall implementation by breaking down the complex task of MC-LAG failure detection into manageable, independent units that can be configured and managed separately.
Data Source
AI summary
Methods and apparatuses for enabling sub-seconds link failure detection in a multi-chassis link aggregation group (MC-LAG) system are described. A first network device of a packet network transmits a Multiprotocol Label Switching (MPLS) encapsulated packet over a first link that is part of the MC-LAG, where the MC-LAG couples the first network device with a second network device and a third network device and the second and third network devices are part of an inter-chassis redundancy (ICR) system. The MPLS encapsulated packet includes a generic associated channel header (ACH) and a payload, and where the payload includes a bidirectional forwarding detection (BFD) control packet.