BFD Echo Service for Link Aggregation Failover

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

During a control plane failover in a network element, BFD sessions are disrupted, leading to false negatives where other network elements believe the failover network element is down, despite its data plane still functioning, causing service disruptions and network churn.

Innovation Solution

Implementing a fault detection service that continues to send fault detection packets on the link during a supervisory failover, using a BFD echo service to survive control plane failures and saving the fault detection state in a standby supervisor to resume the service after failover, ensuring the link is not incorrectly marked as down.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If BFD sessions are processed by the control plane during supervisor failover, then fault detection can be performed, but the BFD sessions are disrupted causing false negatives and service disruptions

Engineering Contradiction:

Improvefault detection accuracyVSAvoidcontrol plane failover time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent segments the fault detection function into two independent parts: control plane BFD sessions and data plane BFD echo packets. The data plane continues to process BFD echo packets independently during control plane failover, preventing disruption of fault detection while the control plane recovers.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary mechanism where the data plane acts as a mediator for BFD echo packet processing. This intermediary continues to forward and process BFD packets even when the control plane is unavailable, bridging the gap during failover transitions.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Adaptability or versatility

If the control plane fails over from active to standby CPU, then system redundancy is maintained, but the BFD sessions are disrupted leading to false fault detection

Engineering Contradiction:

Improvesupervisor redundancyVSAvoidlink status indication accuracy

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent separates control plane functions from data plane functions for BFD processing. The data plane independently handles BFD echo packets, maintaining accurate link status indication during control plane supervisor failover, while the control plane transitions between active and standby states.

Inventive Principle:

Principle #1Segmentation

3Difficulty of detecting and measuring

If BFD packets are not received during control plane failover, then fault detection mechanism works as designed, but it incorrectly identifies the network element as down

Engineering Contradiction:

Improvefault detection sensitivityVSAvoidnetwork element status accuracy

Core Design Contradiction:

Difficulty of detecting and measuringVSReliability

Solution Approach 1:

The data plane serves as an intermediary that continues to process and respond to BFD echo packets during control plane failover. This ensures BFD packets are received and responded to correctly, preventing false fault detection while maintaining the sensitivity of the fault detection mechanism.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS10630564B2System and method of handling a fault detection mechanism with link aggregation groups

Publication Date: 2020.04.21 ARISTA NETWORKS INC

AI summary

A method and apparatus of a network element that configures a fault detection service on a multi-link group of a network element is described. In this embodiment, the network element starts the fault detection service on a first link of the multi-link group. The network element further configures a first packet for the fault detection service and transmits the first packet on the first link. The network element additionally receives the first packet and determines if the first packet was received on the first link. If the first packet was received on a different link, the network element configures a second packet by modifying a packet characteristic of the first packet and transmits the second packet on the first link. If the first packet was received on the first link, the network element marks the fault detection service for the first link as converged.