BFD Session Equivalence Class Queue for Failure Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In scenarios where a public network path experiences gradual signal attenuation, Bidirectional Forwarding Detection (BFD) sessions may report inconsistent states, leading to delayed Down event reporting and subsequent failure processing delays due to the existing detection mechanisms' inability to accurately detect network path failures in a timely manner.

Innovation Solution

A method that utilizes an equivalence class time-window queue to aggregate BFD session events, where a Down event is reported when multiple sessions within a set time-window indicate a failure, ensuring timely detection and reducing reporting delays by managing queue elements and equivalence class sets to avoid repeated notifications.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional BFD session detection mechanism is used, then failure detection speed is fast (milliseconds), but in scenarios of gradual signal attenuation, inconsistent state reporting leads to delayed Down event notification and failure processing delay

Engineering Contradiction:
Improvefailure detection accuracyVSAvoidDown event reporting delay
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent merges multiple BFD session states belonging to the same equivalence class into a unified detection result. When multiple sessions in an equivalence class report Down events, the system aggregates these states and generates a single authoritative Down event notification, eliminating the delay caused by processing individual inconsistent session states separately.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces an equivalence class state machine as an intermediary layer between individual BFD sessions and the failure detection system. This intermediary aggregates session states, resolves inconsistencies through state transitions, and produces unified Down event notifications, thereby solving the timing conflict between fast detection and accurate failure declaration.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If multiple BFD sessions are monitored independently, then comprehensive network path coverage is achieved, but inconsistent state reporting occurs leading to processing delays

Engineering Contradiction:
Improvenetwork path coverageVSAvoidstate management complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments BFD sessions into equivalence classes based on shared network paths. Each equivalence class is managed as an independent unit with its own state machine, allowing comprehensive network coverage while simplifying state management through modular organization. This segmentation reduces the complexity of managing all sessions globally by localizing state transitions within each equivalence class.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS11206176B2Preventing failure processing delay
Publication Date: 2021.12.21 NEW H3C TECH CO LTD
  • US11206176B2 patent drawing
  • US11206176B2 patent drawing
  • US11206176B2 patent drawing

AI summary

A method and device for preventing a failure processing delay are provided in the disclosure. In an example, when the number of queue elements in an equivalence class time-window queue reaches a set threshold (denoted as N) in a set time-window, it means that there are N Bidirectional Forwarding Detection (BFD) sessions in the same equivalence class set, that detect Down events. It thus can be intelligently inferred that a public network path carrying the N BFD sessions breaks down. For processing a failure in time and reducing data stream loss on an upper layer, the present disclosure may allow reporting a corresponding Down event for each BFD session in the equivalence class set to which the N BFD sessions belong.