Active-Active Cluster Forwarding Rule Synchronization via Control Node

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In active-active clusters, the unreliable synchronization of forwarding rules between network devices leads to packet loss and communication service interruptions, as nodes may not receive inbound forwarding rules in time to process response packets correctly.

Innovation Solution

Implementing a control method and a control node that stores forwarding rules independently, allowing each node in the active-active cluster to obtain rules from the control node as needed, ensuring reliable synchronization and reducing the need for rule storage on each node.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the first network processing node sends the inbound forwarding rule to all other network processing nodes in the active-active cluster, then the forwarding rules are synchronized across nodes, but the synchronization reliability is poor and packet loss occurs when nodes do not receive rules in time

Engineering Contradiction:
Improveforwarding rule synchronization reliabilityVSAvoidpacket loss
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The patent introduces a control node as an intermediary between network processing nodes. The control node receives forwarding rules from the first network processing node and distributes them to the second network processing node, ensuring reliable rule synchronization. This mediator approach eliminates the unreliable direct peer-to-peer synchronization between cluster nodes.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The control node pre-establishes a database for storing forwarding rules and maintains readiness to distribute rules before they are needed. When the second network processing node needs a forwarding rule, the control node can immediately provide it, preventing packet loss due to timing issues in rule synchronization.

Inventive Principle:
Principle #10Preliminary action

2Ease of operation

If each network processing node stores all forwarding rules locally, then nodes can process packets independently, but the device complexity and storage requirements increase

Engineering Contradiction:
Improveindependent packet processing capabilityVSAvoidrule storage complexity
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The patent segments the forwarding rule storage function from the network processing nodes and places it in a dedicated control node. Each network processing node only stores or caches rules temporarily needed for processing, while the control node maintains the complete set of forwarding rules in its database, reducing complexity at processing nodes.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The control node acts as a centralized repository and distributor of forwarding rules, allowing network processing nodes to maintain simpler local storage while still having access to all necessary rules through the intermediary control node.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Productivity

If the cluster uses unicast or multicast synchronization for forwarding rules, then rule distribution is implemented, but the synchronization is unreliable and service interruptions occur

Engineering Contradiction:
Improverule distribution efficiencyVSAvoidsynchronization reliability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The control node serves as a reliable intermediary that manages the distribution of forwarding rules to cluster nodes. It receives rules from the first node and ensures they are properly delivered to the second node, replacing the unreliable unicast/multicast direct synchronization between cluster nodes.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements feedback mechanisms where the control node tracks which nodes have received which forwarding rules and can request retransmission if needed, ensuring reliable synchronization without service interruptions.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3965377B1Control method for main master cluster and control node
Publication Date: 2025.02.19 HUAWEI CLOUD COMPUTING TECHNOLOGIES CO LTD
  • EP3965377B1 patent drawingFigure 1~2
  • EP3965377B1 patent drawingFigure 3A
  • EP3965377B1 patent drawingFigure 3B

AI summary

This application provides an active-active cluster control method. The method includes: A control node receives a first query request sent by a first network processing node in an active-active cluster, configures an outbound forwarding rule based on forwarding information that is of a first packet and that is carried in the first query request, generates an inbound forwarding rule whose direction is opposite to that of the outbound forwarding rule, and sends the outbound forwarding rule to the first network processing node. The control node may further receive a second query request sent by a second network processing node in the active-active cluster, determine that forwarding information of a second packet matches the inbound forwarding rule, obtain, based on the forwarding information of the second packet, the recorded inbound forwarding rule, and send the inbound forwarding rule to the second network processing node. Technical solutions provided in this application can implement reliable synchronization of forwarding rules between network devices in the active-active cluster, thereby avoiding problems such as a packet loss and service interruption.