VXLAN PE Device Link Restoration via Pre-negotiated Port Roles

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The complexity of establishing and rapidly restoring entry backup links between all-active gateways in VXLAN networks hinders rapid fault recovery and network reliability during packet transmission.

Innovation Solution

A packet processing method where PE devices with pre-negotiated port statuses are used to rapidly switch between active and standby links, employing E-Trunk and EVPN protocols to manage MAC entries and ensure consistency in packet forwarding, thereby enabling swift link restoration and maintaining network reliability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If an entry backup link is established between all-active gateways to synchronize ARP and MAC entries, then network reliability is ensured, but the process becomes complex and link restoration is slow

Engineering Contradiction:
Improvenetwork reliabilityVSAvoidlink establishment complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent combines the entry backup link with the existing MC-trunk link, merging two separate functions into a single unified link. This eliminates the need for separate backup link establishment while maintaining both entry synchronization and fault tolerance capabilities, thereby reducing complexity without compromising reliability

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The MC-trunk link is designed to serve multiple functions simultaneously: it acts as both the primary active link for packet forwarding and the backup link for entry synchronization. This multi-functionality eliminates the need for dedicated backup link infrastructure, simplifying the overall system while ensuring network reliability through rapid failover

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Reliability

If an entry backup link is established between all-active gateways, then ARP and MAC entries can be synchronized, but link restoration time increases affecting network reliability

Engineering Contradiction:
Improveentry synchronization reliabilityVSAvoidlink restoration time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent pre-configures the MC-trunk link with both active and standby port roles before any fault occurs. The standby port is pre-negotiated and ready to immediately take over when the active port fails, eliminating the need for complex post-fault link establishment and entry synchronization processes, thus reducing restoration time while maintaining entry consistency

Inventive Principle:
Principle #10Preliminary action

3Productivity

If all-active gateway devices use ECMP for load sharing, then network capacity is improved, but fault recovery complexity increases

Engineering Contradiction:
Improvepacket forwarding capacityVSAvoidfault recovery complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the MC-trunk link into distinct active and standby port roles, creating a clear hierarchical structure. This segmentation allows ECMP to operate on the logical link level for load sharing while the physical port roles provide simplified fault recovery paths, decoupling the complexity of ECMP from the fault recovery mechanism

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3531632B1Failure recovery in a network system
Publication Date: 2022.01.19 HUAWEI TECH CO LTD
  • EP3531632B1 patent drawingFigure 1
  • EP3531632B1 patent drawingFigure 2
  • EP3531632B1 patent drawingFigure 3

AI summary

This application provides a packet processing method, a device, and a network system. The network system includes a first PE device, a second PE device, a third PE device, and a CE device. A port that is of the first PE device and that is connected to the CE device is in an activated state, and a port that is of the second PE device and that is connected to the CE device is in a deactivated state. The method includes: when the port that is of the first PE device and that is connected to the CE device is switched from the activated state to the deactivated state, sending, by the first PE device, a MAC entry update packet to the third PE device, to trigger the third PE device to delete a first MAC entry whose packet forwarding destination address is the first PE device, generate, through learning, a second MAC entry whose packet forwarding destination address is the second PE device, and forward a packet based on the second MAC entry. Therefore, after a fault occurs, a link is rapidly restored to forward a packet, and reliability of a VXLAN network is ensured.