Switching Device MAC Address Masking for MCLAG Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In a PBB network with MCLAG switches, the learning information on address tables is unnecessarily altered when frames from the same customer terminal are received by different edge switching devices with different source encapsulation addresses, leading to instability and potential fault detections.
Innovation Solution
The implementation of a relay system with a learning information control unit that prevents the learning of correspondence relations between source customer addresses and source encapsulation addresses in cases where the difference lies only in a specific bit, maintaining consistent address table entries for the same customer address across different edge switching devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If frames from the same customer terminal are received by different edge switching devices with different source encapsulation addresses, then the address table learning function operates normally, but the learning information on address tables is unnecessarily altered leading to instability and potential fault detections
Solution Approach 1:
The patent applies parameter changes by modifying the MAC address comparison process to ignore specific bit positions (bits 0-3 of the encapsulation address) that differ between MCLAG peer devices. This allows the system to recognize that frames with different encapsulation addresses but identical customer MAC addresses should not trigger address table learning updates, thereby preventing unnecessary alterations while maintaining reliable address table stability.
2Adaptability or versatility
If MCLAG is set across two switching devices to achieve redundancy and expand communication band, then fault redundancy and communication expansion are achieved, but learning information instability occurs when frames are received through different devices
Solution Approach 1:
The patent applies local quality by making the MAC address comparison selective rather than uniform. Specifically, it compares only certain bits (ignoring bits 0-3) of the encapsulation address while comparing all bits of the customer MAC address. This localized differentiation allows the system to maintain stability for customer address learning while accommodating the variable encapsulation addresses used by MCLAG peer devices for redundancy and load balancing.
Solution Approach 2:
The patent modifies the comparison parameters by changing which bits of the MAC address are considered significant. By configuring the system to ignore specific bit positions in the encapsulation address that vary between MCLAG peers, it enables stable address table learning despite the presence of multiple encapsulation addresses from different switching devices in the MCLAG group.
3Loss of information
If the address table learns correspondence relations between source customer addresses and source encapsulation addresses, then complete forwarding information is maintained, but false fault detections occur due to unnecessary learning alterations
Solution Approach 1:
The patent applies parameter changes by modifying the learning trigger conditions. Instead of triggering address table learning updates whenever any bit of the source MAC address changes, the system is configured to ignore changes in specific bits (0-3) of the encapsulation address. This selective parameter change allows the address table to maintain complete forwarding information while preventing false fault detections caused by expected variations in MCLAG encapsulation addresses.
Data Source
AI summary
An encapsulation address of a first switching device is set so as to be different only in a mask bit in comparison with that of a second switching device. A learning information control unit of a third switching device does not learn the correspondence relation between a source customer address and a source encapsulation address in a first case and a second case. The first case corresponds to a case in which an encapsulated frame is received at an upper-link port and an encapsulation address corresponding to the source customer address is acquired from an address table. The second case corresponds to a case in which a difference between the source encapsulation address and the encapsulation address acquired from the address table lies only in the mask bit.


