Distributed Spanning Tree Protocol Across Line Cards
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In layer 2 computer networks, the redundancy provided by interconnecting LAN segments using devices like switches can lead to unwanted loops, which negatively impact network performance by causing unnecessary resource consumption and traffic loops, and existing spanning tree protocols struggle to efficiently manage ports across multiple line cards in network devices.
Innovation Solution
Distributing spanning tree processing across multiple line cards within a network device, where each card executes its own instance of the spanning tree algorithm and communicates to determine a single superior port, allowing for efficient management and blocking of ports to prevent loops, while appearing as a single entity to other network devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If spanning tree protocol is executed on a single centralized processor managing all ports, then loop prevention is achieved, but device complexity and computational burden increase with the number of ports
Solution Approach 1:
The patent divides the spanning tree protocol execution into multiple independent instances, each running on a different line card. Each line card executes the spanning tree algorithm locally for its own ports, rather than having a single centralized processor manage all ports. This segmentation reduces the computational burden on any single processor and distributes the complexity across multiple cards.
Solution Approach 2:
The patent introduces a hierarchical dimension to spanning tree protocol execution by creating a relationship between line card level (local superior port determination) and network device level (superior port determination). This multi-level approach allows local processing while maintaining global coordination, effectively distributing complexity across different organizational dimensions.
2Productivity
If multiple line cards each execute spanning tree algorithm independently, then computational complexity is reduced, but coordination between cards becomes necessary
Solution Approach 1:
Each line card performs preliminary spanning tree calculations locally to determine its own superior local port before participating in the overall device-level superior port determination. This preliminary local processing reduces the amount of coordination needed, as each card arrives at the device level with already-processed information about its local port status.
Solution Approach 2:
The patent implements feedback mechanisms where line cards exchange information about their determined superior local ports with other line cards. This feedback loop allows the system to coordinate superior port determination across the device while maintaining the benefits of distributed processing, as each card adjusts its behavior based on information received from other cards.
Data Source
AI summary
Techniques are disclosed that may allow a network device comprising multiple line cards that are each executing a spanning tree algorithm to appear as a single spanning tree protocol entity to other network devices within a bridged local area network (LAN). In one example, a method includes determining, via a first processor on a first card of a network device, a first superior local port on the first card, transmitting, from the first card to a second card of the network device, a data unit specifying the determined first superior local port, and determining, via a second processor on the second card, a superior port on the network device among the first superior local port and a second plurality of ports on the second card based on a second plurality of BPDUs received via the second plurality of ports.