TRILL Network Multi-destination Traffic Convergence via Inactive Tree Pruning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data center networks face challenges in efficiently managing multi-destination traffic, leading to time-synchronized churn in route computation and potential black-holing of end-to-end traffic due to frequent recomputation of multiple trees, which results in disruptions and performance issues.
Innovation Solution
The proposed method involves a communication system that receives and processes Type-Length-Value (TLV) messages to identify inactive multi-destination trees in a TRILL network, allowing for deferred route computation until a predetermined threshold of active trees is reached, thereby reducing network churn and maintaining smooth traffic convergence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If frequent recomputation of multiple trees is performed to manage multi-destination traffic, then route computation is updated, but network churn increases and traffic black-holing occurs
Solution Approach 1:
The system performs preliminary actions by pre-computing and caching route information for multiple destination trees before actual traffic needs arise. When traffic convergence is needed, the system retrieves pre-computed routes from cache rather than performing full recomputation, thereby reducing network churn and preventing traffic black-holing while maintaining convergence stability.
Solution Approach 2:
The patent creates copies of route information and tree structures in cache memory. Instead of recomputing entire routes, the system uses copied route data from previous computations, reducing the time and resources required for route updates while maintaining reliable traffic convergence across the network.
2Adaptability or versatility
If multiple trees are actively maintained for multi-destination traffic, then traffic routing is flexible, but network churn and disruptions increase
Solution Approach 1:
The system extracts and removes inactive or redundant trees from active maintenance. By identifying which trees are actually needed based on current traffic patterns and removing unnecessary trees from active management, the system reduces network churn and disruptions while maintaining routing flexibility for actual traffic demands.
Solution Approach 2:
The patent applies local quality by treating different trees differently based on their actual usage. Active trees that carry traffic are maintained with full flexibility, while inactive trees are pruned or put on hold. This selective approach maintains routing flexibility where needed while improving network stability overall by reducing unnecessary churn across the entire network.
3Measurement precision
If route computation is synchronized across the network, then consistent routing is achieved, but convergence disruptions occur
Solution Approach 1:
The system performs preliminary computation of routes independently across different network nodes before synchronization. Each node pre-computes its own routing table based on local information and cached data, then these pre-computed routes are synchronized. This approach achieves consistent routing while avoiding the disruptions of simultaneous recomputation across the entire network.
Data Source
AI summary
An example method for convergence of multi-destination traffic in a network environment is provided and includes receiving a first type-length-value (TLV) message from a true broadcast root in a Transparent Interconnection of Lots of Links (TRILL) network, where the first TLV message indicates a first subset of multi-destination trees in the TRILL network, receiving a second TLV message from the true broadcast root indicating a second subset of multi-destination trees in the TRILL network, such that a union of the first subset and the second subset indicates at least one inactive multi-destination tree in the TRILL network, and deleting the inactive tree from a hash table of active trees.