Dynamic Hash Index Mapping via Binary Trees for LAG Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network devices face inefficiencies in distributing traffic across link aggregation groups (LAGs) due to uneven utilization of member links, leading to over-utilization and dropped traffic, with conventional re-distribution techniques having high computational complexity.
Innovation Solution
The use of binary trees to dynamically re-assign hash indices and update the hash data structure based on net weights and utilization values, allowing for proportional allocation of traffic across member interfaces, reducing the occurrence of over-utilization and improving resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If conventional re-distribution techniques are used to balance traffic across member links, then traffic distribution uniformity is improved, but computational complexity increases to O(N*N)
Solution Approach 1:
The patent changes the computational approach from conventional O(N*N) re-distribution techniques to a binary tree-based method with O(M/N) complexity. By representing member links as binary tree nodes and using tree traversal algorithms, the system achieves traffic distribution balancing with significantly reduced computational complexity while maintaining distribution uniformity.
Solution Approach 2:
The patent replaces the conventional iterative re-distribution mechanism with a binary tree-based mapping system. Instead of repeatedly calculating and adjusting traffic distribution across N links (mechanical iteration), the system uses binary tree structure where each node represents a link and hash indices are mapped through tree traversal, substituting the mechanical re-distribution process with a more efficient computational model.
2Device complexity
If static hash index mapping is used, then device complexity is reduced, but resource utilization deteriorates due to over-utilization of certain member links
Solution Approach 1:
The patent introduces dynamics into the hash index mapping by using binary trees that can be reconfigured based on link utilization. The system monitors utilization metrics and dynamically adjusts the binary tree structure or repartitions hash indices among member links, transforming a static mapping into a dynamic adaptation mechanism that optimizes resource utilization while maintaining structural simplicity.
Solution Approach 2:
The binary tree-based system enables self-service traffic distribution by automatically monitoring link utilization and reconfiguring mappings without external intervention. The system uses feedback from utilization metrics to autonomously adjust hash index assignments, allowing the network device to self-optimize traffic distribution across member links.
3Adaptability or versatility
If member links have different weights and bandwidths, then system adaptability is improved, but traffic distribution uniformity deteriorates due to uneven utilization
Solution Approach 1:
The patent applies local quality by assigning different weights to binary tree nodes representing member links with different characteristics. The binary tree structure allows each node to have customized weight attributes, enabling the system to accommodate varying link bandwidths and weights while maintaining overall traffic distribution uniformity through weighted tree traversal and hash index mapping.
Data Source
AI summary
Techniques are described for dynamically mapping hash indices to member interfaces of an aggregated interface in a hash data structure. As one example, a network device may compute net weights for the member links of a link aggregation group (LAG) and respective utilization values of hash indices in a hash map for the LAG. The network device may generate binary trees based on the net weights and utilization values of the member links of the LAG, and may map values, e.g., median values, of one or more nodes of the binary trees to member interfaces, and update the hash data structure based on the mapping.