Network Topology Optimization for Balanced Throughput
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computer networks, particularly in data centers, face challenges in achieving balanced interconnectivity between superblocks and spineblocks, leading to variations in fairness and throughput, resulting in suboptimal network performance.
Innovation Solution
A method and system for configuring networks by assigning communication links between nodes in a data center's superblocks and spineblocks using a binary matrix, ensuring balanced interconnectivity by allocating interconnectivity slack links in a way that maximizes minimum throughput across all superblock pairs, using a topology analysis module and network configuration module to determine and disseminate optimal link configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If communication links are assigned between superblocks and spineblocks without optimization, then network connectivity is established, but throughput uniformity and fairness deteriorate
Solution Approach 1:
The patent applies parameter changes by optimizing the link assignment matrix parameters to achieve balanced throughput. Specifically, it adjusts the number of links between superblocks and spineblocks using mathematical optimization to ensure uniform maximum throughput across all superblock pairs, transforming the network topology parameters to resolve the contradiction between connectivity and throughput uniformity.
Solution Approach 2:
The patent employs preliminary action by pre-calculating and pre-assigning the optimal link configuration between superblocks and spineblocks before the network operates. The system computes the optimal assignment matrix in advance and configures the physical links accordingly, eliminating the need for dynamic adjustment and ensuring throughput uniformity from the outset.
2Reliability
If interconnectivity slack links are allocated to maximize minimum throughput, then network fairness improves, but configuration complexity increases
Solution Approach 1:
The patent changes the allocation parameters of interconnectivity slack links by formulating an optimization problem that maximizes the minimum throughput across all superblock pairs. It adjusts the link assignment matrix parameters to achieve fair throughput distribution, resolving the contradiction between network fairness and configuration complexity through mathematical optimization.
Solution Approach 2:
The patent uses a binary matrix representation to model and solve the link assignment problem. This matrix copying approach allows the system to work with an abstract representation of the network topology, compute the optimal configuration, and then replicate this configuration across the physical network, simplifying the management of complexity while achieving fairness.
3Speed
If optical transmission technologies are used for high-speed data links, then transmission speed improves, but network interconnectivity balance deteriorates
Solution Approach 1:
The patent addresses the contradiction by optimizing the topological parameters of the network rather than the physical transmission medium parameters. It adjusts the logical link assignment matrix to ensure balanced interconnectivity while maintaining high-speed optical transmission, separating the speed capability from the balance requirement and resolving the contradiction through topology optimization.
Data Source
AI summary
Systems and methods of configuring a computer network are provided. The network can include a first stage of M nodes and a second stage of N nodes. Each node in the first stage can form F communication links with nodes in the second stage and each node in the second stage can form S communication links with nodes in the first stage. Nodes in the first stage are connected to each other only through communication links connecting nodes in the first stage to nodes in the second stage. Communication links between the nodes can be assigned such that the minimum value of the throughput between any pair of nodes in the first stage is maximized.