Multi-Network Packet Routing via Mesh and Tree Topologies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data processing systems face inefficiencies in packet transmission due to the trade-offs between mesh and tree networks, where mesh networks offer high throughput but long latency, and tree networks provide low latency but low throughput, especially as the number of processing units increases.
Innovation Solution
A multi-network system that selectively injects packets into either a mesh network or a tree network based on hop-count differences, latency measurements, and network congestion, using a network interface to determine the most efficient path and adjust filtering ratios to optimize packet transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a mesh network is used for packet transmission, then packet throughput is improved, but latency increases
Solution Approach 1:
The patent segments the packet transmission path by introducing an intermediate node between the source processing unit and the mesh network. This intermediate node receives packets from the source, determines optimal routing based on hop-count calculations, and selectively injects packets into either the mesh network or tree network. This segmentation allows the system to combine advantages of both network topologies: the mesh network provides high throughput for direct connections, while the tree network provides lower latency for connections requiring intermediate routing, thereby resolving the contradiction between throughput and latency.
2Loss of time
If a tree network is used for packet transmission, then latency is reduced, but packet throughput decreases
Solution Approach 1:
The patent implements dynamic packet routing by having the intermediate node calculate hop-counts for both mesh and tree networks and make real-time decisions about packet injection based on these calculations. The system dynamically adjusts the routing path for each packet based on the destination processing unit and current network conditions. This dynamic approach allows the system to optimize between latency and throughput on a per-packet basis, resolving the contradiction by selecting the appropriate network path according to specific transmission requirements.
3Quantity of substance
If the number of processing units increases, then system capacity is improved, but zero-load latency increases due to longer end-to-end diameter
Solution Approach 1:
The patent introduces an intermediate node that acts as a mediator between processing units and the packet transmission networks. This intermediate node calculates optimal routing paths by determining hop-counts through both mesh and tree networks, and selectively injects packets into the appropriate network based on these calculations. By using this intermediary to intelligently route packets through the tree network for distant destinations (reducing hop-count) and mesh network for nearby destinations (maximizing throughput), the system maintains low zero-load latency even as the number of processing units and end-to-end diameter increase.
Data Source
AI summary
Provided is a multi network and method to transmit packets. The multi network includes a mesh network, a tree network, and a network interface connected to the mesh network and the tree network and configured to transmit, through the mesh network and the tree network, a packet generated by a processing unit, of a processing system having plural processing units, at a starting point to a destination point for another processing unit of the processing system and configured to selectively inject the packet into one of the mesh network and the tree network to transmit the packet to the other processing unit.