Network Transport Selection for Latency-Sensitive Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing networking technologies fail to efficiently manage traffic over multiple WAN transports to optimize performance for varying application requirements, as they often prioritize either low latency or high throughput without considering dynamic conditions and usage costs.
Innovation Solution
A networking device that classifies packets based on their latency sensitivity and selects the most appropriate network transport for each packet or group of packets by estimating the expected latency, taking into account the current conditions of available network transports, including queue depth and throughput, to achieve optimal performance and cost efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single network transport is used for all traffic, then device complexity is reduced, but performance optimization for different application requirements cannot be achieved
Solution Approach 1:
The patent segments network traffic into different classes (interactive, bulk, intermediate) and assigns each class to appropriate network transports based on their performance requirements. This segmentation enables optimized performance for each traffic type while managing complexity through systematic classification rules.
Solution Approach 2:
The patent implements dynamic transport selection where the system continuously monitors network conditions (latency, throughput, queue depth) and adapts transport assignments in real-time. This dynamic approach allows the system to respond to changing network conditions and optimize performance without requiring static, complex configuration.
2Productivity
If multiple network transports are used concurrently, then performance and cost efficiency are improved, but device complexity increases
Solution Approach 1:
The patent performs preliminary classification of traffic into different classes before transport selection. By pre-categorizing traffic based on its performance requirements, the system simplifies the subsequent transport selection process and enables more efficient concurrent use of multiple transports without excessive complexity.
Solution Approach 2:
The patent changes the parameter of transport selection from static to dynamic by considering real-time network conditions such as queue depth, latency measurements, and throughput availability. This parameter-based dynamic selection optimizes productivity while managing complexity through measurable, objective criteria.
3Loss of time
If a low-latency network transport is selected, then latency-sensitive applications perform better, but usage cost increases
Solution Approach 1:
The patent applies local quality by assigning different network transports to different traffic classes based on their specific requirements. Latency-sensitive interactive traffic is routed to low-latency transports, while bulk transfers use cost-effective high-throughput transports. This localized optimization achieves performance benefits without incurring costs for all traffic types.
Solution Approach 2:
The patent applies partial action by using low-latency transports only for the portion of traffic that actually requires low latency (interactive and intermediate classes), rather than applying them to all traffic. This partial application reduces usage costs while maintaining performance for latency-sensitive applications.
4Productivity
If a high-throughput network transport is used, then bulk transfer performance is improved, but latency increases
Solution Approach 1:
The patent applies local quality by matching transport characteristics to traffic requirements: high-throughput transports are assigned to bulk transfer traffic where throughput is the critical parameter, while low-latency transports are assigned to interactive traffic where latency is critical. This localized matching optimizes both throughput and latency for their respective traffic types.
Data Source
AI summary
Systems, methods, and apparatus, including computer-readable media, for enhanced network communication using multiple network connections. In some implementations, a networking apparatus concurrently maintains connectivity to a network through each of multiple network transports. The networking apparatus receives one or more packets to be transmitted over the network and classifies the one or more packets to determine a class of service. The networking apparatus selects one of the multiple network transports to transmit the one or more packets based on (i) the class of service for the one or more packets and (ii) measures of expected latency for transmission of the one or more packets over the respective multiple network transports. The networking apparatus transmits the one or more packets using the selected network transport.