Network Latency Optimization via Dynamic Flow Path Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Packet-switched communications networks face issues with bufferbloat, leading to increased latency and reduced network throughput, particularly affecting high-data-rate applications like gaming, due to excessive buffering of data packets.
Innovation Solution
Implementing a network traffic management technique that selects a flow path based on latency rather than available bandwidth, allowing computing devices to determine and adjust endpoint latency values to optimize packet buffer queue depths and manage network communications paths dynamically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If packet prioritization is applied to minimize delay for high priority packets, then latency for high priority packets is improved, but latency for low priority packets increases and network throughput decreases
Solution Approach 1:
The patent changes the parameter used for flow path selection from bandwidth-based to latency-based. By calculating latency values for multiple flow paths and selecting the path with minimum latency, the system optimizes packet transmission time without excessive buffering, thereby reducing bufferbloat while maintaining network throughput for all packet types.
Solution Approach 2:
The patent implements dynamic flow path selection based on real-time latency measurements. The system continuously monitors network conditions and dynamically adjusts the selected flow path to minimize latency, rather than using static bandwidth allocation or fixed prioritization schemes that cause bufferbloat.
2Loss of time
If quality of service (QoS) and packet prioritization are used to transmit higher priority packets ahead, then latency for prioritized packets is reduced, but performance for low priority traffic flows deteriorates
Solution Approach 1:
The patent segments the network path selection process into multiple independent flow paths, each evaluated based on its latency characteristics. By segmenting and independently optimizing each path based on latency rather than strict prioritization, the system ensures reliable performance for all traffic flows while maintaining low latency for prioritized packets.
Solution Approach 2:
The patent implements feedback mechanisms to monitor actual latency performance and adjust flow path selection accordingly. By continuously measuring latency and using this feedback to refine path selection, the system maintains reliable performance for low priority traffic while still providing low latency for high priority packets, avoiding the performance degradation caused by traditional QoS prioritization.
3Productivity
If excess buffering is used to manage network traffic, then network throughput is improved, but latency and jitter increase
Solution Approach 1:
The patent inverts the traditional approach by not increasing buffer sizes to improve throughput, but instead selecting flow paths based on minimizing latency. This inversion of the conventional wisdom leads to reduced bufferbloat and lower latency while maintaining network throughput through optimal path selection rather than excessive buffering.
Data Source
AI summary
A network may provide latency optimization by configuring respective latency values of one or more network components. A latency manager may receive a request indicative of a maximum latency value of a communications path between two devices, and may determine a particular network latency value. The latency manager may then determine respective endpoint latency values for endpoint devices based on the maximum latency value and network latency values. In addition, buffer characteristics, such as buffer depth at particular devices, may be adjusted in view of the latencies.