Network Congestion Detection via Propagation Time Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network congestion, particularly frame-based bottlenecks, are difficult to predict and manage, as existing methods often require costly infrastructure upgrades or disruption of network switches, and do not effectively address congestion across the entire network.
Innovation Solution
Implementing congestion detection by capturing send and receive times using local clocks of sender and receiver nodes, computing total propagation time, and adjusting transmission routes based on detected congestion, without modifying existing switching infrastructure, by deploying software or firmware updates on nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If congestion detection is implemented at the switch, then congestion can be detected, but it requires costly infrastructure upgrades and disruption of network switches
Solution Approach 1:
The patent introduces timing information as an intermediary carrier that conveys congestion data from the receiver back to the sender. Instead of requiring the switch to perform complex detection, the receiver measures propagation time and embeds this timing information in acknowledgment packets, which the sender then uses to detect congestion. This shifts the detection function from the switch to the end nodes, avoiding infrastructure upgrades.
Solution Approach 2:
The sender and receiver nodes perform congestion detection themselves using their own resources (local clocks and processing capabilities). The receiver self-measures the propagation time and embeds it in acknowledgment packets, while the sender self-detects congestion by analyzing round-trip times. This eliminates the need for the switch to provide detection services, reducing infrastructure complexity.
2Reliability
If existing methods are used to detect congestion, then some congestion can be detected, but they do not effectively address congestion across the entire network
Solution Approach 1:
The patent makes the congestion detection mechanism universal by implementing it at the end nodes (senders and receivers) rather than at specific network infrastructure points. Any node that sends or receives data can participate in detection, providing network-wide coverage. The same mechanism works for all types of congestion (traffic-based and frame-based) and all network locations, enhancing both reliability and adaptability.
3Measurement precision
If switch-based congestion detection is implemented, then congestion can be monitored, but it requires disruption of network switches and affects communication ability
Solution Approach 1:
The patent uses acknowledgment packets as intermediaries to carry timing information from receiver to sender. This allows congestion detection to occur without interrupting normal data flow, as the timing data is embedded in packets that are already part of the communication protocol. The switch continues to forward traffic normally while end nodes perform detection, maintaining communication continuity.
4Ease of manufacture
If nodes deploy software or firmware updates for congestion detection, then implementation cost is reduced, but it requires updating network devices
Solution Approach 1:
The patent implements congestion detection using software or firmware at the node level rather than requiring expensive hardware upgrades to switches. The timing measurement and detection logic are implemented as software components that can be deployed and updated independently. This approach uses cheaper, more flexible software solutions instead of costly hardware modifications, reducing implementation barriers while managing node software complexity.
Data Source
AI summary
A method and corresponding apparatus for detecting network congestion. The method includes capturing, using a local clock of a sender device, a send time of an outgoing packet sent from the sender device to a receiver device through a forward route, and capturing, using the local clock of the sender device, a receive time of an acknowledgment packet sent from the receiver device to the sender device through a backward route. The acknowledgment packet contains timing information, generated using a local clock of the receiver device, for determining an internal latency of the receiver device. A round trip time is computed as a difference between the send time and the receive time. The internal latency is subtracted from the round trip time to compute a total propagation time. If the total propagation time is above a threshold, the forward route and the backward route are changed.