Packet Congestion Information Fields for Precise Multi-Hop Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing congestion control methods, such as ECN, struggle with inaccurate congestion detection due to limited information carried in the packet headers, leading to inefficient network utilization and difficulty in performing precise congestion control, especially in long-distance multi-hop scenarios.
Innovation Solution
Implementing a congestion control method that includes a congestion detection indication and a congestion information field in the packet, allowing devices to obtain and update congestion information, enabling accurate congestion control by carrying a larger amount of information without increasing packet size hop by hop.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ECN technology uses two bits in IP header to mark congestion status, then congestion detection is enabled, but congestion control accuracy is insufficient
Solution Approach 1:
The patent transitions from using only 2 bits in the IP header to utilizing the TCP header's ECN field and extending congestion information to include multiple dimensions such as queue utilization, link utilization, and hop count. This dimensional expansion allows comprehensive congestion information to be carried without increasing packet size excessively, thereby improving measurement precision while preserving information completeness.
Solution Approach 2:
The patent changes the parameters carried in congestion information from simple binary ECN bits to multi-parameter data including queue utilization rate, link utilization rate, and accumulated congestion hop count. By changing these parameters, the system achieves accurate congestion detection and control while managing packet size through selective parameter inclusion.
2Loss of information
If congestion information is added hop by hop in packet forwarding, then comprehensive congestion information is collected, but packet size increases continuously
Solution Approach 1:
The patent merges congestion information from multiple hops into a single consolidated data structure in the TCP header. Instead of adding separate congestion data at each hop, the system combines queue utilization, link utilization, and hop count into one integrated congestion information field, preventing packet size from increasing continuously while maintaining information completeness.
Solution Approach 2:
The patent extracts only the essential congestion parameters (queue utilization, link utilization, hop count) from each network device and places them in a dedicated congestion information field. By extracting only necessary information rather than copying entire device states, the system maintains comprehensive congestion data while controlling packet size growth.
3Device complexity
If only two ECN bits are used for congestion marking, then implementation is simple, but congestion control precision is insufficient
Solution Approach 1:
The patent makes the TCP header's ECN field serve multiple functions: it not only marks congestion status but also carries quantitative congestion information such as queue utilization and link utilization. This multi-functionality allows the system to maintain simple implementation using existing TCP header structures while achieving precise congestion control through enriched information content.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
A congestion control method, apparatus, and system are provided, and pertain to the field of network technologies. The method includes: A first device obtains a first packet. The first packet includes congestion detection indication information and a congestion information field. The congestion detection indication information indicates a device to obtain congestion information corresponding to the device. The congestion information field is used to carry the congestion information. The first device obtains, based on the congestion detection indication information, congestion information corresponding to the first device. The first device updates the congestion information field based on the congestion information corresponding to the first device. In this application, the congestion information field in the first packet can carry a large amount of congestion information, so that an end-side device performs accurate congestion control based on the congestion information carried in the congestion information field.