Network Congestion Management via Priority-Based Packet Dropping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computer networks face congestion issues due to increasing online traffic, leading to dropped packets and interrupted traffic flows, which affect the quality of service (QoS) and quality of experience (QoE) for subscribers, and existing solutions fail to differentiate between critical and low-priority flows, causing global synchronization and unfair resource distribution.
Innovation Solution
The system employs an enhanced Weighted Random Early Detection (WRED) method that determines packet priority based on attributes such as application type, subscriber status, and session attributes, allowing for proactive dropping of low-priority packets during congestion, thereby preventing queue overflow and maintaining high QoS for critical flows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If network capacity is increased to handle growing traffic, then network congestion is reduced, but capital costs increase
Solution Approach 1:
The system performs preliminary congestion detection by monitoring queue depths and determining congestion levels before actual packet loss occurs. By identifying congestion early and implementing preventive measures such as dropping low-priority packets or adjusting transmission rates, the system avoids the need for costly network upgrades to handle peak loads.
Solution Approach 2:
The system dynamically changes network parameters including packet dropping probability, queue management policies, and transmission rates based on real-time congestion conditions. These parameter adjustments allow the network to adapt to varying traffic loads without requiring physical infrastructure changes, thereby reducing capital costs while maintaining productivity.
2Device complexity
If all packets are treated equally during congestion, then simplicity is maintained, but quality of service deteriorates for critical flows
Solution Approach 1:
The system applies different quality levels to different packets based on their priority classification. Critical packets (e.g., VoIP, real-time gaming) receive preferential treatment with lower dropping probabilities and higher queue priorities, while non-critical packets (e.g., file downloads, bulk transfers) are more readily dropped during congestion. This local differentiation maintains QoS for critical flows without requiring complete redesign of packet handling.
Solution Approach 2:
The system segments traffic into multiple priority classes or queues based on application type, protocol, and service requirements. By dividing the traffic stream into segments with different handling policies, the system can apply simplified rules to each segment while achieving complex overall QoS differentiation, thus balancing simplicity and reliability.
3Reliability
If packets are dropped during congestion, then queue overflow is prevented, but packet loss increases affecting subscriber experience
Solution Approach 1:
The system converts the potentially harmful effect of packet dropping into a beneficial congestion control mechanism. By strategically dropping only low-priority packets during congestion while preserving high-priority traffic, the system uses packet loss as a signal to regulate traffic flow and prevent complete queue overflow, thereby maintaining overall network stability and subscriber experience for critical services.
Solution Approach 2:
The system applies partial dropping action by selectively discarding only a portion of packets based on their priority level rather than uniformly dropping all packets or none. This partial action allows the system to prevent queue overflow while minimizing the impact on subscriber experience by preserving essential traffic and only sacrificing non-critical data.
4Reliability
If network upgrades are implemented to eliminate congestion, then quality of experience is improved, but cost-effectiveness decreases
Solution Approach 1:
The system implements preliminary congestion management through intelligent packet prioritization and selective dropping policies that prevent congestion from degrading QoS. By addressing congestion proactively through software-based traffic management rather than waiting for network saturation, the system maintains high quality of experience without requiring expensive physical network upgrades, thereby improving cost-effectiveness.
Data Source
AI summary
A method for congestion management on a computer network including: receiving a packet from a traffic flow; determining at least one attribute associated with the packet; determining a priority level for the packet based on the at least one attribute; determining a queue depth for a queue in a data plane path of the packet; determining whether to send or drop the packet based on the priority level and the queue depth. A system for congestion management including: an incoming packet handler configured to receive a packet; an application detector configured to determine at least one attribute associated with the packet; a policy module configured to determine a priority level for the packet based on the at least one attribute; an enhanced weighted random early detection module configured to determine a queue depth and whether to send or drop the packet based on the priority level and the queue depth.