Flow Control Mechanism for Congestion Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing congestion control protocols, such as TCP and SCTP, suffer from unfair penalization of flows that experience multiple packet drops, leading to random timeouts and inefficient bandwidth allocation, where a flow's congestion window is unfairly reduced, causing it to slow down excessively and potentially stop sending packets, while other flows gain more bandwidth.
Innovation Solution
Implementing a flow control mechanism that monitors bandwidth usage and applies selective packet dropping based on access control policies, using a virtual 'penalty box' to manage congestion, allowing fast retransmit instead of full stop timeouts, ensuring fairness among flows by dynamically adjusting packet drop rates based on bandwidth usage and flow priority.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If random packet dropping is used to control traffic congestion, then bandwidth allocation is simplified, but flow fairness deteriorates causing unnecessary timeouts and excessive slowdowns
Solution Approach 1:
The patent segments the packet dropping mechanism by introducing a penalty box counter that tracks dropped packets per flow separately. Instead of uniform random dropping, each flow receives individualized treatment based on its drop history, dividing the congestion control into per-flow segments that can be managed independently to prevent unnecessary timeouts
Solution Approach 2:
The patent implements feedback by having the router track the number of dropped packets for each flow in a penalty box counter. This feedback mechanism allows the router to adjust future dropping decisions based on past drops, preventing excessive drops that would trigger unnecessary timeouts while maintaining congestion control effectiveness
2Reliability
If multiple packets are dropped in succession from a congestion controlled flow, then congestion control is enforced, but flow sending rate collapses to slow start excessively
Solution Approach 1:
The patent applies beforehand cushioning by introducing a penalty box counter that accumulates drop counts before triggering actual packet drops. This cushioning mechanism allows the system to tolerate some dropped packets without immediately collapsing the flow to slow start, providing a buffer that prevents excessive reaction to temporary congestion
Solution Approach 2:
The patent changes the parameter of packet dropping from random to selective based on penalty box counter values. By modifying the dropping probability parameter dynamically based on per-flow drop history, the system maintains congestion control effectiveness while avoiding the parameter collapse that triggers unnecessary slow start transitions
3Reliability
If all incoming packets are dropped when queue becomes full, then congestion is controlled, but bandwidth utilization efficiency deteriorates
Solution Approach 1:
The patent applies local quality by treating different flows differently in the queue management. Instead of dropping all packets uniformly when the queue is full, the router selectively drops packets from specific flows based on their individual penalty box counters and AC policies, optimizing bandwidth utilization while maintaining congestion control
Data Source
AI summary
A flow in a flow set having an access control (“AC”) policy assigned is monitored. A bandwidth used by the flow is determined. One or more packets associated with the flow are selectively dropped based on at least one of the used bandwidth and the AC policy.