Group Congestion Control for Network Egress Queues
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network congestion control mechanisms in data centers often fail to accurately identify the source of congestion, leading to inefficient rate control measures that may violate service level agreements (SLAs) and inadequately address network overflow issues.
Innovation Solution
Implementing a group congestion control mechanism where network elements transmit a group congestion notification message to multiple transmitters associated with a congested egress queue, allowing for dynamic rate limiting across all transmitters within the congestion group, rather than targeting a single transmitter.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single transmitter is targeted for rate control when congestion occurs, then the congestion control mechanism is simple to implement, but the measurement precision of congestion cause is insufficient and rate control effectiveness is reduced
Solution Approach 1:
The patent segments transmitters into groups based on their association with specific egress queues. Instead of treating all transmitters uniformly or targeting a single transmitter, the system divides the transmitter population into manageable groups that can be individually managed. This segmentation enables more precise congestion control by identifying which group of transmitters is causing congestion for a specific queue, thereby improving measurement precision while maintaining manageable complexity through standardized grouping mechanisms.
Solution Approach 2:
The patent introduces an intermediary mechanism where network elements (switches/routers) act as mediators between transmitters and the congestion control system. These network elements collect congestion information and generate group congestion notification messages that are then distributed to relevant transmitters. This intermediary approach improves measurement precision by providing structured congestion feedback while managing complexity through automated mediation rather than direct transmitter coordination.
2Reliability
If rate control is applied to ensure SLA compliance, then service level agreement compliance is improved, but network throughput is reduced
Solution Approach 1:
The patent applies local quality by differentiating rate control measures based on the specific egress queue and transmitter group involved. Instead of applying uniform rate control to all transmitters, the system tailors control parameters to the local conditions of each queue-transmitter group. This enables SLA compliance for affected queues while preserving throughput for other unaffected traffic, achieving reliability where needed without unnecessarily reducing overall productivity.
Solution Approach 2:
The patent implements dynamic rate control where transmit rates are adjusted in real-time based on current congestion conditions and SLA requirements. The system dynamically modifies rate parameters for different transmitter groups based on observed congestion patterns, allowing rates to be increased when queues are clear and decreased only when necessary to meet SLAs. This dynamic approach maintains SLA compliance while maximizing throughput during normal operating conditions.
3Productivity
If dynamic rate limiting is applied to all transmitters in a congestion group, then network congestion is effectively reduced, but the complexity of managing multiple transmitters increases
Solution Approach 1:
The patent merges multiple transmitters into unified groups that are managed collectively rather than individually. By combining transmitters that share common egress queues into single manageable groups, the system reduces the complexity of managing multiple transmitters while maintaining effective congestion control. The group-based approach allows a single rate control mechanism to affect multiple transmitters simultaneously, improving congestion reduction effectiveness without proportionally increasing management complexity.
Solution Approach 2:
The patent creates a universal group congestion control mechanism that can be applied across different transmitter groups and network conditions. The same basic framework for forming groups, collecting congestion information, and applying rate control works universally regardless of the specific transmitters or queues involved. This multi-functional approach improves congestion reduction effectiveness while avoiding the need to design and manage separate control mechanisms for each transmitter, thereby limiting the increase in complexity.
Data Source
AI summary
At a network-connected device, congestion at an egress queue can be detected. A potential source of congestion can be identified based on characteristics of a packet that caused the egress queue to become congested. The source of congestion can be a congestion group of transmitters. A group congestion message can be sent to the group of transmitters. The message can identify the packet that caused the egress queue to become congested. Transmitters can respond to the message by reducing their peak transmission rate.