Network Congestion Control Using Path-Aware Device Measurements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data center networks experience burst congestion leading to severe bandwidth loss and long-tail latency, which compromises service-level agreement (SLA) performance and increases service failure rates.
Innovation Solution
A network congestion control method that involves a first network device obtaining measurement packets with forwarding device identifiers and measurement values to identify congested devices, determining control policies based on device roles, and implementing path replacement or rate reduction to resolve congestion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If network congestion control is implemented using traditional methods, then bandwidth utilization can be maintained, but service-level agreement (SLA) performance deteriorates due to long-tail latency and burst congestion
Solution Approach 1:
The patent segments service flows into different priority levels (e.g., high-priority and low-priority flows) and applies differentiated congestion control policies to each segment. High-priority flows receive guaranteed bandwidth and lower latency treatment, while low-priority flows share remaining bandwidth, thereby ensuring SLA performance for critical services while maintaining overall network utilization.
Solution Approach 2:
The patent implements local quality by applying different congestion control mechanisms to different network paths or queues based on their specific characteristics. For example, explicit congestion notification (ECN) is applied to high-priority flows while traditional drop policies are used for low-priority flows, ensuring that each flow receives appropriate treatment for its service requirements.
2Reliability
If network congestion control is implemented using traditional methods, then overall network throughput can be maintained, but service differentiation capability deteriorates as all flows are treated equally
Solution Approach 1:
The patent implements local quality by applying different congestion control mechanisms to different network paths or queues based on their specific characteristics. For example, explicit congestion notification (ECN) is applied to high-priority flows while traditional drop policies are used for low-priority flows, ensuring that each flow receives appropriate treatment for its service requirements.
Solution Approach 2:
The patent implements dynamic service differentiation by allowing the network to adapt congestion control policies in real-time based on current network conditions and flow characteristics. The system can dynamically adjust bandwidth allocation, latency thresholds, and congestion response mechanisms according to the specific needs of different service flows, enabling versatile service differentiation.
3Productivity
If congestion control policies are applied uniformly across all network devices, then implementation complexity is reduced, but congestion resolution effectiveness deteriorates due to inability to address location-specific congestion
Solution Approach 1:
The patent applies preliminary action by pre-configuring congestion control policies and parameters in network devices before congestion occurs. Network devices are pre-programmed with ECN thresholds, bandwidth allocation rules, and path selection criteria based on historical data and service requirements. When congestion is detected, these pre-configured policies are immediately activated, enabling rapid response without complex real-time decision-making.
Solution Approach 2:
The patent implements feedback mechanisms where network devices continuously monitor congestion conditions and adjust control policies based on real-time measurements. ECN marks are fed back to senders to reduce transmission rates, and network devices use this feedback to dynamically adjust queue management and path selection, creating a closed-loop control system that improves congestion resolution effectiveness.
Data Source
AI summary
A network congestion control method includes, during transmission of a first service flow, a first network device that obtains a measurement packet including groups of measurement results corresponding to the first service flow, and each group includes an identifier of a forwarding device on a forwarding path of the first service flow and a measurement value corresponding to the forwarding device. The first network device determines, based on the groups, a congested second network device on the forwarding path. The first network device determines a control policy based on location information of the second network device, where the control policy resolves congestion of the first service flow.