Bypassing Congestion in Converged Ethernet Fabric
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network congestion in data centers leads to poor quality of service, including queueing delay, packet loss, and reduced throughput, especially in multitenant environments with server and network virtualization, where traditional Ethernet protocols struggle to manage congestion effectively.
Innovation Solution
The method involves sampling queues across switches to detect congestion, propagating congestion notification messages upstream, and creating virtual queues to reroute packet traffic based on priority levels, allowing switches to proactively bypass downstream congestion points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional Ethernet protocols are used in multitenant datacenters with server and network virtualization, then network infrastructure is maintained, but network congestion occurs leading to poor quality of service
Solution Approach 1:
The patent implements proactive congestion management by monitoring queue depths and propagating congestion notifications upstream before congestion fully develops. This allows upstream switches to take preventive actions such as creating virtual queues and rerouting traffic before packet loss occurs, thereby maintaining quality of service and preventing throughput degradation
2Loss of energy
If congestion is allowed to develop naturally in best-effort networks, then network simplicity is maintained, but packet loss and queueing delay increase
Solution Approach 1:
The patent implements a feedback-based congestion management system where downstream switches monitor queue depths and send congestion notifications upstream. This feedback loop enables dynamic traffic management without requiring complex centralized control, as each switch autonomously responds to congestion signals by creating virtual queues and rerouting traffic based on flow identifiers
Solution Approach 2:
The patent segments traffic into different virtual queues based on flow identifiers and congestion notifications. By dividing traffic streams into manageable segments that can be independently managed and rerouted, the system reduces packet loss without requiring complete reconfiguration of the entire network infrastructure
3Productivity
If upstream switches proactively create virtual queues and reroute traffic, then congestion is bypassed and throughput is maintained, but switch complexity increases
Solution Approach 1:
The patent implements dynamic congestion management where upstream switches create and manage virtual queues on-demand based on received congestion notifications. The system dynamically reroutes traffic flows by modifying forwarding decisions using flow identifiers, allowing switches to adapt to changing congestion conditions without requiring static pre-configuration or complex centralized control
Data Source
AI summary
Embodiments relate to bypassing congestion points in a network. An aspect includes sampling queues of a plurality of switches in a network. When packet congestion is detected at a congestion point of a first switch, the packet flow contributing to the packet congestion is identified. A congestion notification message indicating the identified packet flow is then propagated to upstream switches, which are upstream from the first switch in the network. The congestion notification message is then snooped by the upstream switches. Virtual queues within the upstream switches are associated with the identified packet flow to hold packets associated with the identified packet flow. The packets associated with the identified packet flow are then re-routed to bypass the packet congestion in the first switch.