Ethernet Storage Area Network Congestion Control via Port-Speed Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Ethernet storage area networks lack in-band mechanisms to detect the speed of Ethernet ports, leading to network congestion and service degradation due to mismatched data transmission rates between storage arrays and host systems.
Innovation Solution
Implement a centralized discovery controller (CDC) to obtain the operating speed of endpoints connected to Ethernet ports, allowing endpoints to adjust bandwidth consumption to match the connected adapter speed, thereby mitigating congestion by limiting data transmission to avoid overloading the host adapter.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If the storage array sends data at a high rate, then data transmission speed is improved, but network congestion occurs and service degradation results
Solution Approach 1:
The patent implements a feedback mechanism where the storage array queries the host adapter's maximum transfer capability through the Ethernet network using standardized TCP/IP protocols. The host adapter responds with its speed capability, and the storage array uses this feedback information to dynamically adjust its data transmission rate, preventing network congestion while maintaining optimal performance.
Solution Approach 2:
The system enables self-service by allowing the host adapter to autonomously report its speed capability to the storage array through standardized in-band protocols. This eliminates the need for external configuration or manual intervention, as the storage array automatically receives and processes the speed information to adjust its transmission rate accordingly.
2Reliability
If in-band mechanisms are implemented to detect port speed, then congestion control is improved, but device complexity increases
Solution Approach 1:
The patent achieves congestion control using existing, standardized TCP/IP protocols that are universally supported by Ethernet networks. The same communication infrastructure used for data transmission is leveraged for speed detection, eliminating the need for specialized hardware or proprietary protocols and thus avoiding additional device complexity.
Solution Approach 2:
The Ethernet network itself serves as an intermediary, carrying both data traffic and speed capability information between the storage array and host adapter. This approach uses the existing network infrastructure to convey control information, avoiding the need for separate dedicated control channels or additional complex signaling mechanisms.
Data Source
AI summary
Embodiments herein allow for endpoints, such as a storage array or a host system, to obtain the speed of a corresponding endpoint that is connected to a port or ports of the endpoint. Given port speeds, an endpoint system may adjust the bandwidth of a payload being sent to the corresponding endpoint system to match the connected port's speed. For example, limiting the data being sent from a storage endpoint system to a host endpoint system in cases where the storage array's port is faster than the host adapter helps mitigate or even ensure that the host adapter of the host endpoint system is not overloaded—thereby avoiding congestion. In one or more embodiments, a centralized discovery controller (CDC) may be enhanced to obtain the current operating speed of a port of an endpoint system.