Network Switch Virtual Channel Bottleneck Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Complex storage area networks (SANs) face challenges in managing bottlenecks, stuck virtual channels, and credit loss issues, which can lead to performance degradation and require effective detection and mitigation tools to maintain network reliability and efficiency.
Innovation Solution
A network switch equipped with an application-specific integrated circuit (ASIC) and software that detects stuck virtual channels, slow-drain bottlenecks, and lost credits, providing alerting mechanisms and mitigation techniques such as timeout adjustments and queue flushing to address these issues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If network switches are configured into a switch fabric with multiple interconnected switches to handle complex SAN requirements, then network capacity and connectivity are improved, but network complexity and difficulty of managing bottlenecks and stuck virtual channels increase
Solution Approach 1:
The patent segments the network fabric into multiple independent virtual channels (VC0, VC1, VC2, VC3) that can operate independently. Each virtual channel has its own credit management and flow control mechanisms, allowing the network to handle multiple simultaneous connections without interference. This segmentation enables complex network requirements to be met while maintaining manageable individual channel complexity.
Solution Approach 2:
The patent introduces virtual channels as intermediary layers between physical network connections. These virtual channels act as mediators that manage credit flow, detect stuck states, and coordinate communication between switches. The virtual channel layer provides an abstraction that simplifies the management of complex inter-switch connections by introducing structured intermediate control mechanisms.
2Reliability
If flow control credits are used to manage data transmission between switches, then data loss is prevented, but detection of lost credits and stuck virtual channels becomes more difficult
Solution Approach 1:
The patent implements feedback mechanisms where each virtual channel continuously monitors its credit status and transmission state. When credits are lost or a virtual channel becomes stuck, the system detects this through feedback from credit counters and transmission state monitoring. The feedback loop enables automatic detection of anomalies while maintaining the reliability benefits of credit-based flow control.
Solution Approach 2:
The patent performs preliminary detection actions by continuously monitoring credit counters and transmission states before actual data loss occurs. The system checks for stuck virtual channels and lost credits proactively through regular status verification, enabling early detection and mitigation before reliability is compromised.
3Productivity
If virtual channels are used to manage multiple connections through a single physical port, then network efficiency is improved, but stuck virtual channels can cause performance degradation
Solution Approach 1:
The patent segments the single physical port into multiple independent virtual channels, each capable of simultaneous operation. This segmentation allows efficient utilization of the physical port while isolating potential stuck conditions to individual virtual channels rather than affecting the entire port. Each virtual channel can be independently detected and mitigated when stuck.
Solution Approach 2:
The patent applies local quality by giving each virtual channel its own independent credit management, flow control, and detection mechanisms. When one virtual channel becomes stuck, the local detection and mitigation mechanisms address only that specific channel without affecting other virtual channels on the same physical port, maintaining overall performance stability.
4Ease of operation
If manual detection and correction of network problems are required, then administrators have full control, but response time and network availability are reduced
Solution Approach 1:
The patent implements self-service mechanisms where the network switch automatically detects stuck virtual channels and lost credits, and autonomously mitigates these conditions without requiring manual administrator intervention. The system performs self-diagnosis and self-correction through automatic detection algorithms and mitigation protocols, dramatically reducing response time while maintaining network availability.
Solution Approach 2:
The patent uses feedback mechanisms to automatically notify administrators of detected problems while simultaneously initiating automated mitigation actions. The feedback loop provides administrators with information about network conditions while the system independently responds to issues, combining automated rapid response with administrator oversight and control.
Data Source
AI summary
Manageability tools are provided for allowing an administrator to have better control over switches in a lossless network of switches. These tools provide the ability to detect slow drain and congestion bottlenecks, detect stuck virtual channels and loss of credits, while hold times on edge ASICs to be different from hold times encore ASICs, and mitigate severe latency bottlenecks.