Controller Port Failover via Identifier Reassignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional failover methods in network arrangements, which involve host devices, require changes at the host device level, leading to reduced input/output communications capacity and affecting load balancing, as they mark failed paths as unavailable, whereas performing failover at the controller node level can maintain transparency to the host device and ensure reliable failback with health checks.
Innovation Solution
Implementing failover and failback procedures primarily at the controller node level, where the failover module detects and reassigns port identifiers to maintain communication paths, and performs health checks before reverting to original paths, utilizing mechanisms like NPIV for multiple logical path management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If failover is performed at the host device level, then failover capability is achieved, but input/output communications capacity is reduced and load balancing is affected
Solution Approach 1:
The patent introduces a fabric switch as an intermediary component between the host device and controller nodes. The fabric switch performs failover operations and health checks, acting as a mediator that prevents host devices from directly managing failover. This intermediary approach maintains communication capacity at the host level while enabling failover capability through the fabric switch's port identifier reassignment mechanisms.
2Productivity
If failover is performed at the controller node level, then transparency to host device is maintained and load balancing is preserved, but complexity of failover management increases
Solution Approach 1:
The patent merges failover management functions into the fabric switch, combining path monitoring, health checks, and port identifier reassignment capabilities in a single centralized component. This consolidation reduces the complexity burden on individual controller nodes while maintaining load balancing transparency to host devices, as the fabric switch handles all failover orchestration centrally.
Solution Approach 2:
The patent implements automated health check mechanisms that actively and continuously monitor path availability, accelerating the detection of failures and enabling rapid failover response. This accelerated monitoring approach reduces the time required to detect and respond to failures, improving the overall failover management efficiency despite the increased complexity.
3Reliability
If port identifiers are reassigned during failover, then communication path availability is maintained, but risk of ping-pong failures increases
Solution Approach 1:
The patent implements health checks as a preliminary action before executing failback operations. The fabric switch performs health checks to verify path availability and controller node status before reassigning port identifiers back to their original positions. This preliminary verification prevents premature failback and eliminates ping-pong failures by ensuring the original path is truly stable and available before restoration.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
A failure at a first port of the controller node is detected, where the first port is initially assigned a first port identifier and is associated with a logical path through a communications fabric between the first port and a port at a host device. In response to detecting the failure, the first port identifier is assigned to a second port to cause the logical path to be associated with the second port. In response to detecting resolution of the failure, a probe identifier is assigned to the first port. Using the probe identifier, a health of network infrastructure between the first port and the host device is checked. In response to the checking, the first port identifier is assigned to the first port to cause failback of the logical path to the first port.