Network Switch Buffer Management for Node Failure Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network switch technologies fail to maintain stable communication when a node experiences congestion or failure, leading to buffer memory overflow and impact on other nodes, as they continue to send pause frames even if the destination node's HDD is in failure, causing data stagnation and communication disruptions.
Innovation Solution
A switch with a determining section to assess the node's state based on continuous congestion notification packets and buffer memory usage, deleting data addressed to non-ordinary nodes to free up memory and prevent communication disruptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the network switch continues to send data to a node in failure state, then the node can potentially receive data, but the buffer memory capacity is exhausted and communications of all nodes are affected
Solution Approach 1:
The network switch proactively monitors the operational state of destination nodes before buffer exhaustion occurs. When a node failure is detected, the switch preemptively stops transmitting data to that node and deletes accumulated data in the buffer, preventing buffer overflow and protecting communication stability for other nodes.
Solution Approach 2:
The patent extracts and removes data addressed to failed nodes from the shared buffer memory. This separation of good data (addressed to operational nodes) from bad data (addressed to failed nodes) allows the buffer capacity to be fully utilized for effective communications while preventing waste on failed destinations.
2Quantity of substance
If the network switch stops sending data to a node in failure state, then buffer memory is preserved for other nodes, but data addressed to the failed node stagnates in the buffer
Solution Approach 1:
Instead of waiting for the failed node to recover or for buffer exhaustion to occur naturally, the network switch immediately skips the failed node by detecting its failure state and rapidly deleting accumulated data addressed to it. This rushing through the problem prevents time loss and buffer waste.
Solution Approach 2:
The network switch implements continuous monitoring of node operational states through feedback mechanisms. When a node failure is detected via feedback signals, the switch responds by adjusting its transmission behavior and cleaning the buffer, creating a closed-loop control system that prevents data stagnation.
3Reliability
If the network switch monitors node state continuously, then communication stability is maintained, but system complexity increases
Solution Approach 1:
The network switch performs self-diagnosis and self-correction by autonomously monitoring its own buffer state and detecting node failures. The switch independently determines when to stop transmission and when to delete buffer data without external intervention, simplifying the overall system architecture while maintaining high reliability.
Data Source
AI summary
A switch has a determining section and a memory managing section. The determining section determines whether or not the node is in a non-ordinary state in which received packets cannot be successfully processed, on a basis of a determination whether or not congestion notification packets received from a node have been continuously received during at least a given period and a determination whether or not a quantity of memory used in a buffer memory which accumulates received packets is at least a given value. The memory managing section deletes, in a case where the node is determined to be in the non-ordinary state, data addressed to the node in the non-ordinary state among data accumulated in the buffer memory.