Optical Transceiver Failure Notification Logic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In supercomputer systems, immediate power disconnections lead to high loads on servers and networks due to excessive failure information transmission, which existing techniques are unable to effectively mitigate.
Innovation Solution
An information processing apparatus with an optical transceiver that returns optical signals between channels during power disconnection, allowing for determination of the power state of external nodes and selective transmission of failure notifications, thereby distinguishing between immediate power disconnection and ICC failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If failure information is transmitted for all detected failures in immediate power disconnection scenarios, then complete failure information is recorded, but server and network load increases significantly
Solution Approach 1:
The patent applies preliminary action by checking the power state of external nodes before transmitting failure information. The system proactively determines whether a node is in immediate power disconnection state by detecting optical signal characteristics, and only then decides to suppress failure information transmission. This prevents unnecessary transmissions before they occur, reducing server and network load while maintaining reliability for actual failures.
2Power
If optical signals are monitored to determine power state, then unnecessary failure information can be suppressed, but system complexity increases
Solution Approach 1:
The patent uses optical signals as an intermediary to determine the power state of external nodes without requiring direct communication or complex control systems. The optical transceiver naturally provides signal presence/absence information that serves as a simple indicator of node power state. This intermediary approach avoids adding complex monitoring infrastructure while still enabling intelligent failure information suppression.
3Power
If existing techniques are used to reduce load, then some load reduction is achieved, but immediate power disconnection scenarios are not effectively addressed
Solution Approach 1:
The patent replaces complex mechanical or electronic monitoring systems with optical signal detection to determine node power state. By substituting the detection mechanism with optical signal presence/absence checking, the system achieves reliable immediate power disconnection detection without adding mechanical complexity. This optical approach naturally integrates with existing interconnect infrastructure.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces the load on failure management servers and networks by suppressing unnecessary failure information transmission caused by immediate power disconnections, enhancing processing efficiency and scalability.
Implementation Method 1
an optical transceiver configured to return an optical signal received by a first channel to a second channel
Data Source
AI summary
An information processing apparatus includes an optical transceiver configured to return an optical signal received by a first channel to a second channel at a time of immediate power disconnection of a casing accommodating nodes; a memory; and a processor coupled to the memory, wherein the processor detects an occurrence of a failure in inter-node communications with an external node, when the occurrence of a failure is detected, the processor controls the optical transceiver in order to emit light to the first channel, makes a determination as to whether or not the second channel is enabled to receive the optical signal, and determines whether or not a power source of the external node is off based on the determination, and when the power source of the external node is off, the processor selects a failure notification to transmit to a failure management device.