Redundant Software Event Mitigation for Memory Integrity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory systems face challenges in detecting and correcting single event upsets (SEUs) and hardware failures, which can lead to unintended changes in memory states due to ionizing radiation or other vulnerabilities, affecting data integrity and system performance.
Innovation Solution
A system comprising an event component, a comparison component, and a mitigation component that initiates events in redundant software applications, stores data, compares check values, and re-initiates events to correct errors, ensuring data integrity and preventing corrupted data propagation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If redundant software applications are used to detect and correct SEUs, then data integrity and system reliability are improved, but device complexity and processing time increase
Solution Approach 1:
The patent creates redundant copies of software applications (first and second software applications) that execute the same events and produce identical data outputs. By comparing check values from these copies, the system can detect SEUs without requiring complex error correction hardware, thus improving reliability while keeping the complexity manageable through software-based redundancy.
Solution Approach 2:
The comparison component continuously monitors check values from redundant software executions and provides feedback when discrepancies are detected. This feedback mechanism triggers the mitigation component to re-initiate events, creating a closed-loop system that automatically corrects errors without requiring complex external intervention systems.
2Measurement precision
If check values are compared between redundant software applications, then SEU detection capability is improved, but processing time and computational overhead increase
Solution Approach 1:
The system pre-computes and stores check values (such as checksums or hash values) alongside the data outputs from redundant software applications. This preliminary action allows for rapid error detection by simply comparing pre-computed values rather than re-processing entire data sets, thus improving detection capability while minimizing additional processing time.
Solution Approach 2:
The patent transforms complex data comparison into simpler parameter comparison by using check values (summary parameters) that represent entire data outputs. This parameter transformation reduces the computational overhead from comparing large data sets to comparing compact check values, thereby improving detection precision while reducing processing time.
3Reliability
If events are re-initiated upon check value mismatch, then error correction capability is improved, but system productivity decreases due to additional processing cycles
Solution Approach 1:
The system implements periodic re-initiation of events only when check value mismatches are detected, rather than continuous re-execution. This periodic action approach maintains error correction capability by responding to actual errors while avoiding unnecessary re-processing cycles, thus preserving system throughput and productivity during normal operation.
Solution Approach 2:
The patent extracts and isolates the error correction function into a separate mitigation component that only activates when needed. This extraction allows the main processing system to continue operating at full productivity while the error correction mechanism handles only the specific cases where SEUs are detected, minimizing the impact on overall system throughput.
Data Source
AI summary
Systems and techniques for memory event mitigation in redundant software installations are presented. A system can initiate an event executed by a first software application and a second software application that corresponds to the first software application. First data associated with the first software application can be stored in a first data store and second data associated with the second software application can be stored in a second data store. The system can also compare a first check value for the first data associated with the event and a second check value for the second data associated with the event. Furthermore, the system can re-initiate the event executed by the first software application and the second software application in response to a determination that the first check value does not correspond to the second check value.