面向运行态的内存故障风险评估与自适应诊断方法及系统
By constructing cross-layer differential triplets and drift evaluation sequences, the problem of cross-layer coupling changes in runtime memory access is solved, enabling accurate identification and dynamic response of fault risks, and improving the accuracy and stability of fault determination.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUIRONG ELECTRONIC SYST ENG CO LTD
- Filing Date
- 2026-03-06
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies cannot effectively handle cross-layer coupling changes in runtime memory access between the host layer, virtualization layer, and NUMA topology. This makes fault risk assessment susceptible to access jitter interference and virtualization scheduling offsets, and lacks cross-layer sequence consistency identification and risk level classification response strategies.
By collecting runtime memory monitoring data, performing time synchronization, access jitter smoothing, anomaly removal, interpolation compensation and standardization processing, constructing cross-layer differential triples, identifying cross-layer inconsistent segments, generating cross-layer drift assessment sequences, and performing response operations based on risk level labels to conduct stability assessments.
It enables accurate identification and dynamic response to cross-layer memory fault risks, improves the accuracy and stability of fault determination, reduces the propagation range of potential faults, and enhances online handling capabilities.
Smart Images
Figure CN121807452B_ABST