Memory Resource Allocation Using Error-Aware Defect Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data centers face challenges in managing and efficiently recovering defective memory devices, particularly in large-scale, modular data centers where resources are individually constructed and restructured based on usage, necessitating a system for categorizing and analyzing error information from memory devices to optimize their operation and allocation.
Innovation Solution
A memory system comprising a plurality of memory devices with error correction capabilities and a controller to acquire and categorize error information, along with a compute system to analyze this data and allocate resources based on workload requirements, including the use of spare memory devices for defective ones and independent power management of normal and shared memory devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If memory devices are individually constructed and restructured in large-scale data centers, then resource flexibility and adaptability improve, but device complexity and management difficulty increase
Solution Approach 1:
The patent segments memory devices into normal and shared categories, with shared memory devices serving multiple memory controllers. This segmentation enables flexible resource allocation while simplifying management through standardized sharing protocols and defect isolation mechanisms.
Solution Approach 2:
Shared memory devices are designed to serve multiple memory controllers simultaneously, providing universal functionality. This multi-functionality improves resource utilization and flexibility while the patent manages complexity through coordinated access protocols and defect containment strategies.
2Reliability
If error information is collected and analyzed from memory devices, then reliability and defect detection improve, but system complexity and processing overhead increase
Solution Approach 1:
The patent extracts error information from memory devices and processes it separately through dedicated error management logic. This extraction approach improves defect detection accuracy while managing complexity by isolating analysis functions from core memory operations.
Solution Approach 2:
The system implements feedback mechanisms where error information from memory devices is continuously monitored and analyzed. This feedback loop improves reliability through real-time defect detection while the patent manages complexity through automated response protocols and selective analysis.
3Reliability
If spare memory devices are used to replace defective ones, then system reliability improves, but device complexity and recovery complexity increase
Solution Approach 1:
Spare memory devices are pre-configured and ready before defects occur. The patent implements preliminary actions by maintaining standby shared memory devices that can be quickly activated upon defect detection, improving system availability while reducing recovery complexity through pre-established replacement protocols.
Solution Approach 2:
The patent implements automated mechanisms to discard defective memory devices and recover functionality through spare devices. This approach improves reliability while managing recovery complexity through standardized replacement procedures and defect isolation.
4Loss of energy
If independent power management is implemented for normal and shared memory devices, then energy efficiency improves, but control complexity increases
Solution Approach 1:
The patent implements dynamic power management where shared memory devices can be independently powered on or off based on workload requirements. This dynamic control improves energy efficiency while managing complexity through automated power state transitions and workload-based activation protocols.
Data Source
AI summary
A memory system and a data processing system including the memory system may manage a plurality of memory devices. For example, the data processing system may categorize and analyze error information from the memory devices, acquire characteristic data from the memory devices and set operation modes of the memory devices based on the characteristic data, allocate the memory devices to a host workload, detect a defective memory device among the memory devices and efficiently recover the defective memory device.