Emulated EEE Memory Partitioning for Endurance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Emulated electrically erasable (EEE) memory systems face a challenge in achieving increased endurance without increasing memory size, leading to higher costs due to the need for larger non-volatile memory.
Innovation Solution
The system partitions a volatile memory into two portions, one for high dynamic records that require frequent updates and another for low dynamic records, using a combination of flash memories to emulate a larger non-volatile memory with improved endurance without increasing overall memory size, by selectively storing and updating data in a logically sequential manner.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the size of non-volatile memory is increased to achieve increased endurance, then the endurance of the EEE memory is improved, but the cost of the device increases
Solution Approach 1:
The patent divides the non-volatile memory into multiple partitions (first NVM partition, second NVM partition, etc.) and divides the volatile memory into corresponding portions. Each partition stores data for specific volatile memory portions, allowing selective wear distribution across partitions. This segmentation enables the system to achieve improved endurance for frequently updated data without requiring a proportional increase in total non-volatile memory capacity, thereby controlling costs while maintaining reliability.
2Reliability
If the size of non-volatile memory is increased to provide increased endurance, then the endurance is improved, but the memory size becomes larger
Solution Approach 1:
The patent implements dynamic wear leveling through a wear indicator mechanism that tracks the number of erase cycles for each NVM partition. The system dynamically redirects future erase operations from worn partitions to less-worn partitions based on these indicators. This dynamic adaptation allows the same physical memory size to provide extended endurance by evenly distributing wear across all partitions over time, preventing any single partition from reaching its endurance limit first.
Data Source
AI summary
A method and system wherein a volatile memory is partitioned to have a first percentage of address space dedicated to a first classification of data which is data that is expected to have greater than a predetermined number of times of being modified and a second percentage of address space dedicated to a second classification of data which is data that is expected to have less than the predetermined probability of being modified. Address assignment of data to be stored in the volatile memory is made on a basis of predicted change of the data. Memory addresses of the first and second percentages of address space are respectively assigned to first and second sections of nonvolatile memory. The memory addresses of the first percentage initially consume a smaller percentage of an address map of the first section than the memory addresses of the second percentage of the second section.

