Selectable Write Paths for NVRAM Power Loss Protection Failure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional storage systems fail to utilize storage devices effectively by marking them as failed when a threshold number of dies fail, despite some dies being suitable for data storage, leading to reduced capacity and performance.
Innovation Solution
A storage system that marks individual dies as likely to fail and recovers data from these dies, allowing them to operate in read-only or degraded modes while using remaining dies for data storage, thereby improving data retention, protection, and system capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If storage devices are marked as failed when a threshold number of dies fail, then data protection is improved, but storage capacity and performance deteriorate
Solution Approach 1:
The storage device is segmented into multiple independent dies, allowing individual dies to be marked as failed while the storage device itself remains operational. This segmentation enables selective deactivation of problematic dies while preserving functionality and capacity of healthy dies, resolving the contradiction between data protection and storage capacity utilization.
Solution Approach 2:
Different quality states are applied to different dies within the same storage device. Each die can be independently assessed and marked as likely to fail or suitable for data storage based on its own health status. This local quality differentiation allows the system to maximize usable capacity by utilizing only the healthy portions of the storage device while maintaining data protection through isolation of problematic segments.
2Reliability
If storage devices are marked as failed when a threshold number of dies fail, then data protection is improved, but system performance deteriorates
Solution Approach 1:
By segmenting the storage device into independently manageable dies, the system can isolate failed dies and continue operating with healthy ones. This prevents complete system failure and maintains performance levels based on the capacity of functioning dies, rather than forcing a total shutdown that would eliminate productivity entirely.
Solution Approach 2:
Instead of completely disabling the storage device when any die fails, the system applies partial action by selectively deactivating only the specific dies that are likely to fail. This partial deactivation maintains system performance using the remaining healthy dies, achieving data protection without the excessive penalty of total system shutdown.
3Reliability
If individual dies are marked as likely to fail and used in read-only mode, then data retention is improved, but write capability deteriorates
Solution Approach 1:
Different operational modes are assigned to different dies based on their health status. Healthy dies maintain full read-write capability, while dies marked as likely to fail are restricted to read-only mode. This local quality differentiation preserves data retention for critical information on vulnerable dies while maintaining write capability on reliable dies, achieving both goals simultaneously.
Solution Approach 2:
The storage device is segmented into different functional zones: read-write zones on healthy dies and read-only zones on vulnerable dies. This segmentation allows the system to preserve data on all dies while maintaining write adaptability on the healthy portions, resolving the contradiction between data retention and write capability.
Data Source
AI summary
An indication that power loss protection (PLP) for a non-volatile random access memory (NVRAM) portion of a storage device has failed is received from the storage device. The storage device is marked to indicate the storage device has a PLP failure. A write path to store data at the storage device is modified to avoid storing the data in the NVRAM portion of the storage device that has the PLP failure.