Non-Volatile Memory Metadata for Read Status Differentiation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing non-volatile memory systems face challenges in accurately determining the status of memory locations, leading to misinterpretation of data integrity and failure to distinguish between successful and failed read operations, especially due to phenomena like program disturb which can cause memory locations to change state.
Innovation Solution
Incorporating a fixed predetermined sequence in the metadata field of data vectors stored in non-volatile memory, ensuring that even all-zero vectors are not misinterpreted as valid data, by introducing a non-zero value that differentiates them from no-access vectors and valid codewords, thereby preventing misinterpretation of read operations and ensuring correct data integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed predetermined sequence is added to the metadata field, then the ability to distinguish successful reads from failed reads improves, but the device complexity increases
Solution Approach 1:
A fixed predetermined sequence is pre-defined and embedded in the metadata field of every data vector before storage. This preliminary action ensures that when data is read back, the system can immediately compare the metadata against this known sequence to determine read success, eliminating the need for complex post-read analysis and enabling reliable distinction between successful and failed read operations.
2Quantity of substance
If all-zero vectors are allowed in metadata, then storage efficiency improves, but the ability to distinguish valid data from no-access vectors deteriorates
Solution Approach 1:
The metadata field is given a distinct local quality by incorporating a fixed predetermined sequence that is non-zero. This creates a clear distinction between valid data vectors (which contain the predetermined sequence) and no-access vectors (which would be all-zero or contain invalid patterns). The local quality of the metadata field thus differs from the user data field, enabling reliable status determination without sacrificing storage efficiency.
Data Source
AI summary
Systems and methods are provided for storing data in a portion of a non-volatile memory (“NVM”) such that the status of the NVM portion can be determined with high probability on a subsequent read. An NVM interface, which may receive write commands to store user data in the NVM, can store a fixed predetermined sequence (“FPS”) with the user data. The FPS may ensure that a successful read operation on a NVM portion is not misinterpreted as a failed read operation or as an erased NVM portion. For example, if the NVM returns an all-zero vector when a read request fails, the FPS can include at least one “1” or one “0”, as appropriate, to differentiate between successful and unsuccessful read operations. In some embodiments, the FPS may also be used to differentiate between disturbed data, which passes an error correction check, and correct data.