Systems and methods for identifying counterfeit memory
Inventors
Rahman, M. Tauhidur • Bahar Talukder, Bashir Mohammad Sabquat
Assignees
University of Alabama in Huntsville
Publication Number
US-11139043-B2
Publication Date
2021-10-05
Expiration Date
2040-05-20
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
Due to design variations, process variations, manufacturing variations, and other factors, memory of a certain type manufactured by one manufacturer often exhibits a unique pattern of performance characteristics relative to the patterns of the same type of performance characteristics exhibited by memory from other manufacturers. A system for identifying counterfeit memory is trained to learn the different patterns of performance characteristics for different manufacturers of memory. Thereafter, the system may analyze the performance of a given memory device to determine whether the memory device has been manufactured by a particular manufacturer. Thus, the system is capable of determining whether the memory device is counterfeit (e.g., has been manufactured by an unexpected manufacturer).
Core Innovation
The invention pertains to systems and methods for identifying counterfeit memory based on unique patterns of performance characteristics that arise from design, process, and manufacturing variations specific to each manufacturer. A memory analyzer is trained to learn the different patterns or correlations present in the performance signatures of memory devices produced by various manufacturers. This trained system can subsequently analyze the performance characteristics of a given memory device, such as power-up default values, latency errors, or refresh errors, to determine whether the device originates from a particular manufacturer.
The problem being solved is the widespread proliferation of counterfeit memory chips and modules within the semiconductor supply chain. Counterfeit memory poses significant risks due to its sub-standard quality, reduced reliability, shorter lifespan, and increased vulnerability to security attacks. Detecting counterfeit memory is challenging because such devices can often pass initial testing; thus, an improved technique is needed to identify counterfeit memory rapidly, accurately, and cost-effectively.
The invention provides techniques wherein performance signatures, such as power-up default signatures, latency signatures, and refresh signatures, are collected from memory devices and analyzed for learned correlations that are characteristic of specific manufacturers. These correlations are identified by memory analysis logic, which may utilize machine learning algorithms trained on data from multiple devices of known origin. The system is thereby capable of confirming the manufacturer or flagging memory as counterfeit if the performance signature does not correspond to any learned pattern.
Claims Coverage
There are two independent claims, each covering distinct inventive features related to detecting the origin of a memory device and identifying counterfeit memory using performance signatures and trained analysis logic.
System for identifying memory origin using trained analysis of performance signatures
A system comprising: - A memory device. - A memory analyzer coupled to the memory device, containing at least one processor programmed with memory analysis logic. - The memory analysis logic is trained to learn correlations between multiple performance signatures from memory devices manufactured by a manufacturer. - The logic receives at least one performance signature from the device, where the signature is uniquely generated by the memory device based on at least manufacturing variations. - The logic identifies the origin of the memory device based on the performance signature and the learned correlations. - The system provides an output indicating the origin.
Method for determining memory device origin via analysis of performance signatures
A method comprising: 1. Receiving, at a memory analyzer, at least one performance signature from a memory device, where the performance signature is based on the performance of the memory device in memory operations and is uniquely generated due to manufacturing variations. 2. The memory analyzer is equipped with at least one processor programmed with memory analysis logic trained to learn correlations among performance signatures from known manufacturers. 3. Analyzing the performance signature with respect to the learned correlations. 4. Identifying the origin of the memory device based on this analysis. 5. Providing an output indicating the identified origin.
The independent claims cover both a system and a method that use trained memory analysis logic to process performance signatures from memory devices in order to determine and indicate the origin of those devices, enabling identification of counterfeit memory.
Stated Advantages
The system enables rapid, accurate, and inexpensive detection of counterfeit memory devices.
Identifying memory device origin helps mitigate the deployment of sub-standard, unreliable, or vulnerable memory within electronic systems.
Documented Applications
Testing memory devices within personal electronic devices such as cellular telephones or personal computers to identify the origin or detect counterfeiting.
Use in stand-alone electronic testing devices specifically designed to test memory modules or chips for authenticating their manufacturer origin.
Interested in licensing this patent?