Presentation is loading. Please wait.

Presentation is loading. Please wait.

Extracting Robust Keys from NAND Flash Physical Unclonable Functions Shijie Jia, Luning Xia, Zhan Wang, Jingqiang Lin, Guozhu Zhang and Yafei Ji Institute.

Similar presentations


Presentation on theme: "Extracting Robust Keys from NAND Flash Physical Unclonable Functions Shijie Jia, Luning Xia, Zhan Wang, Jingqiang Lin, Guozhu Zhang and Yafei Ji Institute."— Presentation transcript:

1 Extracting Robust Keys from NAND Flash Physical Unclonable Functions Shijie Jia, Luning Xia, Zhan Wang, Jingqiang Lin, Guozhu Zhang and Yafei Ji Institute of Information Engineering, CAS, Beijing, China ISC 2015 11th September,2015 1

2 Outline  Motivation  NAND Flash memory basics  Robust key generation  Experiments and Evaluation  Conclusion 2

3 Outline  Motivation  NAND Flash memory basics  Robust key generation  Experiments and Evaluation  Conclusion 3

4  NAND Flash memory is ubiquitous  Smart phones, SD cards, USB memory stick, etc.  The keys used by electronic devices need be protected  contain many confidential information  many applications need to identify and authenticate users Motivation 4

5 Physically uncloable function (PUF)  PUFs are hardware primitives which produce unpredictable and instantiation dependent outcomes. Noises exist in the PUFs  PUF responses are generally not perfectly reproducible.  In general, fuzzy extractors (ECC and hash function )are used to ensure the reliability of the responses.  As the error rate increases, the cost of fuzzy extractor is rather high for devices with limited hardware resources. 5

6 Motivation Physically uncloable function (PUF)  PUFs are hardware primitives which produce unpredictable and instantiation dependent outcomes. Noises exist in the PUFs  PUF responses are generally not perfectly reproducible.  In general, fuzzy extractors (ECC and hash function )are used to ensure the reliability of the responses.  As the error rate increases, the cost of fuzzy extractor is rather high for devices with limited hardware resources. 6

7  We focus on  NAND Flash Physical Unclonable Function (NFPUF)  Error reduction techniques  First, we present three methods to extract raw NFPUF numbers.  partial erasure  partial programming  program disturbance  Second, we introduce two methods to select the cells.  bit-map  position-map Motivation 7

8 Outline  Motivation  NAND Flash memory basics  Robust key generation  Experiments and Evaluation  Conclusion 8

9 NAND Flash memory basics  Uncertain States of Cells  on account of variations in manufacturing processes, the threshold voltages vary from cell to cell.  When the threshold voltage is not shifted sufficiently from the programmed state to the erased state, and vice versa, the cell will be in an uncertain state. 9

10 NAND Flash memory basics  Uncertain States of Cells  on account of variations in manufacturing processes, the threshold voltages vary from cell to cell.  When the threshold voltage is not shifted sufficiently from the programmed state to the erased state, and vice versa, the cell will be in an uncertain state. 10

11 NAND Flash memory basics  Disturbance Related to Array Organization  During the programming operations,there exists noises between the adjacent cells.  After many repeating operations, it makes the adjacent cells flip 11

12 NAND Flash memory basics  Disturbance Related to Array Organization  During the programming operations,there exists noises between the adjacent cells.  After many repeating operations, it makes the adjacent cells flip 12

13 Outline  Motivation  NAND Flash memory basics  Robust key generation  Experiments and Evaluation  Conclusion 13

14  Partial Erasure  Erase a block  program a page to “0”  perform fixed number (PENum) of partial erasure operations (Te) to the selected page  record the number of partial erasure operations that the selected cells need to flip 14 Robust key generation ---- Extracting Raw NFPUF Numbers 1 1 1 1 0 0 0 0 0 1 0 0 1 0 0 0 0 1 1 1 0 0 1 0 1 0 2 1 1 0 1 1 1 1 0 3 Raw NFPUF Numbers: 2 1 4 3 1 3 2 4

15  Partial Erasure  Erase a block  program a page to “0”  perform fixed number (PENum) of partial erasure operations (Te) to the selected page  record the number of partial erasure operations that the selected cells need to flip 15 Robust key generation ---- Extracting Raw NFPUF Numbers 1 1 1 1 0 0 0 0 0 1 0 0 1 0 0 0 0 1 1 1 0 0 1 0 1 0 2 1 1 0 1 1 1 1 0 3 Raw NFPUF Numbers: 2 1 4 3 1 3 2 4

16  Partial Programming  Erase a block  perform fixed number (PPNum) of partial programming operations (Tp) to the selected page  record the number of partial programming operations that the selected cells need to flip 16 Robust key generation ---- Extracting Raw NFPUF Numbers 1 1 1 1 1 0 1 1 0 1 1 1 0 1 0 0 1 1 0 1 0 1 2 0 0 1 0 0 0 0 1 3 Raw NFPUF Numbers: 2 1 4 3 1 3 2 4

17  Partial Programming  Erase a block  perform fixed number (PPNum) of partial programming operations (Tp) to the selected page  record the number of partial programming operations that the selected cells need to flip 17 Robust key generation ---- Extracting Raw NFPUF Numbers 1 1 1 1 1 0 1 1 0 1 1 1 0 1 0 0 1 1 0 1 0 1 2 0 0 1 0 0 0 0 1 3 Raw NFPUF Numbers: 2 1 4 3 1 3 2 4

18  Program Disturbance  Erase a block  perform fixed number (PDNum) of programming operations to the selected page  record the number of programming operations that the selected cells in its physically adjacent page need to flip 18 Robust key generation ---- Extracting Raw NFPUF Numbers 1 1 1 1 1 0 1 1 0 1 1 1 0 1 0 0 1 1 0 1 0 1 2 0 0 1 0 0 0 0 1 3 Raw NFPUF Numbers: 2 1 4 3 1 3 2 4

19  Program Disturbance  Erase a block  perform fixed number (PDNum) of programming operations to the selected page  record the number of programming operations that the selected cells in its physically adjacent page need to flip 19 Robust key generation ---- Extracting Raw NFPUF Numbers 1 1 1 1 1 0 1 1 0 1 1 1 0 1 0 0 1 1 0 1 0 1 2 0 0 1 0 0 0 0 1 3 Raw NFPUF Numbers: 2 1 4 3 1 3 2 4

20 20 Robust key generation ---- Extracting Robust Keys from Raw NFPUF Numbers Fluctuations of raw NFPUF numbers from four cells of a MLC type NAND Flash memory chip

21 21 Robust key generation ---- Extracting Robust Keys from Raw NFPUF Numbers

22 22 Robust key generation ---- Extracting Robust Keys from Raw NFPUF Numbers

23 23 Robust key generation ---- Extracting Robust Keys from Raw NFPUF Numbers

24 Outline  Motivation  NAND Flash memory basics  Robust key generation  Experiments and Evaluation  Conclusion 24

25 25 Experiments  Setup  Platform: a Flash test board with an ARM Cortex-M3 controller  Tested NAND Flash memory chips

26 26 Evaluation  The security and reliability of the keys  speed (for performance)  reproducibility (for reliability)  uniqueness (for security)  randomness (for high-entropy)

27 27 Evaluation ——speed The average throughput of the raw NFPUF numbers (Kbits/second) Parameters setting of the raw NFPUF numbers extraction methods

28 28 Evaluation ——speed The average throughput of the raw NFPUF numbers (Kbits/second) Parameters setting of the raw NFPUF numbers extraction methods

29 29 Evaluation ——Reproducibility ( Intra-chip variations )

30 30 Evaluation —— Uniqueness ( Inter-chip variations ) The inter-chip variations of the three proposed extraction methods

31 31 Evaluation —— Randomness The percentage of bit “1” with temperature changes

32 Outline  Motivation  NAND Flash memory basics  Robust key generation  Experiments and Evaluation  Conclusion 32

33 33  First, we proposed three methods to extract raw NFPUF numbers from NAND Flash memory chips.  Second, we utilized the bit-map or position-map method to select the cells with the most reliable relationship of the size between raw NFPUF numbers.  At last, we evaluated the primary characteristics of the key under various temperature and aging conditions.  Our key generator eschews the costly ECC overheads to generate robust and error-free keys.

34 34


Download ppt "Extracting Robust Keys from NAND Flash Physical Unclonable Functions Shijie Jia, Luning Xia, Zhan Wang, Jingqiang Lin, Guozhu Zhang and Yafei Ji Institute."

Similar presentations


Ads by Google