Two Birds With One Stone: Differential Privacy by Low-Power SRAM Memory
Authors: 
Jianqing Liu, Na Gong, Hritom DasAuthors Info & Claims
Jianqing Liu, Na Gong, Hritom DasAuthors Info & ClaimsPages 5706 - 5719
Abstract
The software-based implementation of differential privacy mechanisms has been shown to be neither friendly for lightweight devices nor secure against side-channel attacks. In this work, we aim to develop a hardware-based technique to achieve differential privacy by design. In contrary to the conventional software-based noise generation and injection process, our design realizes local differential privacy (LDP) by harnessing the inherent hardware noise into controlled LDP noise when data is stored in the memory. Specifically, the noise is tamed through a novel memory design and power downscaling technique, which leads to double-faceted gains in privacy and power efficiency. A well-round study that consists of theoretical design and analysis and chip implementation and experiments is presented. The results confirm that the developed technique is differentially private, saves 88.58% system power, speeds up software-based DP mechanisms by more than <inline-formula><tex-math notation="LaTeX">$10^{6}$</tex-math><alternatives><mml:math><mml:msup><mml:mn>10</mml:mn><mml:mn>6</mml:mn></mml:msup></mml:math><inline-graphic xlink:href="liu-ieq1-3382630.gif"/></alternatives></inline-formula> times, while only incurring 2.46% chip overhead and 7.81% estimation errors in data recovery.
References
[1]
Y. Zhou, D. Wilkinson, R. Schreiber, and R. Pan, “Large-scale parallel collaborative filtering for the netflix prize,” in Proc. Int. Conf. Algorithmic Appl. Manage., Springer, 2008, pp. 337–348.
[2]
C. Dwork, F. McSherry, K. Nissim, and A. Smith, “Calibrating noise to sensitivity in private data analysis,” in Proc. Theory Cryptogr. Conf., Springer, 2006, pp. 265–284.
[3]
A. Machanavajjhala, D. Kifer, J. Abowd, J. Gehrke, and L. Vilhuber, “Privacy: Theory meets practice on the map,” in Proc. IEEE 24th Int. Conf. Data Eng., 2008, pp. 277–286.
[4]
F. McSherry and I. Mironov, “Differentially private recommender systems: Building privacy into the netflix prize contenders,” in Proc. 15th ACM SIGKDD Int. Conf. Knowl. Discov. Data Mining, 2009, pp. 627–636.
[5]
Z. Qin, Y. Yang, T. Yu, I. Khalil, X. Xiao, and K. Ren, “Heavy hitter estimation over set-valued data with local differential privacy,” in Proc. ACM SIGSAC Conf. Comput. Commun. Secur., 2016, pp. 192–203.
[6]
Ú. Erlingsson, V. Pihur, and A. Korolova, “RAPPOR: Randomized aggregatable privacy-preserving ordinal response,” in Proc. ACM SIGSAC Conf. Comput. Commun. Secur., 2014, pp. 1054–1067.
[7]
I. Mironov, “On significance of the least significant bits for differential privacy,” in Proc. ACM Conf. Comput. Commun. Secur., 2012, pp. 650–661.
[8]
M. Andrysco, D. Kohlbrenner, K. Mowery, R. Jhala, S. Lerner, and H. Shacham, “On subnormal floating point and abnormal timing,” in Proc. IEEE Symp. Secur. Privacy, 2015, pp. 623–639.
[9]
A. Haeberlen, B. C. Pierce, and A. Narayan, “Differential privacy under fire,” in Proc. USENIX Secur. Symp., 2011.
[10]
J. Liu and N. Gong, “Privacy by memory design: Visions and open problems,” IEEE Micro, vol. 44, no. 1, pp. 49–58, Jan./Feb. 2024.
[11]
S. Y. Yu, H. Jiang, S. Huang, X. Peng, and A. Lu, “Compute-in-memory chips for deep learning: Recent trends and prospects,” IEEE Circuits Syst. Mag., vol. 21, no. 3, pp. 31–56, Third Quarter 2021.
[12]
S. L. Warner, “Randomized response: A survey technique for eliminating evasive answer bias,” J. Amer. Stat. Assoc., vol. 60, no. 309, pp. 63–69, 1965.
[13]
J. Fu, Z. Liao, J. Liu, S. C. Smith, and J. Wang, “Memristor-based variation-enabled differentially private learning systems for edge computing in IoT,” IEEE Internet Things J., vol. 8, no. 12, pp. 9672–9682, Jun. 2021.
[14]
Y. Xu, H. Das, Y. Gong, and N. Gong, “On mathematical models of optimal video memory design,” IEEE Trans. Circuits Syst. Video Technol., vol. 30, no. 1, pp. 256–266, Jan. 2020.
[15]
N. Gong, S. Jiang, J. Wang, B. Aravamudhan, K. Sekar, and R. Sridhar, “Hybrid-cell register files design for improving NBTI reliability,” Microelectronics Rel., vol. 52, no. 9/10, pp. 1865–1869, 2012.
[16]
J. Croon, S. Decoutere, W. Sansen, and H. Maes, “Physical modeling and prediction of the matching properties of MOSFETS,” in Proc. 30th Eur. Solid-State Circuits Conf., 2004, pp. 193–196.
[17]
H. Das, A. A. Haidous, S. C. Smith, and N. Gong, “Flexible low-cost power-efficient video memory with ecc-adaptation,” IEEE Trans. Very Large Scale Integration Syst., vol. 29, no. 10, pp. 1693–1706, Oct. 2021.
[18]
M. Gottscho, A. BanaiyanMofrad, N. D. Dutt, A. Nicolau, and P. Gupta, “DPCS: Dynamic power/capacity scaling for SRAM caches in the nanoscale era,” ACM Trans. Archit. Code Optim., vol. 12, no. 3, pp. 1–27, 2015.
[19]
J. Edstrom, D. Chen, Y. Gong, J. Wang, and N. Gong, “Data-pattern enabled self-recovery low-power storage system for big video data,” IEEE Trans. Big Data, vol. 5, no. 1, pp. 95–105, Mar. 2019.
[20]
Y. Wang, R. Paccagnella, E. T. He, H. Shacham, C. W. Fletcher, and D. Kohlbrenner, “Hertzbleed: Turning power {Side-Channel} attacks into remote timing attacks on x86,” in Proc. 31st USENIX Secur. Symp., 2022, pp. 679–697.
[21]
D. R. Dipta and B. Gulmezoglu, “DF-SCA: Dynamic frequency side channel attacks are practical,” in Proc. 38th Annu. Comput. Secur. Appl. Conf., 2022, pp. 841–853.
[22]
T. Murakami and Y. Kawamoto, “Utility-optimized local differential privacy mechanisms for distribution estimation,” in Proc. 28th {USENIX} Secur. Symp., 2019, pp. 1877–1894.
[23]
T. Wang, J. Blocki, N. Li, and S. Jha, “Locally differentially private protocols for frequency estimation,” in Proc. 26th {USENIX} Secur. Symp., 2017, pp. 729–745.
[24]
D. Yang, D. Zhang, V. W. Zheng, and Z. Yu, “Modeling user activity preference by leveraging user spatial temporal characteristics in lbsns,” IEEE Trans. Syst., Man, Cybern. Syst., vol. 45, no. 1, pp. 129–142, Jan. 2015.
[25]
N. Holohan, S. Braghin, P. Mac Aonghusa, and K. Levacher, “Diffprivlib: The IBM differential privacy library,” 2019,.
[26]
Digi-Key, “Use advanced LDOs to meet IoT wireless sensor power supply design challenges,” 2019. [Online]. Available: https://www.digikey.com/en/articles/use-advanced-ldos-iot-wireless-sensor-power-supply-design
[27]
H. Kwon, D. Kim, Y. H. Kim, and S. Kang, “Variation-aware SRAM cell optimization using deep neural network-based sensitivity analysis,” IEEE Trans. Circuits Syst. I: Reg. Papers, vol. 68, no. 4, pp. 1567–1577, Apr. 2021.
[28]
K. A. Bowman, “Adaptive and resilient circuits: A tutorial on improving processor performance, energy efficiency, and yield via dynamic variation,” IEEE Solid-State Circuits Mag., vol. 10, no. 3, pp. 16–25, Summer 2018.
[29]
J. Tschanz et al., “Adaptive frequency and biasing techniques for tolerance to dynamic temperature-voltage variations and aging,” in Proc. IEEE Int. Solid-State Circuits Conf. Dig. Tech. Papers, 2007, pp. 292–604.
[30]
J. Liu, C. Zhang, B. Lorenzo, and Y. Fang, “DPavatar: A real-time location protection framework for incumbent users in cognitive radio networks,” IEEE Trans. Mobile Comput., vol. 19, no. 3, pp. 552–565, Mar. 2020.
[31]
X. Pei, X. Deng, S. Tian, J. Liu, and K. Xue, “Privacy-enhanced graph neural network for decentralized local graphs,” IEEE Trans. Inf. Forensics Secur., vol. 19, pp. 1614–1629, 2023.
[32]
N. Gai, K. Xue, B. Zhu, J. Yang, J. Liu, and D. He, “An efficient data aggregation scheme with local differential privacy in smart grid,” Digit. Commun. Netw., vol. 8, no. 3, pp. 333–342, 2022.
[33]
C. Ilvento, “Implementing the exponential mechanism with base-2 differential privacy,” in Proc. ACM SIGSAC Conf. Comput. Commun. Secur., 2020, pp. 717–742.
[34]
L. Yang and B. Murmann, “Approximate SRAM for energy-efficient, privacy-preserving convolutional neural networks,” in Proc. IEEE Comput. Soc. Annu. Symp. VLSI, 2017, pp. 689–694.
[35]
J. Hsu et al., “Differential privacy: An economic method for choosing epsilon,” in Proc. IEEE 27th Comput. Secur. Found. Symp., 2014, pp. 398–410.
[36]
L. T. Clark, S. B. Medapuram, and D. K. Kadiyala, “SRAM circuits for true random number generation using intrinsic bit instability,” IEEE Trans. Very Large Scale Integration Syst., vol. 26, no. 10, pp. 2027–2037, Oct. 2018.
Index Terms
- Two Birds With One Stone: Differential Privacy by Low-Power SRAM Memory
Index terms have been assigned to the content through auto-classification.
Recommendations
Next high performance and low power flash memory package structure
In general, NAND flash memory has advantages in low power consumption, storage capacity, and fast erase/write performance in contrast to NOR flash. But, main drawback of the NAND flash memory is the slow access time for random read operations. Therefore,...
Clock Power Analysis of Low Power Clock Gated Arithmetic Logic Unit on Different FPGA
CICN '14: Proceedings of the 2014 International Conference on Computational Intelligence and Communication NetworksThis paper, deals with Latch Free Clock Gating technique for reduction of clock power and total power consumption in Low Power Arithmetic and Logic Unit and we have analysed power reduction on different FPGA devices. Without latch free clock gating ...
Racetrack Memory based hybrid Look-Up Table (LUT) for low power reconfigurable computing
AbstractThe large area and high power consumption are the two main bottlenecks in the conventional SRAM-based Field Programmable Gate Arrays (FPGAs). In recent works, resistive Non-Volatile Memories (NVMs) have been widely proposed to tackle the above ...
Highlights- A hybrid LUT with one non-volatile input and multiple volatile inputs.
- The computation and storage are integrated in the non-volatile input.
- The hybrid LUT reduces the area and leakage power with extra reconfigurable resource.
- ...
Comments
Information & Contributors
Information
Published In
1545-5971 © 2024 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See https://www.ieee.org/publications/rights/index.html for more information.
Publisher
IEEE Computer Society Press
Washington, DC, United States
Publication History
Published: 01 November 2024
Qualifiers
- Research-article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 0Total Downloads
- Downloads (Last 12 months)0
- Downloads (Last 6 weeks)0
Reflects downloads up to 11 Feb 2025