隨著嵌入式系統與智慧邊緣應用的快速發展，資料儲存的安全性與可靠性日益重要。傳統的資料驗證與修復機制多仰賴主機端運算資源進行處理，不僅增加系統延遲，也提高資料搬移的成本與風險。為解決此問題，本研究基於儲存裝置內運算（In-Storage Processing, ISP）概念，於雙核心嵌入式平台上實作一套結合安全雜湊演算法（SHA）與多副本資料備援之資料完整性驗證與自動修復機制。
本系統採用雙核心異質架構，其中一顆核心作為系統處理器，運行 Linux 作業系統，負責接收主機端命令，並將資料與其對應的 SHA 雜湊值一併傳送至 ISP 處理核心。系統處理器在傳送資料前，先以 SHA-256 或 SHA-512 進行雜湊計算，並將所得雜湊值作為資料的數位簽章（Digital Signature），與原始資料一同儲存於三個獨立儲存區域中，實現多副本備援架構，以提升資料可用性。
本研究以人臉辨識應用為例，針對每筆人臉特徵向量計算對應的 SHA 雜湊值，並將其作為資料完整性驗證之依據。透過此方式，可有效判斷特徵資料是否遭到毀損，即使在多副本儲存架構下，仍能確保每筆人臉特徵的一致性與完整性，有效提升生物特徵資料的可信度。
在資料讀取階段，系統會針對三份副本進行雜湊驗證，若發現某一副本異常，將根據其餘兩份資料進行比對，並執行自動化修復。另一方面，ISP 核心則以 Bare-metal 韌體方式運行，負責對本地 Flash 中的單一副本執行定期資料完整性檢查，透過重新計算 SHA-256 雜湊值與既有記錄進行比對，協助系統及早偵測潛在資料異常。
實驗結果顯示，在 無毀損情境下，對 5000 筆人臉特徵向量進行 SHA-256/512 驗證平均耗時僅 943/873 ms。在 模擬單一副本毀損情境下，本研究針對 0 至 1000 筆毀損資料，以每 200 筆為區間進行測試，結果顯示系統能於 1400至2000 ms 的時間內完成修復。

With the growth of embedded systems and edge applications, ensuring secure and reliable data storage has become increasingly important. Traditional verification and recovery rely on host-side processing, causing higher latency and data movement overhead. This work presents an In-Storage Processing (ISP) approach on a dual-core embedded platform, integrating Secure Hash Algorithms (SHA-256/512) with multi-replica redundancy for data integrity verification and automatic recovery. One core, running Linux, handles host communication and hash computation, while the bare-metal ISP core performs periodic integrity checks on local Flash. Using facial recognition as a case study, the system verifies and repairs corrupted replicas by comparing hash values across three copies, ensuring data consistency and improving availability of biometric data.

[1] A. Barbalace and J. Do, “Computational Storage: Where Are We Today?,” vol. Innovative Data Systems Research, Virtual Confernece, pp. 1–6, Jan. 2021.
[2] A. HeydariGorji, M. Torabzadehkashi, S. Rezaei, H. Bobarshad, V. Alves, and P. H. Chou, “In-storage Processing of I/O Intensive Applications on Computational Storage Drives,” Dec. 23, 2021, arXiv: arXiv:2112.12415. doi: 10.48550/arXiv.2112.12415.
[3] M. Torabzadehkashi, “SoC-Based In-Storage Processing: Bringing Flexibility and Efficiency to Near-Data Processing,” Ph.D. Dissertation, University of California, Irvine, United States -- California, 2019. [Online]. Available: https://www.proquest.com/dissertations-theses/soc-based-storage-processing-bringing-flexibility/docview/2394394763/se-2?accountid=12719
[4] T. Wang et al., “NICE: A Nonintrusive In-Storage-Computing Framework for Embedded Applications,” IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst., vol. 43, no. 11, pp. 3876–3887, Nov. 2024, doi: 10.1109/tcad.2024.3446857.
[5] S. Kang et al., “ISP Agent: A Generalized In-storage-processing Workload Offloading Framework by Providing Multiple Optimization Opportunities,” ACM Trans. Archit. Code Optim., vol. 21, no. 1, pp. 1–24, Mar. 2024, doi: 10.1145/3632951.
[6] Y.-C. Liu, K.-C. Hsu, and H.-W. Tseng, “Rethinking Programming Frameworks for In-Storage Processing,” in 2023 60th ACM/IEEE Design Automation Conference (DAC), San Francisco, CA, USA: IEEE, July 2023, pp. 1–6. doi: 10.1109/DAC56929.2023.10247919.
[7] J. Do, S. Sengupta, and S. Swanson, “Programmable Solid-State Storage in Future Cloud Datacenters,” Commun. ACM, vol. 62, no. 6, pp. 54–62, May 2019, doi: 10.1145/3286588.
[8] J. Xue, R. Chen, and Z. Shao, “SoftSSD: Software-defined SSD Development Platform for Rapid Flash Firmware Prototyping,” in 2022 IEEE 40th International Conference on Computer Design (ICCD), Olympic Valley, CA, USA: IEEE, Oct. 2022, pp. 602–609. doi: 10.1109/iccd56317.2022.00094.
[9] D. Gouk, M. Kwon, H. Bae, and M. Jung, “Containerized In-Storage Processing Model and Hardware Acceleration for Fully-Flexible Computational SSDs,” IEEE Comput. Archit. Lett., pp. 1–4, 2024, doi: 10.1109/lca.2023.3289828.
[10] J. Do et al., “Cost-effective, Energy-efficient, and Scalable Storage Computing for Large-scale AI Applications,” ACM Trans. Storage, vol. 16, no. 4, pp. 1–37, Nov. 2020, doi: 10.1145/3415580.
[11] P. K C, R. Soman, and P. Honnavalli, “Validity of Forensic Evidence using Hash Function,” in 2020 5th International Conference on Communication and Electronics Systems (ICCES), COIMBATORE, India: IEEE, June 2020, pp. 823–826. doi: 10.1109/icces48766.2020.9138061.
[12] D. Meiri and A. Kucherov, “Hash-Based Data Recovery from Remote Storage System Responsive to Missing or Corrupted Hash Digest,” US10996871B2, May 04, 2021 Accessed: July 19, 2025. [Online]. Available: https://patents.google.com/patent/US10996871B2/en
[13] L. Beringer, K. Q. Ye, and A. W. Appel, “Verified Correctness and Security of OpenSSL HMAC,” presented at the 24th USENIX Security Symposium, Washington, D.C., USA: USENIX Association, Aug. 2015, pp. 207–221.
[14] P. K. Yien, K. Malik, and S. N. Ramli, “Comparative Study on Hash Function Algorithms for Blockchain Technology,” vol. 5, no. 1, 2024.
[15] National Institute of Standards and Technology (US), “Secure hash standard,” National Institute of Standards and Technology (U.S.), Washington, D.C., NIST FIPS 180-4, 2015. doi: 10.6028/NIST.FIPS.180-4.
[16] Crucial by Micron, “SMART與SSD（固態硬碟）,” Crucial Taiwan. Accessed: July 07, 2025. [Online]. Available: https://www.crucial.tw/articles/about-ssd/smart-and-ssds
[17] R. Micheloni, Ed., 3D Flash Memories. Dordrecht: Springer Netherlands, 2016. doi: 10.1007/978-94-017-7512-0.
[18] R. G. Gallager, “Low-Density Parity-Check Codes,” Ph.D. Dissertation, Massachusetts Institute of Technology, Department of Electrical Engineering. [Online]. Available: https://web.stanford.edu/class/ee388/papers/ldpc.pdf
[19] Q. Li et al., “Characterizing and Optimizing LDPC Performance on 3D NAND Flash Memories,” ACM Trans. Archit. Code Optim., vol. 21, no. 3, pp. 1–26, Sept. 2024, doi: 10.1145/3663478.
[20] “GitHub - openzfs/zfs: OpenZFS on Linux and FreeBSD.” Accessed: July 08, 2025. [Online]. Available: https://github.com/openzfs/zfs
[21] W. A. Bhat and M. A. Wani, “Forensic analysis of B-tree file system (Btrfs),” Digit. Investig., vol. 27, pp. 57–70, Dec. 2018, doi: 10.1016/j.diin.2018.09.001.
[22] O. Rodeh, J. Bacik, and C. Mason, “BTRFS: The Linux B-Tree Filesystem,” ACM Trans. Storage, vol. 9, no. 3, pp. 1–32, Aug. 2013, doi: 10.1145/2501620.2501623.
[23] Su Y.-D., Design and Implementation of an Embedded System-Based In-Storage Processing Prototype, M.S. thesis, Department of Engineering Science, National Cheng Kung University, Tainan, Taiwan, 2024.
[24] Lai T.-Y., Design and Implementation of an Embedded System Based In-Storage Processing Prototype, M.S. thesis, Department of Engineering Science, National Cheng Kung University, Tainan, Taiwan, 2024.
[25] Nuvoton Technology Corp., “MA35D1_Linux_Applications/examples/rpmsg-v2,” GitHub. Accessed: June 28, 2025. [Online]. Available: https://github.com/OpenNuvoton/MA35D1_Linux_Applications/tree/master/examples/rpmsg-v2