SECURE NVM · PHYSICAL SECURITY ASSURANCE

Turn security claims into reproducible evidence

The security value of permanent data and an ephemeral root key must be measured, challenged, improved and scoped against realistic attack models.

OBSERVABLEPOWER · EM · TIMING
DISTURBANCEVOLTAGE · CLOCK · EM · LASER
EXPECTED OUTCOMEFAIL SECURE · NO KEY DISCLOSURE
THREATMETHODOBSERVEMITIGATEEVIDENCEASSURE

FAIL CLOSED失敗即關閉

Claims stay testable, or they stay unmarked主張必須可測,否則維持未標

Named-condition figures do not upgrade themselves into a certification.具名條件下的數字,不會自行升級成認證。

Side-channel evidence is a measurement problem. It does not name a product defect by itself.旁路證據是量測問題,本身不構成產品缺陷名稱。

00 / OPERATING PRINCIPLE

State what is known—and mark what is not

PUBLIC FACTVENDOR CLAIMANALYST INFERENCEVERIFIEDUNKNOWN

01 / PHYSICAL ATTACK SURFACE

Observe, disturb, read: three different evidence problems

Attack classes are not a fear list. They turn product assumptions into testable security requirements.

OBSERVE

Side-Channel Analysis (DPA/CPA)

Measure transient power (DPA/CPA), electromagnetic emissions (EM), timing or contactless traces to verify that complementary differential read pairs and algorithmic masking attenuate leakage during PUF reconstruction, key derivation, and AES execution.

ASSETSPUF response · root/derived key · plaintext window · DPA/CPA attenuation
DISTURB

Fault Injection (LFI/Glitch)

Deploy backside laser fault injection (LFI), voltage glitch, clock disturbance, or EMFI to challenge authentication, lifecycle transitions, control flow (preventing instruction skip), crypto cores, and single-cycle zeroization.

OUTCOMEScheck bypass · DFA · instruction skip · state corruption
READ

Invasive Analysis & Active Shielding

Defend against FIB microprobing and backside silicon thinning. Deploy active top-metal shielding/mesh to trigger single-cycle key zeroization upon mesh breach, paired with address/data scrambling so physical readout cannot yield a usable root secret.

BOUNDARYactive mesh · FIB resistance · ciphertext · scrambling · layout

01b / INVASIVE ATTACK PHYSICS & DEFENSE CHAIN

Invasive Attack Physics: FIB-PVC, TEM Limits & Active Shielding Chain

Physical invisibility does not equate to cryptographic security. Analyze ion-beam contrast, cross-section blind spots, and multi-layer mitigation. Fixed impedance, hit-rate, or single-cycle zeroization figures are literature-level illustrations — not product certification.

MICROSCOPIC CONTRAST

FIB-PVC Passive Voltage Contrast Physics

Following delayering to contacts/metal, a focused Ga⁺ ion beam scans the unpowered die surface, inducing state-dependent charge dissipation:

  • 未編程單元 (高阻 >10¹¹ Ω):閘極介電質完整,離子束沉積正電荷無法洩放,表面電位升高壓制二次電子 (SE) 逸出,在探測器呈暗態 (Dark)。Unprogrammed Cell (>10¹¹ Ω): Intact gate oxide traps positive charge, creating a surface retarding field that suppresses secondary electrons (Dark Contrast).
  • 已編程 AntiFuse (<100 Ω):介電質滲透擊穿形成的矽微絲提供歐姆通路,正電荷瞬間排入基板接地,二次電子產額極高,呈亮態 (Bright)。Programmed AntiFuse (<100 Ω): Ohmic breakdown filament discharges positive charge to substrate ground; high secondary electron yield produces Bright Contrast.
  • 防禦因果:斷電狀態物理可讀;安全依賴陣列位址擾碼、密文存儲與 PUF 短暫根金鑰保護。Defense Implication: Physical readout is viable at rest; security relies on address scrambling, ciphertext, and ephemeral PUF root keys.
CROSS-SECTION LIMIT

Microscopic Limits of TEM Cross-Sectioning

TEM is often misconstrued as an omnipotent extraction tool, but device physics and geometry create severe throughput bottlenecks:

  • 微絲奈米尺度:介電質擊穿微絲直徑僅約 3~8 nm,且在閘極三維空間中呈隨機滲透路徑 (Stochastic 3D Percolation Path)。Nanoscale Filament: Breakdown filaments measure only 3–8 nm in diameter along a stochastic 3D percolation trajectory.
  • 幾何命中率極限:在先進 FinFET / GAA 結構下,FIB 盲切製作 30 nm 超薄樣品 (Lamella) 正好切中該微絲的機率低於 10⁻⁴。Geometric Blind Spot: In FinFET/GAA nodes, blind FIB lamella preparation (30 nm thickness) hits the specific filament cross-section with probability <10⁻⁴.
  • 紀律界限:不可宣稱「TEM 看不到代表安全」;專業攻擊者會轉向 FIB-PVC 或 sMIM,而非盲切 TEM。Discipline Boundary: Never claim 'invisible under TEM equals secure'; attackers employ FIB-PVC or sMIM rather than blind TEM.
ACTIVE MITIGATION

Active Metal Shielding & Multi-layer Defense Chain

Defense against invasive tampering must not rely on a single physical property, but rather on an orchestrated defense-in-depth chain:

  • 主動金屬屏蔽 (Active Top-Metal Mesh):頂層部署差分蛇形感測線,動態注入偽隨機碼 (PRBS),監控互容抗與阻抗漂移。Active Top-Metal Mesh: Top metal layers deploy differential serpentine lines energized with dynamic PRBS to monitor capacitive/impedance drift.
  • 單週期急毀迴路 (Zeroization Loop):偵測到探針侵入或網格切斷時,硬體比較器在單週期內觸發電容放電,清除揮發性工作金鑰。Single-Cycle Zeroization: Breach detection triggers immediate capacitive discharge, zeroizing all ephemeral volatile working keys in a single clock cycle.
  • 殘餘防線:即使攻擊者完全逆向 OTP 陣列實體狀態,所獲數據僅為無主金鑰保護的高熵密文,無法還原任何明文資產。Residual Security: Even full physical extraction of the OTP yields only high-entropy ciphertext unreadable without the erased PUF root.
互動認證實驗室 · 實體攻擊潛力評估 SECURITY LAB · PHYSICAL ATTACK POTENTIAL EVALUATOR

Common Criteria (ISO/IEC 15408 / AVA_VAN.5) 實體攻擊防護力與計分試算器 Common Criteria (ISO/IEC 15408 / AVA_VAN.5) Attack Potential Evaluator

Beyond High (AVA_VAN.5)
硬體縱深防禦措施 (Hardware Countermeasures):HARDWARE COUNTERMEASURES:
攻擊者基準潛力分RAW ATTACK POTENTIAL
—
硬體防禦增益加分DEFENSE BONUS
—
有效抗攻擊總潛力EFFECTIVE POTENTIAL
—
Common Criteria 評估門檻進度條 (AVA_VAN.1 至 AVA_VAN.5):COMMON CRITERIA RATING BAR: VAN.1 (0) | VAN.2 (16) | VAN.3 (21) | VAN.4 (26) | VAN.5 (31+)
—
密碼學隨機性 · 原生熵源與統計檢定 CRYPTOGRAPHIC RANDOMNESS · HARDWARE ENTROPY & STATISTICAL TESTING

晶片實體不可複製功能 (PUF) 空間隨機性與 NIST SP 800-22 統計檢驗器 Physical Unclonable Function (PUF) Spatial Randomness & NIST SP 800-22 Evaluator

SP 800-22 檢定示意 · 頻率法簡化熵 SP 800-22 Test Illustration · Frequency-Only Entropy Estimate

以 1,024 位元模擬樣本示意空間分佈與偏壓,計算 SP 800-22 中四項檢定的簡化實作(單位元頻率 Monobit、區塊頻率 Block Frequency、遊程 Runs、累計和 Cusum),並依 0/1 頻率估計簡化熵。各預設情境用於比較 AntiFuse、SRAM 啟動態與差動失配的假設分佈,不代表產品量測,也不判定晶片信任根安全性或標準符合性。 Uses a simulated 1,024-bit sample to illustrate spatial distribution and bias with simplified implementations of four SP 800-22 tests (Frequency Monobit, Block Frequency, Runs, Cumulative Sums) and a frequency-only entropy estimate. Presets compare assumed distributions for AntiFuse, SRAM startup states, and differential mismatch. They are not product measurements and do not determine root-of-trust security or standards conformance.

⚠️ 示意模擬模型與技術邊界說明:⚠️ Schematic Model & Technical Boundary Notice: 四項統計檢定與 H∞ = −log₂(max(p, 1−p)) 都是教學示意;頻率法簡化熵未評估完整的相關性、預測性或熵源設計,不等同 SP 800-90B 評估。通過所選檢定不能證明密碼學安全,也不能取得 FIPS 140-3/ISO/IEC 19790 認證。實體 PUF 仍需驗證跨製程、電壓、溫度與老化的穩定性、洩漏及重建機制,並依適用制度取得實作證據。 The four tests and H∞ = −log₂(max(p, 1−p)) are teaching illustrations. The frequency-only estimate does not fully assess correlation, predictability, or entropy-source design and is not an SP 800-90B assessment. Passing the selected tests does not establish cryptographic security or FIPS 140-3 / ISO/IEC 19790 certification. Physical PUFs still require evidence for stability across process, voltage, temperature and aging, leakage, reconstruction, and the applicable evaluation scheme. NIST SP 800-22 · NIST 熵源驗證NIST Entropy Validation
漢明權重 (1-比例)HAMMING WEIGHT
50.02 %
目標 50% ±1%Target 50% ±1%
頻率法簡化熵 H∞FREQUENCY-ONLY ENTROPY
0.985 bit
教學參考值 ≥0.9;非 90B 門檻Illustrative reference ≥0.9; not a 90B threshold
所選檢定通過數SELECTED TESTS PASSED
4 / 4
α = 0.01 門檻α = 0.01 threshold
累計和偏離 ZCUSUM EXCURSION
Z = 12
漫步軌跡極值Random walk max
🛡️ 【教學樣本 · 所選 4 項檢定通過】P-value 均 ≥ 0.01。漢明權重 = 50.02%,頻率法簡化熵估計 = 0.985 bit/bit。結果不能判定熵源安全性、SP 800-90B 符合性或 FIPS 140-3/ISO/IEC 19790 認證。 [Illustrative Sample — 4 Selected Tests Passed] All P-values ≥ 0.01. Hamming weight = 50.02%, frequency-only entropy estimate = 0.985 bit/bit. These results do not establish entropy-source security, SP 800-90B conformance, or FIPS 140-3 / ISO/IEC 19790 certification.

01B / 後量子密碼與抗側信道防禦01B / PQC & SIDE-CHANNEL DEFENSE

硬體根信任 (RoT) 與後量子密碼 PUF 熵品質與差分功率分析 (DPA) 模擬器 Hardware Root of Trust (RoT) & PQC PUF Entropy Quality & DPA Defense Simulator

以差分感測、電流偽裝與時脈抖動的教學假設,示意 Pearson 相關係數 ρmax、金鑰揭示採樣次數 (MTD) 與防護評分。MTD 及高/中/基本/低分級皆為啟發式,不是 FIPS、SESIP、PSA、EVITA 或 ASIL 的要求,也不能推導認證、實測攻擊抵抗力或儲存技術優劣。 Illustrates Pearson correlation ρmax, Measurements to Disclosure (MTD), and protection scores using teaching assumptions for differential sensing, current blinding, and clock jitter. MTD and high/medium/basic/low categories are heuristics, not FIPS, SESIP, PSA, EVITA, or ASIL requirements. They do not establish certification, measured attack resistance, or storage-technology superiority.

電路雜訊擾動標準差 (Circuit Noise Sigma)Circuit Noise Sigma (σ) σ = 4.8
揭示量測次數 MTDTrace MTD
> 10,000,000
峰值相關係數 ρmaxPeak Correlation
0.003
晶片內原生錯誤率Native Intra-BER
0.0000%
安全韌性總評分Security Score
95 / 100

側信道相關性與 PUF 熵分佈視覺化 Side-Channel Correlation & PUF Entropy Visualization

Loading security metrics...

01B / 航太與國防極限輻照物理01B / AEROSPACE & DEFENSE RADIATION HARDNESS

電離輻射清空浮閘電荷;
歐姆物理微絲具天然抗電離屏障
Ionizing radiation drains floating gates;
ohmic microfilaments provide natural immunity

在木星深空強輻射帶(TID > 2.5 Mrad)與地球同步軌道(GEO 150 krad)極端任務中,傳統浮閘與電荷陷阱因光生載子快速中和而失效。0-Mask AntiFuse 透過永久金屬化歐姆微絲,提供天然抗電荷流失之物理特性與巨集極低軟錯誤率。 In extreme deep-space Jovian radiation belts (TID > 2.5 Mrad) and GEO orbits (150 krad), conventional floating gates and charge traps fail as photo-generated carriers neutralize stored electrons. 0-Mask AntiFuse delivers intrinsic resistance to charge loss and low macro soft-error susceptibility via permanent ohmic filaments.

航太與國防極限輻照環境下 NVM 技術之耐受度特性對比Aerospace and Defense Radiation Hardness NVM Technology Comparison
記憶體技術Memory Technology 儲存物理機制Storage Mechanism 總電離劑量耐受 (TID)TID Tolerance 重離子單一事件翻轉 (SEU)Heavy-Ion SEU 單一事件門鎖 (SEL) 門檻SEL LET Threshold 航太太空認證評級Space Qualification
0-Mask AntiFuse OTP 永久歐姆金屬化微絲Permanent Ohmic Filament > 2.5 Mrad(Si) 微絲無電荷 (巨集極低 SEU)Filament immune (Low macro SEU) > 55 MeV·cm²/mg 深空與高軌航太優選Deep Space & Orbit Architecture
STT-eMRAM 磁性穿隧接面 (MTJ)Magnetic Tunnel Junction ~ 500 krad(Si) LETth ≈ 28(需 EDAC 加固)LETth ≈ 28 (Needs EDAC) > 75 MeV·cm²/mg LEO / GEO 軌道耐受LEO / GEO Qualified
ReRAM (OxRAM) 金屬氧化物氧空缺微絲Oxygen-Vacancy Filament ~ 300 krad(Si) LETth ≈ 22(高阻態易擾動)LETth ≈ 22 (HRS Drift) > 60 MeV·cm²/mg 低軌衛星評估中LEO Evaluation
商用浮閘 eFlash 基準Commercial eFlash Baseline 浮閘俘獲靜電荷Trapped Gate Charge 15–25 krad(Si) 極度敏感 (LETth < 6.5)Sensitive (LETth < 6.5) < 35 MeV·cm²/mg 未抗輻加固之商規標準(高輻射環境易早期失效)Standard Terrestrial Grade (Early failure under radiation)
⚠️ 示意模擬模型與技術邊界說明:⚠️ Illustrative Model & Technical Boundary Notice: 本模擬器採用簡化電離輻射劑量 (TID) 閾值漂移與 Weibull 重離子截面積模型,用以示意儲存物理機制(歐姆微絲、磁性穿隧接面、氧空缺與浮閘電荷)在高能粒子撞擊下的巨集可靠度趨勢。真實航太任務須依 MIL-STD-883 方法 1019 或 ESA/ESCC 規範進行實體重離子迴旋加速器與鈷-60 (Co-60) 射源晶圓輻照測試;晶片級 Class-S / QML-V 評級屬於封裝與篩選認證體系,非單一 IP 固有屬性。 This simulator employs simplified TID threshold shift and Weibull heavy-ion cross-section models to illustrate macro-level reliability trends across storage mechanisms (ohmic filaments, MTJ, oxygen vacancies, and trapped charges) under ionizing radiation. Actual flight missions require physical heavy-ion cyclotron and Cobalt-60 source testing per MIL-STD-883 Method 1019 or ESA/ESCC standards. Space ratings such as Class-S / QML-V apply to hermetic package screening systems rather than intrinsic IP properties.
抗輻射架構評估 · 歐姆微絲抗電離模擬器 RAD-HARD ARCHITECTURE EVALUATION · OHMIC FILAMENT SIMULATOR

極限輻照耐受度 (TID/SEE) 與歐姆微絲抗電離動態模擬工作台 Extreme Radiation Hardness (TID/SEE) & Ohmic Filament Simulator

MIL-STD-883 / ESA 規範參考模型 MIL-STD-883 / ESA Reference Model

微觀動態模擬在總電離劑量(0–1,000 krad(Si))與宇宙重離子(LET 0–100 MeV·cm²/mg)衝擊下,介電質電洞捕獲、閾值電壓漂移 ΔVth、位元感測裕度窗退化與軌道 FIT 失效率。數值量化揭露 0-Mask AntiFuse 歐姆微絲在非電荷儲存機制下的極低失效率與深空級可靠度。 Simulates dielectric hole trapping, threshold voltage drift ΔVth, read sense margin window degradation, and orbital FIT rates under Total Ionizing Dose (0–1,000 krad(Si)) and cosmic heavy ions (LET 0–100 MeV·cm²/mg). Quantifies the exceptionally low failure rate and deep-space reliability of 0-Mask AntiFuse ohmic filaments based on non-charge storage physics.

總電離劑量 TID:TID DOSE: 150 krad
重離子 LET:HEAVY-ION LET: 65 MeV
閾值電壓漂移 ΔVthDELTA VTH SHIFT
—
陷阱電洞電荷累積Trapped hole accumulation
剩餘感測電壓裕度REMAINING MARGIN
—
讀取訊號窗 (下限 50mV)Sense window (>50mV)
感測裕度留存比率MARGIN RETENTION
—
相對於照射前初值Relative to pre-rad
SEU 翻轉截面積 σSEU CROSS SECTION
—
重離子單一事件翻轉Heavy-ion upset prob
太空軌道 FIT 率ORBITAL FIT RATE
—
FIT / Mbit 任務失效率FIT / Mbit failures
宇航認證耐受評級RAD-HARD RATING
—
ESA / NASA 級別ESA / NASA Level
Aerospace Radiation Physics Solver
Loading Aerospace Radiation Hardness calculations...

01C / 後量子密碼硬體信任根01C / POST-QUANTUM HARDWARE ROOT-OF-TRUST

格密碼金鑰膨脹 50–150 倍;
分層種子展開守住非揮發性矽面積
Lattice keys inflate by 50–150x;
tiered seed derivation preserves non-volatile silicon area

NIST FIPS 203 (ML-KEM) 與 FIPS 204 (ML-DSA) 後量子密碼標準之私鑰達 1,632–4,896 Bytes。若以晶片內 eNVM 直接存儲完整多項式,面積將造成嚴重成本懲罰。本模擬器展示 0-Mask AntiFuse 存儲 256/512-bit 根種子,透過硬體 PRNG 展開多項式矩陣,結合高階遮罩防禦 DPA 側信道洩漏。 NIST FIPS 203 (ML-KEM) and FIPS 204 (ML-DSA) PQC standards require 1,632–4,896 bytes of private key material. Storing full polynomial matrices in on-chip eNVM imposes prohibitive silicon area penalties. This simulator demonstrates how 0-Mask AntiFuse secures a 256/512-bit root seed with hardware PRNG expansion and higher-order masking to mitigate DPA side-channel leakage.

⚠️ 示意模擬模型與技術邊界說明:⚠️ Schematic Model & Technical Boundary Notice: 本工作台之金鑰儲存預算、矽晶圓面積與 DPA 側信道最小洩漏採樣次數(MTD)為基於 NIST FIPS 203/204 規格與 ISO/IEC 17825 側信道理論之微架構評估模型,非單一實體 ASIC 晶圓實測報告。實際晶圓面積受代工廠 PDK 標準單元庫(CPP×MMP)及周邊電路排版影響;側信道防護等級取決於整體加密模組之動態暫存器防護與板級佈局,不能單獨由非揮發性記憶體介質保證。FIPS 140-3 與 SESIP 評估對象為最終密碼模組 TOE,本技術僅作為硬體安全基礎元件。 The key storage budgets, silicon areas, and DPA MTD figures in this workbench are microarchitectural evaluation models based on NIST FIPS 203/204 specifications and ISO/IEC 17825 theoretical principles, rather than physical ASIC measurement certificates. Actual silicon footprint is determined by foundry PDK standard cell libraries (CPP x MMP) and peripheral layout; side-channel resistance depends on cryptographic accelerator runtime shielding and system layout, and cannot be guaranteed by NVM media alone. Formal FIPS 140-3 and SESIP certifications evaluate complete module TOEs, for which NVM serves as a foundational building block.
PQC 原生私鑰全儲存與分層 AntiFuse 種子展開架構特性對比PQC Raw Key Storage vs Tiered AntiFuse Seed Derivation Architecture Comparison
儲存架構方案Storage Architecture 非揮發性記憶體預算NVM Budget 矽晶圓面積縮減Area Reduction 晶圓測試時間 (CP) 估算Wafer Test Time (CP Est.) DPA 側信道 MTD 估算DPA Defense MTD (Est.) 支援之模組安全評估基礎Module Security Baseline
0-Mask AntiFuse 根種子衍生0-Mask AntiFuse Seed Derivation 32–64 Bytes (256/512-bit) 50x–126x (98%–99%) < 15 ms / die (32B 燒錄32B prog) > 2.5e7 traces (高階遮罩估算Higher-order masking est.) 支援高抗 DPA 架構設計(可協助達成 FIPS 140-3 / SESIP 評估)Supports high-DPA resistance architectures (assists FIPS 140-3 / SESIP evaluations)
原生多項式矩陣全容量儲存Raw Polynomial Direct Storage 1,632–4,896 Bytes 1.0x (基準,面積龐大Baseline, high area) > 850 ms / die (全密鑰燒錄Full key prog) < 1.0e4 traces (讀出軌跡暴露Exposed trace) 常規防護(需外部電路補強以抗 DPA)Standard baseline (requires external logic hardening against DPA)
⚠️ 示意模擬模型與技術邊界說明:⚠️ Illustrative Model & Technical Boundary Notice: 本模擬器採用簡化矽面積估算與 DPA 側信道洩漏指數模型,用以示意 NIST FIPS 203/204 後量子密碼演算法在原生私鑰全儲存與分層根種子展開架構下的面積與測試時間折衷。文中推導之 MTD(最小洩漏採樣次數)為理論簡化估算,實際側信道防禦強度需經 ISO/IEC 17825 實驗室 TVLA 實測檢驗;FIPS 140-3 與 SESIP 認證屬系統級晶片 TOE 產品評估範疇,不可由單一儲存介質直接推定。 This simulator utilizes simplified silicon area models and exponential DPA leakage estimations to illustrate trade-offs between full raw key storage and tiered root seed derivation for NIST FIPS 203/204 PQC algorithms. The derived MTD figures are theoretical estimates; actual side-channel defense efficacy requires empirical TVLA testing per ISO/IEC 17825 standards. FIPS 140-3 and SESIP certifications apply to comprehensive system-level TOE products and cannot be inferred from the storage medium alone.
PQC 密碼工程架構評估 · 信任根金鑰儲存試算器 PQC CRYPTO ARCHITECTURE EVALUATION · HARDWARE ROT SIMULATOR

後量子密碼 (ML-KEM / ML-DSA) 金鑰儲存預算與側信道防禦試算器 Post-Quantum (ML-KEM/ML-DSA) Key Storage & Side-Channel Simulator

NIST FIPS 203 / 204 標準 NIST FIPS 203 / 204 Standards

基於 NIST FIPS 203/204 規格評估 ML-KEM-512/768/1024 與 ML-DSA-65/87 在全私鑰存儲與 0-Mask AntiFuse 根種子(32/64 Bytes)衍生架構下的矽晶圓面積開銷、測試時間、以及結合布林/算術遮罩時 DPA 側信道最小洩漏採樣次數(MTD 估算值)。 Evaluates silicon area overhead, CP test times, and modeled DPA Minimum Traces to Disclose (MTD) between raw key storage and 0-Mask AntiFuse root seed (32/64 Bytes) expansion across ML-KEM-512/768/1024 and ML-DSA-65/87 algorithms based on NIST FIPS 203/204 specifications.

原生私鑰容量RAW KEY SIZE
—
完整多項式向量Full poly vector
實體種子儲存預算SEED BUDGET
—
AntiFuse 根金鑰AntiFuse root key
面積縮減與節省比AREA REDUCTION
—
矽面積開銷節約Silicon area saved
DPA 破解採樣門檻DPA MTD TRACES
—
側信道抗攻擊防護Side-channel metric
FIPS 信任根評級FIPS ROT RATING
—
硬體安全保證Hardware assurance
Post-Quantum Hardware Security Solver
Loading Post-Quantum Storage calculations...
後量子硬體信任根 · 雙錨點架構評估PQC HARDWARE ROOT OF TRUST · DUAL-ANCHOR EVALUATION

SRAM PUF/AntiFuse 混合信任根與輔助資料防護SRAM PUF / AntiFuse Hybrid Root of Trust & Helper Data Protection

教學示意 · 非商用晶片驗證Teaching Model · No Commercial Silicon Validation

本段提出可供評估的混合架構:SRAM PUF 重建易失根值,AntiFuse OTP 保存輔助資料的可信摘要或驗證公鑰;若採 HMAC,須另保護其秘密金鑰,HMAC 標記也不是一般雜湊。是否另存 PQC 秘密種子,取決於金鑰生命週期需求;若種子永久儲存,便不能宣稱所有根秘密皆在斷電後不留存。「0-Mask」須有特定製程/IP 證據,並非 AntiFuse 的通用屬性。以下標準與論文支持演算法及威脅分析,不證明此組合已商用、通過認證或具有資訊理論層級的整體硬體安全。 This teaching architecture combines a volatile root reconstructed by SRAM PUF with AntiFuse OTP holding a trusted helper-data digest or verification public key. HMAC additionally requires a protected secret key; its tag is not an ordinary hash. Persisting PQC secret seeds is a lifecycle choice and precludes claiming that all root secrets are absent at rest. “0-Mask” requires process/IP-specific evidence. The cited standards and papers support algorithm and threat analysis, not commercial deployment, certification or information-theoretic assurance of the complete hardware.

01 · 區分熵損失與主動竄改01 · Entropy Loss & Active Tampering

對 [n, k] 二元線性碼,症候可由 n−k 個獨立校驗位元表示;secure sketch 的熵損失須依建構與來源模型估算,不能把所有 BCH 都寫成每糾正 t 個錯誤即洩漏 7t 位元。公開輔助資料的剩餘熵分析與主動竄改防護是兩個問題。示意設計以可信摘要/簽章驗證、裝置綁定及版本政策拒絕未授權資料;也可評估穩健模糊擷取器,仍須驗證失敗處理與實作側信道。 For a binary [n, k] linear code, a syndrome can be represented by n−k independent parity bits. Secure-sketch entropy loss depends on the construction and source model; 7t bits is not a universal BCH rule. Residual-entropy analysis and active-tamper protection are separate. The model uses trusted digest/signature verification, device binding and version policy, or can evaluate a robust fuzzy extractor; failure handling and implementation side channels still require validation.

線性碼與熵界線原始論文Linear Codes & Entropy Bounds · 穩健擷取器Robust Extractors

02 · PQC 種子與實際儲存預算02 · PQC Seeds & Storage Budget

FIPS 204 的 ML-DSA.KeyGen 使用 32 位元組 ξ;ML-DSA-87 編碼私鑰為 4,896 位元組。FIPS 203 的 ML-KEM.KeyGen 使用 d、z 各 32 位元組,完整重建種子合計 64 位元組。重建須依各標準 KeyGen 與指定雜湊/XOF 流程,不能用任意 AES-CTR 擴張替代。種子須受私鑰等級保護並符合亂數要求;PUF 根值不會自動滿足 FIPS 金鑰生成要求。儲存位元組減少也不等於矽面積等比例縮小,仍須計入輔助資料、工作記憶體、加速器、延遲及輪替。 FIPS 204 ML-DSA.KeyGen uses a 32-byte ξ; the ML-DSA-87 encoded private key is 4,896 bytes. FIPS 203 ML-KEM.KeyGen uses 32-byte d and z, totaling 64 bytes for regeneration. Follow each standard’s KeyGen and specified hash/XOF operations, not arbitrary AES-CTR expansion. Seeds need private-key protection and compliant randomness; a PUF root is not automatically a compliant KeyGen input. Byte savings do not imply proportional silicon-area savings: helper data, working memory, accelerators, latency and rotation also matter.

NIST FIPS 203 §3.3 / §7.1 · NIST FIPS 204 §5.1 / Table 2

03 · 斷電殘留與環境可靠度03 · Remanence & Environmental Reliability

SRAM 斷電後的資料衰減受時間、溫度與電路影響,不能宣稱立即無殘留;應驗證重建根值與工作記憶體清除。PUF 重建錯誤率與 AntiFuse 讀取裕量均須使用目標製程的環境及老化數據。260°C 回焊耐受與零漂移未有本架構的實測佐證,不能當成通用特性;兩種元件並用也不會自動抵抗探針、DPA 或故障注入。此處要求對應證據,不提供未經量測的可靠度數值。 SRAM decay after power loss depends on time, temperature and circuitry; immediate absence of remanence is not assured. Validate root and working-memory zeroization. PUF reconstruction errors and AntiFuse read margins require process-specific environmental and aging evidence. This model has no measured support for 260°C reflow tolerance or zero drift. Combining the components does not automatically resist probing, DPA or fault injection; no unmeasured reliability figures are assigned here.

USENIX TARDIS:SRAM 殘留實驗USENIX TARDIS: SRAM Remanence Experiments

05 / HARDWARE TROJAN & OPTICAL PEM

Hardware Trojan, Optical PEM & PUF Verification Simulator

Physical & statistical simulation: Hot-carrier luminescence, backside silicon transmission, SNSPD optical SNR, trojan detection probability, and 0-Mask PUF Hamming distribution.

⚠️ 示意模擬模型與技術邊界說明:⚠️ Illustrative Model & Technical Boundary Notice: 本模擬器所採用之熱載子發光 (PEM) 光子計數、矽基板近紅外穿透率、SNSPD 光學信噪比與木馬檢出率為簡化物理與統計推論模型(參考 Becker 2013 與 Schlösser 2012 相關文獻),非實驗室特定破壞性逆向工程機台(如 PHEMOS-1000 / Thermo Fisher Meridian)之實測保證數據。光學側信道發射強度受操作電壓、金屬層覆蓋幾何及晶圓背面研磨厚度影響;防禦能力應透過實體晶片光學顯微檢驗與全生命週期供應鏈管控多層驗證。 The hot-carrier luminescence (PEM) photon emission counts, backside NIR transmission, SNSPD optical SNR, and trojan detection probabilities in this simulator are simplified physical and statistical models (referencing Becker 2013 & Schlösser 2012), not empirical test guarantees from specific failure analysis instruments (such as PHEMOS-1000 / Meridian). Physical emission signatures depend on operating voltage, metal layer stack geometry, and backside thinning depth; defense posture must be verified through physical chip optical analysis and multi-tier supply chain chain-of-custody controls.
*註:數值為文獻典型值與工程模型估算假設(參考 Becker 2013 與 Schlösser 2012),非產品保證。認證等級屬完整 TOE 評估範疇。 *Note: Values represent academic typical figures and engineering assumptions (per Becker 2013 & Schlösser 2012). Certification tiers apply to complete TOE evaluations.
信任根儲存架構Root-of-Trust Architecture 光罩增加道數Mask Adders 估算光子通量Estimated Photon Flux 屏蔽衰減估算Optical Shielding 木馬防護邊界Trojan Defense Boundary 晶片間漢明距離 (Inter-HD)Inter-Chip HD 架構防禦角色Architectural Role
0-Mask AntiFuse 原生 PUF0-Mask AntiFuse Native PUF 0 道 (標準邏輯製程)0 Masks (Standard CMOS) < 0.1 ph/s (微安級微絲)< 0.1 ph/s (Microamp filament) ~ 48 dB (高衰減模型)~ 48 dB (High attenuation) 消除客製層光罩插入面;仍需防範標準層摻雜木馬Eliminates dedicated mask vector; requires dopant defenses 50.02% (典型均勻)50.02% (Uniform) 高抗測餘量 (需結合 TOE)High Margin (Requires TOE)
STT-MRAM (垂直 MTJ)STT-MRAM (Perpendicular MTJ) 3 ~ 5 道 (BEOL 金屬層)3 ~ 5 Masks (BEOL) ~ 85 ph/s~ 85 ph/s ~ 28 dB~ 28 dB 後段磁穿隧結隱蔽性高;需防範外部磁場與 MOKEBEOL MTJ concealment; requires MOKE / magnetic shielding 49.80%49.80% 中等抗測 (需磁屏蔽)Moderate (Needs mag shield)
SRAM 啟動狀態 PUFSRAM Power-Up PUF 0 道 (標準邏輯製程)0 Masks (Standard CMOS) ~ 150 ph/s (翻轉瞬間)~ 150 ph/s (Switching) ~ 18 dB~ 18 dB 免額外光罩;需防範供電噪訊與雷射 BBI 誘騙No mask adders; requires anti-glitch & laser BBI filtering 49.30% (Intra-BER ~ 4.8%)49.30% (Intra-BER ~ 4.8%) 商用基線 (需強韌 ECC)Commercial Baseline (Needs ECC)
傳統 eFlash 儲存金鑰Conventional eFlash Key 8 ~ 12 道 (高壓厚氧層)8 ~ 12 Masks (HV Oxide) > 1000 ph/s (高壓發射)> 1000 ph/s (Pump Emission) < 10 dB< 10 dB 具專屬高壓光罩攻擊面;高壓電荷泵需發射遮蔽Dedicated HV mask vector; requires pump shielding 無 (非 PUF 靜態寫入)N/A (Non-PUF Static) 需增強主動屏蔽網Requires Active Shielding
零信任供應鏈與光學安全威脅建模結論:Zero-Trust & Optical Security Modeling Finding: 在半導體供應鏈安全中,代工廠惡意篡改與背面光學側信道 (PEM/TRE) 構成關鍵攻擊面。0-Mask AntiFuse PUF 藉由「零額外光罩」消除專屬非揮發層的光罩插入途徑;微安級讀取電流亦在物理上降低熱載子發光強度,大幅拉高時間分辨光子發射 (TRE) 的捕捉難度。但依據 Becker 2013 與 Schlösser 2012 文獻,任何單一技術皆無法完全免疫標準邏輯層摻雜木馬或極端背面薄化下的探測,仍需配合全生命週期防禦鏈與系統級 TOE 驗證。詳細文獻與證據總帳請參見 Evidence V22 ↗。 In semiconductor supply-chain security, untrusted foundry manipulation and backside optical side-channels (PEM/TRE) represent significant threat surfaces. 0-Mask AntiFuse PUF eliminates dedicated mask-insertion vectors targeting non-volatile layers; microampere read currents physically suppress hot-carrier luminescence, raising acquisition thresholds for time-resolved emission (TRE) attacks. However, as demonstrated by Becker 2013 and Schlösser 2012, no single medium immunizes against standard-cell dopant Trojans or aggressive backside thinning; full lifecycle governance and system-level TOE evaluations remain essential. Refer to Evidence V22 ↗ for the complete research ledger.

02 / ATTACK-WINDOW LIFECYCLE

The root has no permanent address; risk still has windows

Select a lifecycle phase to inspect assets, attack surface, test question and expected outcome.

以下互動呈現假設架構的敏感時窗與預期控制目標,不是實際執行安全控制或產品測試結果。This interaction describes sensitive windows and intended controls in an assumed architecture; it does not execute security controls or report product test results.

03 / COUNTERMEASURE PATTERNS

A defense is not a feature list; it is a validated causal chain

Review the full control chain item by item.

SECURITY BLOCKCLAIMTESTEXPECTED EVIDENCELIMIT
SRAM PUFRoot is not storedSCA · FI · PVT · agingReconstruction, leakage, failure modeSensitive intermediates exist at runtime
AES / KDFOTP readout does not equal plaintextDPA/CPA · DFA · protocol abuseAttack cost and residual riskImplementation and integration determine outcome
ANTIFUSE OTPRetains ciphertext and statereadout · microscopy · mappingObservable content and exploitabilityDo not assume cells are unreadable
PQC / HYBRID ROTHybrid classical & quantum-resistant bindingML-KEM · ML-DSA · SCA · Fault injectionConstant-time execution, higher-order DPA maskingEscalated key and ciphertext sizes increase NVM footprint
ACTIVE SHIELDDetects physical delayering and microprobingFIB breach · line open · impedance/capacitance driftDynamic PRBS excitation, single-cycle zeroizationPackaging stress and environmental shifts may induce false triggers
CONTROLLEROwns access and lifecycle policyglitch · reset · sequence abuseFail-secure, atomicity, error handlingSoC integration can break IP assumptions

04 / ASSURANCE MATURITY · M1–M6

Assurance is not binary; maturity has levels

M1–M6 describes assurance maturity; it is distinct from the Evidence Ledger's E1–E4 source classes. Select a level to see acceptable outputs and claims that remain out of scope.

06 / ROLE-BASED LEARNING

From understanding attacks to reviewing an assurance plan

07 / KNOWLEDGE GOVERNANCE

Preserve traceable knowledge for SharePoint and Copilot

CLASSIFY

Public · Internal · NDA

SCOPE

IP · Subsystem · SoC · Device

TRACE

Claim · Source · Review date · Owner

QUALIFY

Applicability · Limitation · Unknown

The content model maps to SharePoint List fields so Copilot can return sources, applicability and unknowns with each answer.

PUBLIC SOURCE DISCIPLINE

Learn from professional methods without implying endorsement

Riscure and Keysight are cited as public sources only. This site does not claim partnership, tool use or certification. Actionable exploit parameters remain outside the public site.

後量子密碼晶片防禦模型POST-QUANTUM CRYPTO DEFENSE MODEL

PQC 金鑰儲存 1064nm 抗雷射注入 (LFI) 物理防禦模擬器PQC Key Storage 1064nm Laser Fault Injection (LFI) Defense Simulator

⚠️ 示意模擬模型與技術邊界說明:本模型基於矽基底 1064nm 近紅外光電效應與雷射脈衝注入閥值計算,模擬主動金屬屏蔽層之光電信號抑制與熔斷清零響應,供晶片安全架構防護設計參考。 ⚠️ Illustrative Simulation Model & Technical Boundary Note: Grounded in silicon 1064nm photoelectric carrier generation equations; models active mesh attenuation and tamper defense.

在後量子密碼學 (PQC) 晶片中,儲存私鑰的 NVM 容易遭受 1064nm 近紅外超短脈衝雷射攻擊(LFI)。雷射穿透矽基板引發微觀光電突波電流,可能導致位元翻轉洩漏金鑰。本工具評估主動頂層金屬網格屏蔽、差動感測與即時安全熔斷機制對抗 LFI 攻擊之防禦裕度。 Post-Quantum Cryptography (PQC) private keys are vulnerable to 1064nm near-IR Laser Fault Injection (LFI). Focused pulses create photocurrent transients that induce bit flips. This tool evaluates active top-metal mesh shielding and rapid tamper zeroization defenses.

透射有效能量Effective Energy
0.0003 nJ
金鑰翻轉竄改機率Upset Probability
0.0%
安全防禦裕度Security Margin
+36.2 dB
主動安全清零延遲Zeroization Delay
4.5 ns
載入中...Loading...