Pith. sign in

REVIEW 3 major objections 78 references

Accelerating Confidential Databases with Crypto-free Mappings

T0 review · 3 major / 0 minor · reviewed 2026-07-13 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Confidential databases can replace per-field encryption on the query path with simple trusted-domain index lookups and still keep secrets out of the cloud.

desk verdict Solid systems fix for a real CDB bottleneck: crypto-free FID mappings beat HEDB hard on the measured path, with production GaussDB integration and only presentation/side-channel caveats. read the letter →

arxiv 2603.18836 v3 pith:FW3J5TUM submitted 2026-03-19 cs.CR cs.DB

Wenxuan Huang , Zhanbo Wang , Mingyu Li This is my paper · ORCID
classification cs.CRcs.DB
keywords confidentialdatabasestrustedexecutionenvironmentscrypto-freemappingsfieldidentifierssplitarchitectureTPC-HTPC-Ccross-domainconsistency
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Modern confidential databases keep most of the engine outside the trusted hardware and only put expression operators inside it. That split forces a decrypt-compute-encrypt round trip on every cross-domain call, plus ciphertext expansion that inflates I/O and memory. This paper argues the real work of those ciphertexts is only indirection—pointing from the untrusted database into plaintext held inside the trusted domain—and that indirection need not be cryptographic. FEDB therefore stores data-independent field identifiers in the database and resolves them with ordinary put/get operations against a mapping store of plaintext secrets that lives only inside the trusted privacy zone. Encryption is deferred to asynchronous block-level eviction for data at rest. The result, measured against a state-of-the-art baseline on TPC-C and TPC-H, is large reductions in query latency and storage while still supporting standard SQL, ACID, DBA maintainability, and the paper’s confidentiality and integrity claims. Techniques from the design have been integrated into a production cloud database.

What carries the argument

Crypto-free mappings: a trusted-domain mapping store that binds data-independent field identifiers (FIDs) to plaintext secrets, so the untrusted DBMS manipulates only opaque FIDs and the privacy zone replaces decrypt/encrypt with constant-time Get/Put, with temporary versus permanent partitions, locality-aligned layout, and a two-phase commit protocol that enforces external synchrony.

What would settle it

Re-run the same TPC-H suite and microbenchmarks on an equivalent dual-TEE split, replace the mapping-store Get/Put path with the baseline’s field-level AES-GCM path while holding batch size, memory, and locality fixed, and check whether the reported speedups (up to ~78× on TPC-H, ~1.8× on TPC-C) and storage reductions disappear or reverse.

Watch

Extended reading notes

Core claim

The performance crisis of modern split-architecture confidential databases is not inherent to their security goals. By decoupling indirection from protection—using lightweight, data-independent field identifiers for cross-domain references and reserving encryption for data at rest—crypto-free mappings remove synchronous en/decryption from the critical path and cut ciphertext expansion, yielding up to 78× lower TPC-H latency and up to 1.8× higher TPC-C throughput versus HEDB while preserving functionality, maintainability, and the stated threat-model guarantees.

Load-bearing premise

Keeping plaintext secrets resident for long periods inside the trusted privacy zone does not open a meaningfully worse digital side-channel surface than the long-lived encryption keys modern confidential databases already hold, provided page-table isolation and existing key-isolation practices are in place.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 0 minor

Summary. The manuscript presents FEDB (named ZENO in the front-matter abstract), a split-architecture confidential database that replaces ciphertext-based cross-domain references with crypto-free field identifiers (FIDs). FIDs are stored in the untrusted DBMS and resolved via a TEE-resident mapping store (Put/Get) to plaintext secrets, so en/decryption leave the query critical path; encryption is deferred to asynchronous block-level protection at rest. The design inherits a dual-zone (integrity/privacy) architecture in the style of HEDB, adds temporary vs. permanent mapping partitions, locality-aligned partitioning and prefetch, and a prepare/FlushLog/commit protocol that enforces external synchrony for cross-domain ACID. Evaluation on ARM Secure-EL2 against plaintext PostgreSQL and HEDB reports up to ~1.8× TPC-C throughput, up to 78.0× lower TPC-H latency (body), large storage savings, microbenchmark sensitivity results, recovery timing, and a security argument (active integrity, FID plaintext-independence, L-security leakage profile, side-channel discussion).

Significance. If the results hold under the stated threat model, this is a practically important systems contribution: modern maintainable CDBs are already shipping, and synchronous field-level crypto plus ciphertext expansion are the dominant adoption barriers the paper profiles. The core insight—decoupling indirection from protection, in the spirit of capabilities/handles—is clean and transferable. Strengths include a profiled motivation against HEDB, a concrete API that existing operator proxies can adopt with small changes, end-to-end TPC-C/TPC-H plus storage/recovery/ablation evidence on ARM, an explicit external-synchrony correctness criterion, and a claimed production path (open-source integration into GaussDB). These are the right artifacts for a systems security venue.

major comments (3)
  1. Front-matter abstract vs. body are inconsistent on load-bearing result claims. The abstract names the system ZENO and reports TPC-H speedups of up to 53.1× (ARM S-EL2) and 94.7× (x86 TDX) vs. HEDB, plus a “real-world industrial workload.” The body names FEDB, reports up to 78.0× TPC-H and 1.8× TPC-C on ARM Secure-EL2 (§7.1, Fig. 6, Fig. 5), and does not present end-to-end TPC-H/TPC-C or industrial-workload numbers on TDX—only micro-operation cycle counts in Table 4. Before acceptance, the abstract, title branding, and §7 must describe the same system and only claim platforms/workloads that are actually evaluated (or the missing experiments must be added).
  2. §8 “Side-channel considerations” is the main security residual behind the performance win: the privacy zone now holds long-lived plaintext fields, not only keys. The argument that this is comparable to permanent key residency in modern CDBs, and that ARM Secure-EL2 page-table isolation (or equivalent) plus existing key-isolation mitigations suffice, is plausible but largely qualitative. For the central claim that security is preserved rather than traded, the paper should more sharply state what is and is not in scope (e.g., which digital side channels are assumed mitigated by the TEE configuration used in §7), and avoid implying multi-TEE generality beyond what isolation properties are actually required.
  3. §5’s external-synchrony commit protocol (prepare → privacy-zone FlushLog as commit #1 → DBMS WAL as commit #2) is the correctness linchpin for durable FID→secret mappings. The abort path (logical delete + offline physical reclaim) and crash/GC discussion are directionally right, but the paper should state more precisely the failure model for partial FlushLog, concurrent multi-connection commits, and how orphan GC is driven without violating the invariant that every user-visible FID still maps to a valid secret. A short correctness argument or invariant list tied to Figure 4 would make this load-bearing mechanism reviewable rather than narrative-only.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: empirical systems design and measured speedups against external baselines (plaintext PostgreSQL, HEDB), not definitional or fitted predictions.

full rationale

FEDB/ZENO is a systems paper whose load-bearing claims are architectural (crypto-free FID mappings that decouple indirection from protection, dual-zone commit/WAL for external synchrony) and empirical (TPC-C/TPC-H/microbenchmark throughput and latency vs. plaintext PostgreSQL and HEDB on ARM Secure-EL2, plus storage and recovery measurements). The FID indistinguishability argument is a standard plaintext-independent allocation argument, not a tautology that redefines the target. Reuse of HEDB operators and dual-zone structure is ordinary prior-work inheritance with overlapping authorship; the new mapping store, partitioning, prefetch, and consistency protocol are independently specified and evaluated. There are no fitted parameters renamed as predictions, no uniqueness theorems imported from the authors, no ansatz smuggled via self-citation, and no renaming of a known empirical pattern as a first-principles derivation. Naming/number mismatches between abstract and body are presentation issues, not circular reductions. Score 0 with empty steps is the correct outcome.

Assumptions & free parameters 3 free parameters · 5 assumptions · 3 invented entities

The central performance and security claims rest on standard TEE guarantees, an accepted modern-CDB leakage/maintainability tradeoff, and several design mechanisms the authors introduce (FIDs, partitioned mapping store, external-synchrony commit). Configuration knobs (batch size, FID bit layout, partition policy) affect absolute numbers but are not fitted parameters that define the claim. No physical constants or curve-fitting parameters underwrite the result.

free parameters (3)
  • RPC batch size = 256 fields
    Evaluation uses batch size 256 for aggregation-style operators; absolute speedups depend on this engineering choice though the crypto-free advantage remains with batching off.
  • FID bit layout (partition prefix + offset) = 16 + 48 bits (default)
    Default 16-bit partition prefix and 48-bit offset are configurable; they trade addressable partitions against storage/lookup efficiency and are chosen by design, not fitted to hide a failed claim.
  • Mapping-store partition policy = DB-specific alignment
    Per-heap-file (PostgreSQL) or per-page (InnoDB-style) partitioning is a hand-chosen locality policy that improves measured throughput (+7.0% in ablation) but is not a statistical fit.
assumptions (5)
  • domain assumption TEE hardware provides memory isolation of privacy-zone state and remote attestation of the loaded code before secrets are provisioned.
    Stated in threat model and §8; without it, plaintext mapping-store contents are not confidential.
  • domain assumption Adversaries may control all software outside TEEs and observe I/O/network patterns, but TEE defects, physical DRAM probing, DoS, and DBMS logic bugs are out of scope.
    Explicit threat-model exclusions in §3; security claims are conditional on this boundary.
  • ad hoc to paper External synchrony is a sufficient correctness criterion: user-visible FID state must always map to valid secrets; orphan secrets in the mapping store are allowed.
    Adopted in §5 from Nightingale et al.; underwrites the abort/commit/recovery design rather than full internal cross-domain linearizability at all times.
  • standard math FID allocation is a stateful counter independent of plaintext, so the adversary’s FID view is identical for any two equal-length insertion sequences of secrets.
    Core of the passive-adversary indistinguishability argument in §8; standard for non-message-dependent handles.
  • domain assumption HEDB’s dual-zone, operator-in-TEE architecture is representative of industry modern CDBs for both performance bottlenecks and security goals.
    Used throughout motivation and evaluation; gains are primarily reported relative to HEDB.
invented entities (3)
  • Crypto-free mapping / FID independent evidence
    purpose: Replace ciphertext pointers with data-independent identifiers so cross-domain RPCs use put/get instead of en/decrypt.
    Central mechanism; implemented as 64-bit IDs into a linear/slab mapping store. Independent evidence via open implementation and benchmarks, not only narrative.
  • Partitioned mapping store (temporary vs permanent) independent evidence
    purpose: Manage ephemeral intermediate results vs durable fields with different reclamation and persistence rules without dangling FIDs.
    Introduced in §4.1 to solve lifetime inference; evaluated in technique sensitivity (+2.4% temporary, +7.0% permanent).
  • Dual-zone external-synchrony commit protocol (prepare / commit#1 FlushLog / commit#2 DBMS WAL) independent evidence
    purpose: Ensure secrets are durable before FIDs become user-visible and support crash recovery without breaking the FID→secret invariant.
    New cross-domain protocol in §5; recovery cost measured (2.1 s mapping WAL vs 13.2 s DBMS).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Accelerating Confidential Databases with Crypto-free Mappings." pith.science (2026). https://pith.science/paper/FW3J5TUM

@misc{pith2026260318836,
  author       = {Pith},
  title        = {Pith review of: Accelerating Confidential Databases with Crypto-free Mappings},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FW3J5TUM}},
  note         = {Machine review of arXiv:2603.18836}
}
read the original abstract

Confidential databases (CDBs) enable secure queries over sensitive data in untrusted cloud environments using confidential computing hardware. While adoption is growing, widespread deployment is hindered by high overheads from frequent synchronous cryptographic operations, which cause significant computational and I/O bottlenecks. ZENO is a novel CDB design that removes cryptographic operations from the critical path. It introduces crypto-free mappings that maintain data-independent identifiers within the database while securely mapping them to plaintext secrets in a trusted domain. This paradigm shift yields substantial performance gains across industry-standard benchmarks (TPC-C, TPC-H) and a real-world industrial workload. Specifically, ZENO speeds up TPC-H queries by up to 53.1x on ARM S-EL2 and 94.7x on x86 TDX compared to HEDB. ZENO's optimization techniques have been integrated into GaussDB.

Figures

Figures reproduced from arXiv: 2603.18836 by the authors.

Figure 1
Figure 1. In modern CDBs, a cross-domain invocation involves two decryptions and one encryption, incurring high CPU overhead. This gap arises from a deliberate system design choice made by industry deployments: rather than enclosing the entire DBMS within a TEE, which would prevent database adminis￾trators from performing essential maintenance tasks, modern CDBs execute only critical operators (e.g., numeric SUM oper￾ations) … view at source ↗
Figure 2
Figure 2. (a), the en/decryption operations, along with I/O amplification overhead, contribute considerably to the overall execution time, accounting for 10.5-62.6% and 13.9-25.5% across different workloads. Combined, they cause an overall latency slowdown of up to 5.3× compared to plaintext. Fig￾ure 2 (b) also shows that the field-level encryption incurs 1.4× to 3× storage overhead compared to plaintext. The remaining 8.6-26… view at source ↗
Figure 3
Figure 3. High-level architecture of FEDB. data remains visible only within TEE domains (§ 8). Insights. Current practice of modern CDBs employs a mapping-based protection scheme, using ciphertexts as point￾ers to plaintext data in TEEs. This design conflates two or￾thogonal concerns: indirection (cross-domain references) and protection (data confidentiality). Because ciphertexts serve both roles, this coupling necessitates e… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Timeline of a transaction execution in FEDB, illustrating the handling of (a) abort and (b) commit paths. quires only that the external state remains consistent—internal states may temporarily diverge as long as they never violate external invariants. In FEDB, the mapp…
Figure 5
Figure 5. Figure 5: TPC-C throughput under different client connections with 128 warehouses. FEDB achieves up to 1.8× speedup over HEDB [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Execution time of TPC-H queries of HEDB and FEDB (with and without batching), normalized to the execution time of plaintext PostgreSQL (shown in parentheses). FEDB achieves speedups ranging from 1.0× to 78.0× compared to HEDB. ro rw wo 0.0 0.5 1.0 Normalized TPS point …
Figure 7
Figure 7. Figure 7: Microbenchmark throughput under different configurations. FEDB consistently achieves higher throughput than HEDB. of the integrity and privacy zones) to match different sizes of the dataset. Each experiment runs a 120-second warm￾up followed by a 600-second measurement…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

78 extracted references · 2 linked inside Pith

  1. [1]

    https://aws.amazon.com/rds/ aurora/, 2025

    Amazon Aurora. https://aws.amazon.com/rds/ aurora/, 2025

    2025

  2. [2]

    https://azure.microsoft

    Azure SQL Database. https://azure.microsoft. com/en-us/products/azure-sql/database/, 2025

    2025

  3. [3]

    https://cloud.google.com/ sql, 2025

    Google Cloud SQL. https://cloud.google.com/ sql, 2025

    2025

  4. [4]

    https://preyproject.com/blog/ worst-data-security-breaches, 2025

    Recent data breaches 2025: latest cybersecurity inci- dents & lessons. https://preyproject.com/blog/ worst-data-security-breaches, 2025

    2025

  5. [5]

    https://www.hhs.gov/ hipaa/for-professionals/privacy/index.html, 2024

    The HIPAA Privacy Rule. https://www.hhs.gov/ hipaa/for-professionals/privacy/index.html, 2024

    2024

  6. [6]

    https://www.pcisecuritystandards.org/ standards/, 2024

    Payment Card Data Security Standard (PCI-DSS). https://www.pcisecuritystandards.org/ standards/, 2024

    2024

  7. [7]

    Trusteddb: a trusted hard- ware based database with privacy and data confidential- ity

    Sumeet Bajaj and Radu Sion. Trusteddb: a trusted hard- ware based database with privacy and data confidential- ity. InProceedings of the ACM SIGMOD Conference, 2011

    2011

  8. [8]

    Orthogonal security with cipherbase

    Arvind Arasu, Spyros Blanas, Ken Eguro, Raghav Kaushik, Donald Kossmann, Ravishankar Ramamurthy, and Ramarathnam Venkatesan. Orthogonal security with cipherbase. InProceedings of the Conference on Innovative Data Systems Research (CIDR), 2013

    2013

Show all 78 references
  1. [9]

    En- clavedb: A secure database using SGX

    Christian Priebe, Kapil Vaswani, and Manuel Costa. En- clavedb: A secure database using SGX. InProceedings of the IEEE Symposium on Security and Privacy (S&P), 2018

    2018

  2. [10]

    Stealthdb: a scalable encrypted database with full SQL query support.Proceedings of the Privacy Enhancing Technologies Symposium (PETS), 2019

    Dhinakaran Vinayagamurthy, Alexey Gribov, and Sergey Gorbunov. Stealthdb: a scalable encrypted database with full SQL query support.Proceedings of the Privacy Enhancing Technologies Symposium (PETS), 2019

    2019

  3. [11]

    A, and Florian Ker- schbaum

    Benny Fuhry, Jayanth Jain H. A, and Florian Ker- schbaum. Encdbdb: Searchable encrypted, fast, com- pressed, in-memory database using enclaves. InAnnual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN), 2021

    2021

  4. [12]

    Panagiotis Antonopoulos, Arvind Arasu, Kunal D. Singh, Ken Eguro, Nitish Gupta, Rajat Jain, Raghav Kaushik, Hanuma Kodavalla, Donald Kossmann, Niko- las Ogg, Ravi Ramamurthy, Jakub Szymaszek, Jeffrey Trimmer, Kapil Vaswani, Ramarathnam Venkatesan, and Mike Zwilling. Azure SQL ...

    2020

  5. [13]

    Operon: An encrypted database for ownership- preserving data management

    Sheng Wang, Yiran Li, Huorong Li, Feifei Li, Chengjin Tian, Le Su, Yanshan Zhang, Yubing Ma, Lie Yan, Yuanyuan Sun, Xuntao Cheng, Xiaolong Xie, and Yu Zou. Operon: An encrypted database for ownership- preserving data management. InProceedings of the International Conference on...

    2022

  6. [14]

    Full encryption: An end to end encryption mechanism in gaussdb

    Liang Guo, Jinwei Zhu, Jiayang Liu, and Kun Cheng. Full encryption: An end to end encryption mechanism in gaussdb. InProceedings of the International Conference on Very Large Data Bases (VLDB), 2021

    2021

  7. [15]

    Encrypted databases made secure yet maintainable

    Mingyu Li, Xuyang Zhao, Le Chen, Cheng Tan, Huorong Li, Sheng Wang, Zeyu Mi, Yubin Xia, Feifei Li, and Haibo Chen. Encrypted databases made secure yet maintainable. InProceedings of the USENIX Sympo- sium on Operating Systems Design and Implementation (OSDI), 2023

    2023

  8. [16]

    Secudb: An in- enclave privacy-preserving and tamper-resistant rela- tional database

    Xinying Yang, Cong Yue, Wenhui Zhang, Yang Liu, Beng Chin Ooi, and Jianjun Chen. Secudb: An in- enclave privacy-preserving and tamper-resistant rela- tional database. 2024

    2024

  9. [17]

    AMD Secure Encrypted Virtualization (SEV)

    AMD. AMD Secure Encrypted Virtualization (SEV). https://developer.amd.com/sev/, 2025

    2025

  10. [18]

    com/content/www/us/en/developer/tools/ software-guard-extensions/overview.html, 2025

    Intel Software Guard Extensions.https://www.intel. com/content/www/us/en/developer/tools/ software-guard-extensions/overview.html, 2025

    2025

  11. [19]

    https://www.intel

    Intel Trust Domain Extensions. https://www.intel. com/content/www/us/en/developer/articles/ technical/intel-trust-domain-extensions. html, 2025

    2025

  12. [20]

    https://www

    Confidential Compute Architecture. https://www. arm.com/architecture/security-features/ arm-confidential-compute-architecture, 2025

    2025

  13. [21]

    https://aws.amazon.com/ ec2/nitro/nitro-enclaves/, 2025

    AWS Nitro Enclaves. https://aws.amazon.com/ ec2/nitro/nitro-enclaves/, 2025

    2025

  14. [22]

    Keystone: an open frame- work for architecting trusted execution environments

    Dayeol Lee, David Kohlbrenner, Shweta Shinde, Krste Asanovic, and Dawn Song. Keystone: an open frame- work for architecting trusted execution environments. InProceedings of the ACM European Conference on Computer Systems (EuroSys). ACM, 2020

    2020

  15. [23]

    Scal- able memory protection in the PENGLAI enclave

    Erhu Feng, Xu Lu, Dong Du, Bicheng Yang, Xueqiang Jiang, Yubin Xia, Binyu Zang, and Haibo Chen. Scal- able memory protection in the PENGLAI enclave. In Proceedings of the USENIX Symposium on Operating Systems Design and Implementation (OSDI), 2021

    2021

  16. [24]

    Twinvisor: Hardware-isolated con- fidential virtual machines for ARM

    Dingji Li, Zeyu Mi, Yubin Xia, Binyu Zang, Haibo Chen, and Haibing Guan. Twinvisor: Hardware-isolated con- fidential virtual machines for ARM. InProceedings of the ACM Symposium on Operating Systems Principles (SOSP), 2021

    2021

  17. [25]

    Guerney D. H. Hunt, Ramachandra Pai, Michael V . Le, Hani Jamjoom, Sukadev Bhattiprolu, Rick Boivie, Laurent Dufour, Brad Frey, Mohit Kapur, Kenneth A. Goldman, Ryan Grimm, Janani Janakirman, John M. Ludden, Paul Mackerras, Cathy May, Elaine R. Palmer, Bharata Bhasker Rao, Law...

    2021

  18. [26]

    Design and verification of the arm confidential compute archi- tecture

    Xupeng Li, Xuheng Li, Christoffer Dall, Ronghui Gu, Jason Nieh, Yousuf Sait, and Gareth Stockwell. Design and verification of the arm confidential compute archi- tecture. InProceedings of the USENIX Symposium on Operating Systems Design and Implementation (OSDI), 2022

    2022

  19. [27]

    https://cloud.google.com/security/products/ confidential-computing?hl=en, 2025

    Confidential VM overview - Google Cloud. https://cloud.google.com/security/products/ confidential-computing?hl=en, 2025

    2025

  20. [28]

    https://learn.microsoft.com/ en-us/azure/confidential-computing/ confidential-computing-enclaves, 2022

    Build with SGX enclaves - Azure Virtual Ma- chines. https://learn.microsoft.com/ en-us/azure/confidential-computing/ confidential-computing-enclaves, 2022

    2022

  21. [29]

    https://www.ibm.com/solutions/ confidential-computing, 2025

    IBM Cloud - Confidential computing solu- tions. https://www.ibm.com/solutions/ confidential-computing, 2025

    2025

  22. [30]

    https://www.alibabacloud.com/help/en/ack/ ack-managed-and-ack-dedicated/user-guide/ tee-based-confidential-computing/, 2025

    Alibaba Cloud - TEE-based confidential computing. https://www.alibabacloud.com/help/en/ack/ ack-managed-and-ack-dedicated/user-guide/ tee-based-confidential-computing/, 2025

    2025

  23. [31]

    Hakan Hacigümüs, Sharad Mehrotra, and Balakrishna R. Iyer. Providing database as a service. IEEE Computer Society, 2002

    2002

  24. [32]

    Ribeiro, Nuno Santos, and Nuno O

    Pedro S. Ribeiro, Nuno Santos, and Nuno O. Duarte. Db- store: A trustzone-backed database management system for mobile applications. InInternational Conference on E-Business and Telecommunication Networks, 2018

    2018

  25. [33]

    https: //www.manuelrigger.at/dbms-bugs/, 2025

    Bugs found in Database Management Systems. https: //www.manuelrigger.at/dbms-bugs/, 2025

    2025

  26. [34]

    Spectre attacks: Exploiting speculative execution

    Paul Kocher, Jann Horn, Anders Fogh, Daniel Genkin, Daniel Gruss, Werner Haas, Mike Hamburg, Moritz Lipp, Stefan Mangard, Thomas Prescher, Michael Schwarz, and Yuval Yarom. Spectre attacks: Exploiting speculative execution. pages 1–19. IEEE. 11

  27. [35]

    Hints and principles for computer system design.arXiv preprint arXiv:2011.02455, 2020

    Butler Lampson. Hints and principles for computer system design.arXiv preprint arXiv:2011.02455, 2020

    2011 arXiv

  28. [36]

    Don Dennis.A capability architecture

    T. Don Dennis.A capability architecture. PhD thesis, Purdue University, USA, 1980

    1980

  29. [37]

    Digital Press, 2014

    Henry M Levy.Capability-based computer systems. Digital Press, 2014

    2014

  30. [38]

    The unix time-sharing system.Communications of the ACM, 17(7):365–375, 1974

    Dennis M Ritchie and Ken Thompson. The unix time-sharing system.Communications of the ACM, 17(7):365–375, 1974

    1974

  31. [39]

    Accetta, Robert V

    Michael J. Accetta, Robert V . Baron, William J. Bolosky, David B. Golub, Richard F. Rashid, Avadis Tevanian, and Michael Young. Mach: A new kernel founda- tion for UNIX development. InProceedings of the USENIX Summer Conference, Altanta, GA, USA, June

  32. [40]

    USENIX Association, 1986

    1986

  33. [41]

    Sanjay Ghemawat and Jeff Dean. Leveldb. http:// code.google.com/p/leveldb, 2011

    2011

  34. [42]

    Rocksdb.http://rocksdb.org, 2013

    Facebook. Rocksdb.http://rocksdb.org, 2013

    2013

  35. [43]

    Mdb: A memory-mapped database and backend for openldap

    Howard Chu. Mdb: A memory-mapped database and backend for openldap. InProceedings of the 3rd Inter- national Conference on LDAP , Heidelberg, Germany, volume 35, page 34, 2011

    2011

  36. [44]

    The slab allocator: An object- caching kernel memory allocator

    Jeff Bonwick et al. The slab allocator: An object- caching kernel memory allocator. InUSENIX summer, volume 16. Boston, MA, USA, 1994

    1994

  37. [45]

    Rethink the sync.ACM Transactions on Computer Systems (TOCS), 2008

    Edmund B Nightingale, Kaushik Veeraraghavan, Pe- ter M Chen, and Jason Flinn. Rethink the sync.ACM Transactions on Computer Systems (TOCS), 2008

    2008

  38. [46]

    https://developer.arm.com/ documentation/ddi0487/latest/, 2025

    Arm architecture reference manual for a-profile architecture. https://developer.arm.com/ documentation/ddi0487/latest/, 2025

    2025

  39. [47]

    https://www.kernel.org/ doc/html/latest/admin-guide/device-mapper/ dm-integrity.html, 2025

    Linux dm-integrity. https://www.kernel.org/ doc/html/latest/admin-guide/device-mapper/ dm-integrity.html, 2025

    2025

  40. [48]

    https://www.kernel.org/doc/ html/latest/admin-guide/device-mapper/ dm-crypt.html, 2025

    Linux dm-crypt. https://www.kernel.org/doc/ html/latest/admin-guide/device-mapper/ dm-crypt.html, 2025

    2025

  41. [49]

    Sysbench benchmark tool, 2021

    Oracle. Sysbench benchmark tool, 2021

    2021

  42. [50]

    https://trustedfirmware-m.readthedocs

    Rollback protection in TF-M secure boot. https://trustedfirmware-m.readthedocs. io/en/latest/design_docs/booting/secure_ boot_rollback_protection.html, 2020

    2020

  43. [51]

    Sebastian Angel, Aditya Basu, Weidong Cui, Trent Jaeger, Stella Lau, Srinath T. V . Setty, and Sudheesh Singanamalla. Nimble: Rollback protection for confi- dential cloud services. InProceedings of the USENIX Symposium on Operating Systems Design and Imple- mentation (OSDI), 2023

    2023

  44. [52]

    Rollbaccine: Herd Immunity against Storage Rollback Attacks in TEEs.Proceedings of the ACM SIGMOD Conference, 2026

    David Chu, Aditya Balasubramanian, Dee Bao, Natacha Crooks, Heidi Howard, Lucky E Katahanas, and Sou- janya Ponnapalli. Rollbaccine: Herd Immunity against Storage Rollback Attacks in TEEs.Proceedings of the ACM SIGMOD Conference, 2026

    2026

  45. [53]

    Probabilistic en- cryption & how to play mental poker keeping secret all partial information

    Shafi Goldwasser and Silvio Micali. Probabilistic en- cryption & how to play mental poker keeping secret all partial information. InProceedings of the Fourteenth An- nual ACM Symposium on Theory of Computing, STOC ’82, page 365–377

  46. [54]

    Rishabh Poddar, Tobias Boelter, and Raluca A. Popa. Arx: An encrypted database using semantically secure encryption. InProceedings of the International Confer- ence on Very Large Data Bases (VLDB), 2019

    2019

  47. [55]

    Core slicing: closing the gap between leaky confidential vms and bare-metal cloud

    Ziqiao Zhou, Yizhou Shan, Weidong Cui, Xinyang Ge, Marcus Peinado, and Andrew Baumann. Core slicing: closing the gap between leaky confidential vms and bare-metal cloud. InProceedings of the USENIX Sympo- sium on Operating Systems Design and Implementation (OSDI), pages 247–26...

    2023

  48. [56]

    Sharing is leaking: blocking transient-execution attacks with core-gapped confidential vms

    Charly Castes and Andrew Baumann. Sharing is leaking: blocking transient-execution attacks with core-gapped confidential vms. InProceedings of the International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), 2024

    2024

  49. [57]

    Controlled-channel attacks: Deterministic side channels for untrusted operating systems

    Yuanzhong Xu, Weidong Cui, and Marcus Peinado. Controlled-channel attacks: Deterministic side channels for untrusted operating systems. pages 640–656. IEEE Computer Society, 2015

    2015

  50. [58]

    Fang, Chia-Che Tsai, and Raluca Ada Popa

    Dayeol Lee, Dongha Jung, Ian T. Fang, Chia-Che Tsai, and Raluca Ada Popa. An off-chip attack on hardware enclaves via the memory bus. InProceedings of the USENIX Security Symposium, pages 487–504. USENIX Association, 2020

    2020

  51. [59]

    How to build a trusted database system on untrusted storage

    Umesh Maheshwari, Radek Vingralek, and William Shapiro. How to build a trusted database system on untrusted storage. InProceedings of the USENIX Sym- posium on Operating Systems Design and Implemen- tation (OSDI), pages 135–150. USENIX Association, 2000

    2000

  52. [60]

    Popa, Catherine M

    Raluca A. Popa, Catherine M. S. Redfield, Nickolai Zel- dovich, and Hari Balakrishnan. Cryptdb: protecting confidentiality with encrypted query processing. InPro- ceedings of the ACM Symposium on Operating Systems Principles (SOSP), 2011. 12

    2011

  53. [61]

    Frans Kaashoek, Samuel Madden, and Nickolai Zeldovich

    Stephen Tu, M. Frans Kaashoek, Samuel Madden, and Nickolai Zeldovich. Processing analytical queries over encrypted data. InProceedings of the International Conference on Very Large Data Bases (VLDB), 2013

    2013

  54. [62]

    HE3DB: an efficient and elastic encrypted database via arithmetic-and-logic fully homomorphic encryption

    Song Bian, Zhou Zhang, Haowen Pan, Ran Mao, Zian Zhao, Yier Jin, and Zhenyu Guan. HE3DB: an efficient and elastic encrypted database via arithmetic-and-logic fully homomorphic encryption. InProceedings of the ACM Conference on Computer and Communications Security (CCS), 2023

    2023

  55. [63]

    HEDA: multi-attribute unbounded aggregation over homomor- phically encrypted database

    Xuanle Ren, Le Su, Zhen Gu, Sheng Wang, Feifei Li, Yuan Xie, Song Bian, Chao Li, and Fan Zhang. HEDA: multi-attribute unbounded aggregation over homomor- phically encrypted database. InProceedings of the Inter- national Conference on Very Large Data Bases (VLDB), 2022

    2022

  56. [64]

    Engorgio: An arbitrary- precision unbounded-size hybrid encrypted database via quantized fully homomorphic encryption

    Song Bian, Haowen Pan, Jiaqi Hu, Zhou Zhang, Yun- hao Fu, Jiafeng Hua, Yunyi Chen, Bo Zhang, Yier Jin, Jin Dong, and Zhenyu Guan. Engorgio: An arbitrary- precision unbounded-size hybrid encrypted database via quantized fully homomorphic encryption. InProceed- ings of the USENI...

    2025

  57. [65]

    Arcedb: An arbitrary-precision encrypted database via (amor- tized) modular homomorphic encryption

    Zhou Zhang, Song Bian, Zian Zhao, Ran Mao, Haoyi Zhou, Jiafeng Hua, Yier Jin, and Zhenyu Guan. Arcedb: An arbitrary-precision encrypted database via (amor- tized) modular homomorphic encryption. InProceed- ings of the ACM Conference on Computer and Com- munications Security (C...

    2024

  58. [66]

    Hermes: High-performance homo- morphically encrypted vector databases.arXiv preprint arXiv:2506.03308, 2025

    Dongfang Zhao. Hermes: High-performance homo- morphically encrypted vector databases.arXiv preprint arXiv:2506.03308, 2025

    2025 arXiv

  59. [67]

    Oblix: An efficient oblivious search index

    Pratyush Mishra, Rishabh Poddar, Jerry Chen, Alessan- dro Chiesa, and Raluca Ada Popa. Oblix: An efficient oblivious search index. In2018 IEEE symposium on security and privacy (SP), pages 279–296. IEEE, 2018

    2018

  60. [68]

    Obladi: Oblivious serializable transactions in the cloud

    Natacha Crooks, Matthew Burke, Ethan Cecchetti, Sitar Harel, Rachit Agarwal, and Lorenzo Alvisi. Obladi: Oblivious serializable transactions in the cloud. InPro- ceedings of the USENIX Symposium on Operating Sys- tems Design and Implementation (OSDI), pages 727– 743, 2018

    2018

  61. [69]

    Compass: encrypted semantic search with high accuracy

    Jinhao Zhu, Liana Patel, Matei Zaharia, and Raluca Ada Popa. Compass: encrypted semantic search with high accuracy. InProceedings of the USENIX Symposium on Operating Systems Design and Implementation (OSDI), pages 915–938, 2025

    2025

  62. [70]

    Opaque: An oblivious and encrypted distributed analyt- ics platform

    Wenting Zheng, Ankur Dave, Jethro G Beekman, Raluca Ada Popa, Joseph E Gonzalez, and Ion Stoica. Opaque: An oblivious and encrypted distributed analyt- ics platform. InProceedings of the USENIX Sympo- sium on Networked Systems Design and Implementation (NSDI), 2017

    2017

  63. [71]

    Secure data analytics in apache spark with fine-grained policy enforcement and isolated exe- cution

    Byeongwook Kim, Jaewon Hur, Adil Ahmad, and By- oungyoung Lee. Secure data analytics in apache spark with fine-grained policy enforcement and isolated exe- cution. InProceedings of the Network and Distributed System Security Symposium (NDSS). The Internet Soci- ety, 2025

    2025

  64. [72]

    Chia-Che Tsai, Jeongseok Son, Bhushan Jain, John McAvey, Raluca Ada Popa, and Donald E. Porter. Civet: An efficient java partitioning framework for hardware enclaves. InProceedings of the USENIX Security Sym- posium, 2020

    2020

  65. [73]

    Vc3: Trustworthy data analytics in the cloud using sgx

    Felix Schuster, Manuel Costa, Cédric Fournet, Christos Gkantsidis, Marcus Peinado, Gloria Mainar-Ruiz, and Mark Russinovich. Vc3: Trustworthy data analytics in the cloud using sgx. InIEEE S&P, 2015

    2015

  66. [74]

    Eyers, Rüdiger Kapitza, Pe- ter R

    Sergei Arnautov, Bohdan Trach, Franz Gregor, Thomas Knauth, André Martin, Christian Priebe, Joshua Lind, Divya Muthukumaran, Dan O’Keeffe, Mark Stillwell, David Goltzsche, David M. Eyers, Rüdiger Kapitza, Pe- ter R. Pietzuch, and Christof Fetzer. SCONE: secure linux containers...

    2016

  67. [75]

    Ofir Weisse, Valeria Bertacco, and Todd M. Austin. Re- gaining lost cycles with hotcalls: A fast interface for SGX secure enclaves. InProceedings of the Interna- tional Symposium on Computer Architecture (ISCA), 2017

    2017

  68. [76]

    SPE- ICHER: securing lsm-based key-value stores using shielded execution

    Maurice Bailleu, Jörg Thalheim, Pramod Bhatotia, Christof Fetzer, Michio Honda, and Kapil Vaswani. SPE- ICHER: securing lsm-based key-value stores using shielded execution. InProceedings of the USENIX Con- ference on File and Storage Technologies (FAST), 2019

    2019

  69. [77]

    Pietzuch

    Jörg Thalheim, Harshavardhan Unnibhavi, Christian Priebe, Pramod Bhatotia, and Peter R. Pietzuch. rkt-io: a direct I/O stack for shielded execution. InProceedings of the ACM European Conference on Computer Systems (EuroSys), 2021

    2021

  70. [78]

    Shieldstore: Shielded in- memory key-value storage with sgx

    Taehoon Kim, Joongun Park, Jaewook Woo, Seungheun Jeon, and Jaehyuk Huh. Shieldstore: Shielded in- memory key-value storage with sgx. InProceedings of the ACM European Conference on Computer Systems (EuroSys), 2019. 13

    2019

Pith tools

Reviewed July 13, 2026 · model on record in the stance chip above.