pith. sign in

arxiv: 2506.03308 · v5 · submitted 2025-06-03 · 💻 cs.CR · cs.DB

Hermes: Efficient Global Homomorphic Aggregation over Mutable Packed Ciphertexts

Dongfang Zhao This is my paper

Pith reviewed 2026-05-19 10:31 UTC · model grok-4.3

classification 💻 cs.CR cs.DB
keywords fully homomorphic encryptionpacked ciphertextsglobal aggregationin-place mutabilityencrypted databasesSIMD packinghomomorphic updatesaggregation queries
0
0 comments X

The pith

Hermes packs precomputed aggregates into each ciphertext to enable constant-time global aggregations and in-place updates on encrypted databases.

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

The paper presents Hermes as a system for efficient aggregation queries and dynamic tuple updates on homomorphically encrypted databases. It introduces a SIMD-aware packed data model that embeds precomputed aggregate statistics directly into each ciphertext. This design supports constant-time global aggregations without expensive Galois automorphisms. Homomorphic algorithms based on polynomial slot masking and shifting provide true in-place mutability while preserving security under the IND-CPA model. The work scopes itself to unconditional global aggregations to deliver both high performance and dynamic updates simultaneously.

Core claim

Hermes embeds precomputed aggregate statistics into each packed ciphertext using a SIMD-aware model, allowing global aggregations in constant time without Galois automorphisms, and provides provably secure homomorphic algorithms for polynomial slot masking and shifting to enable in-place mutability of tuples in encrypted databases.

What carries the argument

The SIMD-aware packed data model that embeds precomputed aggregate statistics directly into each ciphertext, together with polynomial slot masking and shifting algorithms for updates.

If this is right

  • Global aggregation queries run in constant time on packed encrypted data.
  • Tuple insertions and deletions occur in place without full ciphertext re-encryption.
  • Query throughput improves on analytical workloads compared to standard FHE approaches.
  • The system integrates directly as loadable functions in MySQL for practical use.
  • Performance gains appear on TPC-H and real-world datasets for dynamic encrypted databases.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The packing technique could reduce overhead in other FHE applications that need frequent statistical summaries.
  • Hybrid approaches might combine Hermes-style packing with separate mechanisms for conditional or grouped queries.
  • Wider adoption could narrow the practical gap between encrypted and plaintext database analytics.
  • Testing on larger-scale dynamic workloads would clarify how the constant-time property scales.

Load-bearing premise

The design assumes workloads are limited to unconditional global aggregations to enable constant-time performance and in-place updates.

What would settle it

Measuring whether a global aggregation on the Hermes MySQL implementation completes without Galois automorphisms and in time independent of data volume would test the constant-time claim.

read the original abstract

Fully Homomorphic Encryption (FHE) promises the ability to compute over encrypted data without revealing sensitive contents. However, enabling high-frequency updates and statistical analysis in outsourced databases remains elusive due to the structural mismatch between mutable database records and the cryptographically expensive mutability of FHE ciphertexts. This paper presents Hermes, a prototype system tailored for efficient aggregation queries and dynamic tuple updates on homomorphically encrypted databases. The core design of Hermes is twofold. First, to amortize FHE costs and accelerate unconditional aggregations, Hermes introduces a SIMD-aware packed data model that embeds precomputed aggregate statistics directly into each ciphertext, enabling constant-time global aggregations without expensive Galois automorphisms. Second, to support true in-place mutability, we develop homomorphic algorithms based on polynomial slot masking and shifting, which are provably secure under the standard IND-CPA model. We scope Hermes to unconditional global aggregations to achieve both high performance and in-place updates simultaneously, two properties that prior FHE database systems have not delivered at scale. Hermes is implemented as a suite of C++ loadable functions in MySQL. Extensive evaluations on the TPC-H benchmark and three real-world datasets demonstrate significant performance improvements in query throughput, tuple insertions, and tuple deletions compared to conventional FHE implementations, validating its efficacy for highly dynamic and analytical workloads.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper presents Hermes, a prototype system for homomorphically encrypted databases supporting efficient unconditional global aggregation queries and dynamic tuple updates. The core contributions are a SIMD-aware packed data model that embeds precomputed aggregate statistics into each ciphertext to enable constant-time global aggregations without Galois automorphisms, and homomorphic algorithms based on polynomial slot masking and shifting to achieve true in-place mutability. These are claimed to be secure under the standard IND-CPA model. The system is implemented as C++ loadable functions in MySQL and evaluated on the TPC-H benchmark plus three real-world datasets, reporting significant improvements in query throughput, insertions, and deletions relative to conventional FHE approaches.

Significance. If the performance and security claims hold, Hermes would represent a meaningful step toward practical FHE databases by simultaneously delivering high-frequency updates and fast global statistics, two properties that prior systems have struggled to combine at scale. The explicit scoping to unconditional global aggregations is a constructive design choice that directly enables the constant-time and mutability properties within well-defined limits, and the MySQL integration provides a concrete path to deployment.

major comments (2)
  1. [Security Analysis] Security section: the claim that the slot-masking and shifting algorithms are 'provably secure under the standard IND-CPA model' is central to the mutability contribution, yet the manuscript provides only a high-level argument rather than a full reduction or game-based proof; this absence prevents verification of the security claim that underpins the in-place update feature.
  2. [Evaluation] Evaluation section and associated tables: the reported throughput and latency improvements lack detailed experimental controls (number of independent runs, noise-growth measurements, variance, and explicit comparison against prior FHE database systems restricted to the same unconditional-global scope), which is load-bearing for the central performance claims.
minor comments (2)
  1. [Abstract] The abstract and introduction should explicitly list the three real-world datasets used in the evaluation for reproducibility.
  2. [Design] Notation for slot indices and masking polynomials could be introduced earlier and used consistently to improve readability of the algorithmic descriptions.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thoughtful and constructive review. We appreciate the positive assessment of Hermes's potential contribution toward practical FHE databases and the recognition of our design scoping. We address each major comment below, outlining the revisions we will make to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Security Analysis] Security section: the claim that the slot-masking and shifting algorithms are 'provably secure under the standard IND-CPA model' is central to the mutability contribution, yet the manuscript provides only a high-level argument rather than a full reduction or game-based proof; this absence prevents verification of the security claim that underpins the in-place update feature.

    Authors: We agree that the current high-level argument, while grounded in the composition of standard FHE operations (linear masking and shifting via additions, constant multiplications, and rotations that preserve the IND-CPA property under RLWE), would benefit from greater formality. In the revised manuscript we will expand the Security section with an explicit game-based proof. We will define a sequence of games showing that any efficient adversary against the slot-masking or shifting algorithms can be used to break the IND-CPA security of the underlying FHE scheme, with concrete advantage bounds. This will include the formal security game definitions and the reduction steps. revision: yes

  2. Referee: [Evaluation] Evaluation section and associated tables: the reported throughput and latency improvements lack detailed experimental controls (number of independent runs, noise-growth measurements, variance, and explicit comparison against prior FHE database systems restricted to the same unconditional-global scope), which is load-bearing for the central performance claims.

    Authors: We acknowledge the value of these controls for reproducibility. In the revised evaluation section and tables we will report: averages and standard deviations over 10 independent runs per experiment; explicit noise-growth measurements (in bits) for each operation; and variance statistics. For comparisons, we will add a dedicated subsection that includes direct or normalized performance figures against prior FHE database systems whose query scope is restricted to unconditional global aggregations, while clearly stating any remaining differences in supported functionality or implementation assumptions. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The Hermes design introduces a new SIMD-aware packed ciphertext model that embeds precomputed aggregates to enable constant-time global sums without Galois operations, paired with polynomial slot masking and shifting for in-place updates. These are explicit engineering choices scoped to unconditional global aggregations, implemented as MySQL loadable functions, and validated through TPC-H and real-world dataset experiments. No derivation step reduces by construction to fitted parameters, self-referential definitions, or load-bearing self-citations; the central claims rest on the proposed algorithms and reported performance gains rather than tautological inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Based on abstract only; no explicit free parameters or invented entities are described. The main assumption is standard cryptographic security.

axioms (1)
  • domain assumption Homomorphic algorithms based on polynomial slot masking and shifting are provably secure under the standard IND-CPA model.
    Stated directly in the abstract as the security foundation for the mutability feature.

pith-pipeline@v0.9.0 · 5759 in / 1186 out tokens · 35517 ms · 2026-05-19T10:31:25.971374+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Reliable Non-Leveled Homomorphic Encryption for Web Services

    cs.CR 2025-08 unverdicted novelty 4.0

    Proposes a reliable non-leveled FHE scheme with new encoding and an algebraic reliability layer to reduce overhead and add built-in error correction for web services.

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages · cited by 1 Pith paper

  1. [1]

    Adi Akavia, Neta Oren, Boaz Sapir, and Margarita Vald. 2023. CSHER:ASystemforCompactStoragewithHE-Retrieval. In Proceedings of the 32nd USENIX Conference on Security Symposium . USENIX Association

    work page 2023

  2. [2]

    Accessed October 30, 2016

    Apache Lucene. Accessed October 30, 2016. https://lucene.apache.org

    work page 2016

  3. [3]

    Ahmad Al Badawi, Jack Bates, Flavio Bergamaschi, David Bruce Cousins, Saroja Erabelli, Nicholas Genise, Shai Halevi, Hamish Hunt, Andrey Kim, Yongwoo Lee, Zeyu Liu, Daniele Micciancio, Ian Quah, Yuriy Polyakov, Saraswathy R.V., Kurt Rohloff, Jonathan Saylor, Dmitriy Suponitsky, Matthew Triplett, Vinod Vaikuntanathan, and Vincent Zucca. 2022. OpenFHE: Open...

    work page 2022

  4. [4]

    Accessed 2022

    Bitcoin Trade History. Accessed 2022. https://www.blockchain.com/charts/trade- volume

    work page 2022

  5. [5]

    Marcelo Blatt, Alexander Gusev, Yuriy Polyakov, Kurt Rohloff, and Vinod Vaikun- tanathan. 2020. Optimized homomorphic encryption solution for secure genome- wide association studies. BMC Medical Genomics 13 (2020), 1–13

    work page 2020

  6. [6]

    Fabian Boemer, Rosario Cammarota, Daniel Demmler, Thomas Schneider, and Hossein Yalame. 2020. MP2ML: A mixed-protocol machine learning framework for private inference. In Proceedings of the 15th international conference on avail- ability, reliability and security. 1–10

    work page 2020

  7. [7]

    Jean-Philippe Bossuat, Christian Mouchet, Juan Troncoso-Pastoriza, and Jean- Pierre Hubaux. 2021. Efficient Bootstrapping for Approximate Homomorphic Encryption with Non-sparse Keys. In Advances in Cryptology – EUROCRYPT 2021, Anne Canteaut and François-Xavier Standaert (Eds.). Springer International Publishing, Cham, 587–617

    work page 2021

  8. [8]

    Zvika Brakerski. 2012. Fully homomorphic encryption without modulus switch- ing from classical GapSVP. Annual cryptology conference (2012), 868–886

    work page 2012

  9. [9]

    Zvika Brakerski, Craig Gentry, and Vinod Vaikuntanathan. 2014. (Leveled) fully homomorphic encryption without bootstrapping. ACM Transactions on Computation Theory (TOCT) 6, 3 (2014), 1–36

    work page 2014

  10. [10]

    Jung Hee Cheon, Andrey Kim, Miran Kim, and Yongsoo Song. 2017. Homomor- phic Encryption for Arithmetic of Approximate Numbers.Advances in Cryptology – ASIACRYPT 2017 (2017), 409–437

    work page 2017

  11. [11]

    Ilaria Chillotti, Nicolas Gama, Mariya Georgieva, and Malika Izabachène. 2018. TFHE: Fast Fully Homomorphic Encryption over the Torus. Cryptology ePrint Archive, Paper 2018/421. https://eprint.iacr.org/2018/421

    work page 2018

  12. [12]

    Accessed 2023

    Chroma. Accessed 2023. http://trychroma.com/

    work page 2023

  13. [13]

    Accessed 2022

    Covid-19 Data. Accessed 2022. https://covidtracking.com/data/download/ national-history.csv

    work page 2022

  14. [14]

    Accessed 2023

    Faiss. Accessed 2023. http://github.com/facebookresearch/faiss

    work page 2023

  15. [15]

    Junfeng Fan and Frederik Vercauteren. 2012. Somewhat Practical Fully Homo- morphic Encryption. Cryptology ePrint Archive, Paper 2012/144 (2012). https: //eprint.iacr.org/2012/144

    work page 2012

  16. [16]

    Craig Gentry. 2009. Fully homomorphic encryption using ideal lattices. Proceed- ings of the forty-first annual ACM symposium on Theory of computing (2009)

    work page 2009

  17. [17]

    Antonio Guimarães, Edson Borin, and Diego F. Aranha. 2024. MOSFHET: Opti- mized Software for FHE over the Torus. IACR Cryptol. ePrint Arch. 2022 (2024),

    work page 2024

  18. [18]

    https://api.semanticscholar.org/CorpusID:248661190

    work page

  19. [19]

    Hakan Hacigümüş, Bala Iyer, Chen Li, and Sharad Mehrotra. 2002. Executing SQL over Encrypted Data in the Database-Service-Provider Model. In Proceedings of the 2002 ACM SIGMOD International Conference on Management of Data (Madison, Wisconsin) (SIGMOD ’02). Association for Computing Machinery, New York, NY, USA, 216–227. https://doi.org/10.1145/564691.564717

    work page doi:10.1145/564691.564717 2002

  20. [20]

    Boyoung Han, Yeonghyeon Kim, Jina Choi, Hojune Shin, and Younho Lee. 2023. Fully Homomorphic Privacy-Preserving Naive Bayes Machine Learning and Classification. In Proceedings of the 11th Workshop on Encrypted Computing & Applied Homomorphic Cryptography (Copenhagen, Denmark) (W AHC ’23). Association for Computing Machinery, New York, NY, USA, 91–102. htt...

    work page doi:10.1145/3605759.3625262 2023

  21. [21]

    Accessed 2022

    HElib. Accessed 2022. https://github.com/homenc/HElib

    work page 2022

  22. [22]

    Accessed 2022

    Human Genome Databases. Accessed 2022. http://hgdownload.soe.ucsc.edu/ goldenPath/hg38/database/

    work page 2022

  23. [23]

    Ilia Iliashenko and Vincent Zucca. 2021. Faster homomorphic comparison opera- tions for BGV and BFV. Proceedings on Privacy Enhancing Technologies 2021, 3 (2021), 246–264

    work page 2021

  24. [24]

    Gunawi, Cody Hammock, Joe Mambretti, Alexander Barnes, François Halbach, Alex Rocha, and Joe Stubbs

    Kate Keahey, Jason Anderson, Zhuo Zhen, Pierre Riteau, Paul Ruth, Dan Stanzione, Mert Cevik, Jacob Colleran, Haryadi S. Gunawi, Cody Hammock, Joe Mambretti, Alexander Barnes, François Halbach, Alex Rocha, and Joe Stubbs. 2020. Lessons Learned from the Chameleon Testbed. In Proceedings of the 2020 USENIX Annual Technical Conference (USENIX ATC ’20). USENIX...

    work page 2020

  25. [25]

    Rishabh Poddar, Tobias Boelter, and Raluca Ada Popa. 2019. Arx: An Encrypted Database using Semantically Secure Encryption. Proc. VLDB Endow. 12, 11 (2019), 1664–1678. https://doi.org/10.14778/3342263.3342641

    work page doi:10.14778/3342263.3342641 2019

  26. [26]

    Raluca Ada Popa, Catherine Redfield, Nickolai Zeldovich, and Hari Balakrishnan

    work page

  27. [27]

    In Proceedings of the Twenty-Third ACM Symposium on Operating Systems Principles

    CryptDB: protecting confidentiality with encrypted query processing. In Proceedings of the Twenty-Third ACM Symposium on Operating Systems Principles . ACM, 85–100

    work page

  28. [28]

    Accessed 2023

    Qdrant. Accessed 2023. http://qdrant.tech/

    work page 2023

  29. [29]

    Savvas Savvides, Darshika Khandelwal, and Patrick Eugster. 2020. Effi- cient Confidentiality-Preserving Data Analytics over Symmetrically Encrypted Datasets. Proc. VLDB Endow. 13, 8 (April 2020), 1290–1303. https://doi.org/10. 14778/3389133.3389144

    work page arXiv 2020

  30. [30]

    Microsoft SEAL (release 3.7)

    SEAL 2021. Microsoft SEAL (release 3.7). https://github.com/Microsoft/SEAL. Microsoft Research, Redmond, WA

    work page 2021

  31. [31]

    Tawose, Jun Dai, Lei Yang, and Dongfang Zhao

    Olamide T. Tawose, Jun Dai, Lei Yang, and Dongfang Zhao. 2023. Toward Efficient Homomorphic Encryption for Outsourced Databases through Parallel Caching. Proceedings of the ACM on Management of Data (SIGMOD) , Article 66 (May 2023), 23 pages. https://doi.org/10.1145/3588920

    work page doi:10.1145/3588920 2023

  32. [32]

    Sharon Hu, and Taeho Jung

    Jonathan Takeshita, Dayane Reis, Ting Gong, Michael Niemier, X. Sharon Hu, and Taeho Jung. 2023. Accelerating Finite-Field and Torus FHE via Compute-Enabled (S)RAM. IEEE Trans. Comput. (2023), 1–14. https://doi.org/10.1109/TC.2023. 3301116

    work page doi:10.1109/tc.2023 2023

  33. [33]

    Accessed 2023

    Vald. Accessed 2023. http://vald.vdaas.org/

    work page 2023

  34. [34]

    Michiel Van Beirendonck, Jan-Pieter D’Anvers, Furkan Turan, and Ingrid Ver- bauwhede. 2023. FPT: A Fixed-Point Accelerator for Torus Fully Homomor- phic Encryption. In Proceedings of the 2023 ACM SIGSAC Conference on Com- puter and Communications Security (Copenhagen, Denmark) (CCS ’23). As- sociation for Computing Machinery, New York, NY, USA, 741–755. h...

    work page doi:10.1145/3576915.3623159 2023

  35. [35]

    Accessed 2023

    Weaviate. Accessed 2023. http://weaviate.io/

    work page 2023

  36. [36]

    Chuangxian Wei, Bin Wu, Sheng Wang, Renjie Lou, Chaoqun Zhan, Feifei Li, and Yuanzhe Cai. 2020. AnalyticDB-V: A Hybrid Analytical Engine towards Query Fusion for Structured and Unstructured Data. Proc. VLDB Endow. 13, 12 (aug 2020), 3152–3165. https://doi.org/10.14778/3415478.3415541

    work page doi:10.14778/3415478.3415541 2020

  37. [37]

    Guowen Xu, Guanlin Li, Shangwei Guo, Tianwei Zhang, and Hongwei Li. 2023. Secure Decentralized Image Classification With Multiparty Homomorphic En- cryption. IEEE Transactions on Circuits and Systems for Video Technology 33, 7 (2023), 3185–3198

    work page 2023

  38. [38]

    Wen Yang, Tao Li, Gai Fang, and Hong Wei. 2020. PASE: PostgreSQL Ultra-High- Dimensional Approximate Nearest Neighbor Search Extension. In Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data (Portland, OR, USA) (SIGMOD ’20). Association for Computing Machinery, New York, NY, USA, 2241–2253. https://doi.org/10.1145/3318464.3386131

    work page doi:10.1145/3318464.3386131 2020

  39. [39]

    Chaoqun Zhan, Maomeng Su, Chuangxian Wei, Xiaoqiang Peng, Liang Lin, Sheng Wang, Zhe Chen, Feifei Li, Yue Pan, Fang Zheng, and Chengliang Chai. 2019. AnalyticDB: Real-Time OLAP Database System at Alibaba Cloud. Proc. VLDB Endow. 12, 12 (aug 2019), 2059–2070. https://doi.org/10.14778/3352063.3352124 16

    work page doi:10.14778/3352063.3352124 2019