Abstract
Secure computation provides cryptographic protocols for collaborative applications with private inputs and outputs. In this paper, we examine a collection of protocols for secure evaluation of polynomials using secure floating-point arithmetic. The main goal is to provide a comparative analysis of their construction, complexity, performance, and tradeoffs in different application settings. The analysis demonstrates the performance gains that can be obtained by evaluating the polynomials using optimized secure multi-operand arithmetic instead of relying on generic constructions based on two-operand arithmetic. It also examines the relations between performance and complexity metrics for different execution environments (LAN, Internet), floating-point precision, and problem sizes. These protocols are part of a framework for secure multiparty computation with fixed-point and floating-point numbers based on Shamir secret sharing and related techniques.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
\(\mathsf {Div2mP}\) returns \(\bar{c}=\lfloor \bar{a}/2^m \rfloor + u\), where \(u\in \{0,1\}\) and \(u=1\) with probability \(p = \frac{\bar{a} \bmod 2^m}{2^m}\). For example, if \(\bar{a}=46\) and \(m=3\) then \(\bar{a}/2^m = 5.75\); the output is \(\bar{c}=6\) with probability \(p=0.75\) or \(\bar{c}=5\) with probability \(1-p=0.25\).
- 2.
This bound is determined as follows: Let \(\{\bar{v}_i\}_{i=0}^m\) the significands of \(\{\hat{a}_i \hat{x}^i\}_{i=0}^m\) after radix-point alignment, with up to \(\ell + \lceil i / 2^\theta \rceil \) bits, and \(\sigma = \lceil m / 2^\theta \rceil \). \(\mathsf {SumFL}\) computes \(\sum _{i=0}^m \bar{v}_i = \bar{v}_0 + \sum _{k=0}^{\sigma -2} \sum _{t=1}^{2^\theta } \bar{v}_{k 2^\theta + t} + \sum _{t=1}^{m \bmod 2^\theta } \bar{v}_{(\sigma -1) 2^\theta + t}< 2^\ell + 2^{\theta + \ell + 1} \sum _{k=0}^{\sigma -2} 2^{k} + 2^{\ell +\sigma }(m \bmod 2^\theta ) < 2^{\ell + \sigma + \theta +1}\). So the maximum bitlength is \(\ell + \lceil m / 2^\theta \rceil + \theta + 1\) bits.
- 3.
For example, if \(m=64\) and \(\theta = 0\) the modulus grows by 128 bits. If \(\theta = 3\), it grows by 16 bits, at the cost of fully normalizing 7 out of 127 multiplications.
- 4.
For example, if \(m=64\) and \(t=64\) the modulus grows by 62 bits. If \(t=16\) it grows by 14 bits and \(\beta =t'=4\). However, if \(t=16\) then \(\mathsf {ProdFL2}\) needs 6 more rounds and \(\beta (2\ell +4(t-1))\) more interactive primitives.
- 5.
Differences between the measured values in Table 4 and those computed based on Table 2 are due to simplified complexity formulas, implementation tradeoffs between round optimization and modularity, and precomputation optimizations. Table 4 lists between brackets the minimum number of rounds computed using the exact formulas.
References
Aliasgari, M., Blanton, M., Zhang, Y., Steele, A.: Secure computation on floating point numbers. In: 20th Annual Network and Distributed System Security Symposium (NDSS 2013) (2013)
Aliasgari, M., Blanton, M., Bayatbabolghani, F.: Secure computation of hidden Markov models and secure floating-point arithmetic in the malicious model. Int. J. Inf. Secur. 16(6), 577–601 (2017). https://doi.org/10.1007/s10207-016-0350-0
Aly, A., et al.: SCALE and MAMBA documentation. https://homes.esat.kuleuven.be/~nsmart/SCALE/. Accessed Apr 2020
Aly, A., Smart, N.P.: Benchmarking privacy preserving scientific operations. In: Deng, R.H., Gauthier-Umaña, V., Ochoa, M., Yung, M. (eds.) ACNS 2019. LNCS, vol. 11464, pp. 509–529. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-21568-2_25
Bogdanov, D., Kamm, L., Laur, S., Sokk, V.: Rmind: a tool for cryptographically secure statistical analysis. IEEE Trans. Dependable Secure Comput. 15(03), 481–495 (2018)
Bogdanov, D., Niitsoo, M., Toft, T., Willemson, J.: High-performance secure multi-party computation for data mining applications. Int. J. Inf. Secur. 11(6), 403–418 (2012). https://doi.org/10.1007/s10207-012-0177-2
Catrina, O.: Round-efficient protocols for secure multiparty fixed-point arithmetic. In: 12th International Conference on Communications (COMM 2018), pp. 431–436. IEEE (2018)
Catrina, O.: Optimization and tradeoffs in secure floating-point computation: products, powers, and polynomials. In: 6th Conference on the Engineering of Computer Based Systems (ECBS 2019), pp. 7:1–7:10. ACM (2019)
Catrina, O.: Evaluation of floating-point arithmetic protocols based on Shamir secret sharing. In: Obaidat, M.S. (ed.) ICETE 2019. CCIS, vol. 1247, pp. 108–131. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-52686-3_5
Catrina, O.: Optimizing secure floating-point arithmetic: sums, dot products, and polynomials. Proc. Rom. Acad. (Ser. A) 21(1), 21–28 (2020)
Catrina, O.: Performance analysis of secure floating-point sums and dot products. In: 13th International Conference on Communications (COMM 2020), pp. 465–470. IEEE (2020)
Cramer, R., Damgård, I., Ishai, Y.: Share conversion, pseudorandom secret-sharing and applications to secure computation. In: Kilian, J. (ed.) TCC 2005. LNCS, vol. 3378, pp. 342–362. Springer, Heidelberg (2005). https://doi.org/10.1007/978-3-540-30576-7_19
Cramer, R., Damgård, I., Nielsen, J.B.: Secure Multiparty Computation and Secret Sharing. Cambridge University Press, Cambridge (2015)
Damgård, I., Thorbek, R.: Non-interactive proofs for integer multiplication. In: Naor, M. (ed.) EUROCRYPT 2007. LNCS, vol. 4515, pp. 412–429. Springer, Heidelberg (2007). https://doi.org/10.1007/978-3-540-72540-4_24
Kamm, L., Willemson, J.: Secure floating point arithmetic and private satellite collision analysis. Int. J. Inf. Secur. 14(6), 531–548 (2015). https://doi.org/10.1007/s10207-014-0271-8
Kerik, L., Laud, P., Randmets, J.: Optimizing MPC for robust and scalable integer and floating-point arithmetic. In: Clark, J., Meiklejohn, S., Ryan, P.Y.A., Wallach, D., Brenner, M., Rohloff, K. (eds.) FC 2016. LNCS, vol. 9604, pp. 271–287. Springer, Heidelberg (2016). https://doi.org/10.1007/978-3-662-53357-4_18
Knuth, D.E.: The Art of Computer Programming, volume 2: Seminumerical Algorithms, 3rd edn. Addison-Wesley, Boston (1997)
Acknowledgements
Part of this work was supported by POC72/1/2, nr.127454, “SECREDAS Support Project”, contract 7/1.1.3H/6.01.2020, associated to the EUs Horizon 2020 ECSEL Joint Undertaking research project SECREDAS.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Catrina, O. (2021). Complexity and Performance of Secure Floating-Point Polynomial Evaluation Protocols. In: Bertino, E., Shulman, H., Waidner, M. (eds) Computer Security – ESORICS 2021. ESORICS 2021. Lecture Notes in Computer Science(), vol 12973. Springer, Cham. https://doi.org/10.1007/978-3-030-88428-4_18
Download citation
DOI: https://doi.org/10.1007/978-3-030-88428-4_18
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-88427-7
Online ISBN: 978-3-030-88428-4
eBook Packages: Computer ScienceComputer Science (R0)