Paper 2019/954

CPA on Hardware Implementation of COLM Authenticated Cipher and Protect it with DOM Masking Scheme

Mohsen Jahanbani, Zeinolabedin Norouzi, and Nasour Bagheri

Abstract

Authenticated encryption schemes provide both confidentiality and integrity services, simultaneously. Correlation power analysis (CPA) can be a thread for authenticated ciphers, like all physical implementations of any cryptographic system. In this paper, for the first time, a three-steps CPA attack against COLM, one of the winners of CAESAR, is presented to indicate its vulnerability. For this purpose, in this research paper, this authenticated encryption scheme is implemented on the FPGA of the SAKURA-G board and, by measuring and collecting 1,800 power traces, a successful CPA attack with zero value power model has been mounted on it. In addition, a protected hardware architecture for the COLM is proposed to make this design secure against first-order CPA attacks. To this end, a domain-oriented masking (DOM) scheme with two inputs/outputs share is used to protect the COLM. To verify the security of these countermeasures, we mounted a first and second-order CPA attack and a non-specified t-test on the protected COLM.

Metadata
Available format(s)
PDF
Category
Implementation
Publication info
Preprint. MINOR revision.
Keywords
Authenticated CipherCOLMCPADOMMasking.
Contact author(s)
na bagheri @ gmail com
mo jahanbani2009 @ gmail com
History
2019-08-21: received
Short URL
https://ia.cr/2019/954
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2019/954,
      author = {Mohsen Jahanbani and Zeinolabedin Norouzi and Nasour Bagheri},
      title = {CPA on Hardware Implementation of COLM Authenticated Cipher and Protect it with DOM Masking Scheme},
      howpublished = {Cryptology ePrint Archive, Paper 2019/954},
      year = {2019},
      note = {\url{https://eprint.iacr.org/2019/954}},
      url = {https://eprint.iacr.org/2019/954}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.