Detailed Information

Cited 0 time in webofscience Cited 0 time in scopus
Metadata Downloads

Correcting errors in private keys obtained from cold boot attacks

Authors
Lee, H.T.Kim, H.Baek, Y.-J.Cheon, J.H.
Issue Date
Dec-2012
Keywords
Cold Boot Attack; Discrete Logarithm; RSA; Side Channel Attack
Citation
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v.7259 LNCS, pp 74 - 87
Pages
14
Journal Title
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume
7259 LNCS
Start Page
74
End Page
87
URI
https://scholarworks.bwise.kr/cau/handle/2019.sw.cau/64930
DOI
10.1007/978-3-642-31912-9_6
ISSN
0302-9743
1611-3349
Abstract
Based on the cold boot attack technique, this paper proposes a new algorithm to obtain the private key of the discrete logarithm (DL) based cryptosystems and the standard RSA from its erroneous value. The proposed algorithm achieves almost the square root complexity of search space size. More precisely, the private key of the DL based system with 160-bit key size can be recovered in 2 43.24 exponentiations while the complexity of the exhaustive search is 2 71.95 exponentiations if the error rate is given by 10%. In case of the standard RSA with 1024-bit key size, our algorithm can recover the private key with 2 49.08 exponentiations if the error rate is given by 1%. Compared with the efficiency of some algorithms [7,6] to recover the private key in RSA using Chinese Remainder Theorem, the recoverable error rate of our algorithm is quite small. However, our algorithm requires only partial information of the private key d while other algorithms require additional information such as partial information of factors of the RSA modulus N. The proposed algorithm can also be used for breaking countermeasure of differential power analysis attack. In the standard RSA, one uses the randomized exponent instead of the decryption exponent d with the random value r. When the size of a random value r is 26-bit, it can be shown that the randomized exponent can be recovered with 2 49.30 exponentiations if the error rate is 1%. Finally, we also consider the breaking countermeasure that splits the decryption exponent d into d 1 and d 2 of same size.
Files in This Item
There are no files associated with this item.
Appears in
Collections
College of Software > School of Computer Science and Engineering > 1. Journal Articles

qrcode

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.

Related Researcher

Researcher Lee, Hyungtae photo

Lee, Hyungtae
소프트웨어대학 (소프트웨어학부)
Read more

Altmetrics

Total Views & Downloads

BROWSE