Research Article
Research of CPA Attack Methods Based on Ant Colony Algorithm
@INPROCEEDINGS{10.1007/978-3-030-90019-9_14, author={Xiaoyi Duan and You Li and Jianmin Tong and Xiuying Li and Siman He and Peishu Zhang}, title={Research of CPA Attack Methods Based on Ant Colony Algorithm}, proceedings={Security and Privacy in Communication Networks. 17th EAI International Conference, SecureComm 2021, Virtual Event, September 6--9, 2021, Proceedings, Part I}, proceedings_a={SECURECOMM}, year={2021}, month={11}, keywords={Power analysis attack CPA (Correlation Power Analysis) AES algorithm Ant Colony Algorithm}, doi={10.1007/978-3-030-90019-9_14} }- Xiaoyi Duan
You Li
Jianmin Tong
Xiuying Li
Siman He
Peishu Zhang
Year: 2021
Research of CPA Attack Methods Based on Ant Colony Algorithm
SECURECOMM
Springer
DOI: 10.1007/978-3-030-90019-9_14
Abstract
The Power analysis attack is an effective method of attacking encryption devices for leakage of side-channel information. CPA (Correlation Power Analysis) is a common method. The traditional method of Power Analysis Attack, which is only one-byte key, is analyzed in one attack and repeats multiple operations to obtain the whole secret key. In this way, a successful attack needs more power curves. In this paper, a new attack method is proposed to select the optimal secret key group through the Ant Colony Algorithm and attack all the bytes of the secret key simultaneously. It can greatly eliminate the influence of the channel noise and improve the efficiency of the attack. To prove the effectiveness of this new method, the AES algorithm as an example is implemented on the MEGA16 microcontroller. The power consumption curve of the AES algorithm with a fixed secret key and random plaintext is collected, and the power consumption is analyzed separately by the original method and the new method. As a result, the success rate of the original method is only 10.981% when using 4000 power curves; however, the new one is up to 100%, which is increased by 89.019%. When the power curves do not exceed 3000, the success rate of the original method is zero. However, the success rate of the new method can reach 34.375% even if only 1500 power curves are used. The new method is more effective than the original one. Being affected by parameters, the attack time of the new method is not consistent but much less than the original method.