A Novel Points of Interest Selection Method For SVM-based Profiled Attacks

Abstract—Currently, one of the most powerful side channel attacks (SCA) is profiled attack. Machine learning algorithms, for example support vector machine (SVM), are currently used to improve the effectiveness of the attack. One issue of using SVM-based profiled attack is extracting points of interest (POIs), or features from power traces. Our work proposes a novel method for POIs selection of power traces based on the combining variational mode decomposition (VMD) and Gram-Schmidt orthogonalization (GSO). VMD is used to decompose the power traces into sub-signals (modes) and POIs selection process based on GSO is conducted on these sub-signals. As a result, the selected POIs are used for SVM classifier to conduct profiled attack. This attack method outperforms other profiled attacks in the same attack scenario. Experiments were performed on a trace data set collected from the Atmega8515 smart card with AES-128 run on the Sakura-G/W side channel evaluation board and the DPA Contest v4 dataset to verify the effectiveness of our method in reducing number of power traces for the attacks, especially with noisy power traces.Tóm tắt—Hiện nay, tấn công mẫu được xem là một trong những tấn công kênh kề (SCA) mạnh. Các thuật toán học máy, ví dụ như máy vector hỗ trợ (SVM), thường được sử dụng để nâng cao hiệu quả của tấn công mẫu. Một thách thức đối với tấn công mẫu sử dụng SVM là cần phải tìm được các điểm thích hợp (POI) hay các đặc trưng từ vết điện năng tiêu thụ. Công trình nghiên cứu này đề xuất một phương pháp mới đề tìm POI của vết điện năng tiêu thụ bằng cách kết hợp kỹ thuật phân tích mode biến phân (VMD) và quá trình trực giao hóa Gram-Schmidt (GSO). Trong đó, VMD được sử dụng để phân tách vết điện năng tiêu thụ thành các tín hiệu con còn gọi là VMD mode và việc lựa chọn POIs trên VMD mode này được thực hiện dựa trên quá trình GSO. Dựa trên phương pháp lựa chọn POIs này, chúng tôi đề xuất phương pháp tấn công mẫu sử dụng SVM có hiệu quả tốt hơn các tấn công mẫu khác ở cùng kịch bản tấn công. Các thí nghiệm tấn công được thực hiện trên tập dữ liệu được thu thập từ thẻ thông minh Atmega8515 cài đặt AES-128 chạy trên nền tảng thiết bị tấn công kênh kề Sakura-G/W và tập dữ liệu DPA Contest v4, để chứng minh tính hiệu quả của phương pháp của chúng tôi, trong việc giảm số lượng vết điện năng tiêu thụ cần cho cuộc tấn công, đặc biệt trong trường hợp các điện năng tiêu thụ có nhiễu.

Pay Attention to Raw Traces: A Deep Learning Architecture for End-to-End Profiling Attacks

With the renaissance of deep learning, the side-channel community also notices the potential of this technology, which is highly related to the profiling attacks in the side-channel context. Many papers have recently investigated the abilities of deep learning in profiling traces. Some of them also aim at the countermeasures (e.g., masking) simultaneously. Nevertheless, so far, all of these papers work with an (implicit) assumption that the number of time samples in raw traces can be reduced before the profiling, i.e., the position of points of interest (PoIs) can be manually located. This is arguably the most challenging part of a practical black-box analysis targeting an implementation protected by masking. Therefore, we argue that to fully utilize the potential of deep learning and get rid of any manual intervention, the end-to-end profiling directly mapping raw traces to target intermediate values is demanded.In this paper, we propose a neural network architecture that consists of encoders, attention mechanisms and a classifier, to conduct the end-to-end profiling. The networks built by our architecture could directly classify the traces that contain a large number of time samples (i.e., raw traces without manual feature extraction) while whose underlying implementation is protected by masking. We validate our networks on several public datasets, i.e., DPA contest v4 and ASCAD, where over 100,000 time samples are directly used in profiling. To our best knowledge, we are the first that successfully carry out end-to-end profiling attacks. The results on the datasets indicate that our networks could get rid of the tricky manual feature extraction. Moreover, our networks perform even systematically better (w.r.t. the number of traces in attacks) than those trained on the reduced traces. These validations imply our approach is not only a first but also a concrete step towards end-to-end profiling attacks in the side-channel context.

EFFICIENT CNN-BASED PROFILED SIDE CHANNEL ATTACKS

Profiled side-channel attacks are now considered as powerful forms of attacks used to break the security of cryptographic devices. A recent line of research has investigated a new profiled attack based on deep learning and many of them have used convolution neural network (CNN) as deep learning architecture for the attack. The effectiveness of the attack is greatly influenced by the CNN architecture. However, the CNN architecture used for current profiled attacks have often been based on image recognition fields, and choosing the right CNN architectures and parameters for adaption to profiled attacks is still challenging. In this paper, we propose an efficient profiled attack for on unprotected and masking-protected cryptographic devices based on two CNN architectures, called CNNn, CNNd respectively. Both of CNN architecture parameters proposed in this paper are based on the property of points of interest on the power trace and further determined by the Grey Wolf Optimization (GWO) algorithm. To verify the proposed attacks, experiments were performed on a trace set collected from an Atmega8515 smart card when it performs AES-128 encryption, a DPA contest v4 dataset and the ASCAD public dataset