A Study of Template Clustering in the Side Channel Template Analysis
Tae-Sung Kim · 2018
The side channel analysis reveals the secret key by exploiting the side information that occurs during the execution of the encryption algorithm. Template analysis, one of the side channel analysis, is to learn the characteristics of the device before attack in relation to the side information. The characteristics of a device are represented by templates and this is the result of learning. The number of templates is determined by a criterion. For example, if the byte length is a reference, 256 templates are created. If one template attack fails, the procedure of changing the criteria and creating a new template should be repeated. In this paper, we propose clustering method of templates with similarity. Clustering prevents repetitive template creation and attack. It is expected that the template analysis using the proposed method will have high accuracy and efficiency.