| Popis: |
This thesis presents an algorithm which contributes to black box penetration testing of embedded systems. It automates the identification process for a selected range of digital communication protocols which are common in the automotive field. Often, the missing knowledge about the communication protocols used on a Printed Circuit Board (PCB) is the only form of protection which prevents an attacker from decoding voltage signals into data for eavesdropping or injection of malicious data. For the protocol identification process, the characteristics for each of the selected protocols are identified first. Then, an algorithm is developed which finds those characteristics automatically using a statistical approach and the search for repeating patterns. The resulting algorithm automatically identifies the functionality and the used protocol type of a data line, based on the found signal characteristics. This automation reduces the effort for finding weaknesses on PCBs and supports the development of exploits for embedded systems. |
