Ibraheem Frieslaar is currently finalising and submitting his PhD thesis at the end of October 2017. This has been concluded in The Department of Computer Science at Rhodes University. He offers a unique skill set as he has specialities in digital signal processing in the cyber security domain, particularly the electromagnetic area where electromagnetic information is utilized to locate vulnerabilities within security systems.
Talk: Using Electromagnetic Emissions to Intercept AES-128 Cryptographic Keys from a Raspberry
The research is the first of its kind as it investigated the utilisation of multi-threading as a software based countermeasure to mitigate Side Channel Analysis (SCA) attacks. Multi-threading is mainly used to increase performance and runtime execution of an algorithm or program. This investigation is novel, as there has not been a software based countermeasure relying on multi-threading to mitigate SCA attacks. Furthermore, the research has been tested on a more fully featured system in the form of the popular Raspberry Pi.
This research has demonstrated that existing Electromagnetic (EM) analysis techniques and approaches used against embedded hardware system with regards to SCA attacks can be utilised against a fully featured device such as the Raspberry Pi. The most important and novel of these contributions is the introduction of a multi-threading software based countermeasure to mitigate SCA attacks on a Raspberry Pi. These multi-threads are comprised of various mathematical operations to generate EM noise which obfuscates the execution of AES-128 algorithm.
This research implemented a novel EM noise generator known as FRIES to assist in obfuscating data captured in the EM field. The noise generator comprises of hiding the AES-128 algorithm within the EM noise generated by the Secure Hash Algorithm from the libcrypto++ and OpenSSL library. While the hash function were generating EM noise, the cryptographic implementation of AES-128 algorithm could not be visibly seen. Furthermore, the intentional leakage was captured and the results indicated that no secret information was recovered while the proposed noise generator was in place.
This novel countermeasure is not limited to cryptography alone, as it can be applied to other operations such as a login process of a device. Since this is a software based solution, developers can integrate this solution either within or alongside their own applications. Their application can have the ability to execute the noise generator at program startup or at critical segments of the algorithm.
The presentation will focus on the following:
- The basics of SCA analysis and its relevance.
- The impact of stolen encryption keys.
- How existing SCA technique can be utilised on new hardware such the Raspberry Pi.
- New approaches this research followed to recover AES-128 encryption keys from the Raspberry Pi.
- The novel software based countermeasure that this research developed to mitigate SCA attacks.
- Integration of the proposed countermeasure with other applications.