Towards a Scientific Basis for Digital Forensics

In this talk, I outline work on a scientific basis for digital forensics. Firstly, I outline the failure of forensic science in general and digital forensics in particular to satisfy the requirements for a scientific discipline. I will then demonstrate my architectural security model, which provides a conceptual reasoning space that can be adapted to model forensic investigations. I intend to specify the model using the KAOS methodology to characterise existing digital forensics models, informed by my conceptual theory, to create a practical model with a firm theoretical basis. If time permits, I will also discuss reasoning about credit card fraud in the Event Calculus. We use abductive reasoning to infer possible causes of fraud from known observations, which we intend to apply to forensic investigations. Finally, I hope to discuss the relevance of this work to the so-called ‘science of security’ that is currently in vogue.