This master thesis addresses various aspects of secure software development for embedded systems where the threat model consists of malicious input data that come from a communication network or sensors and exploit an error in software implementation. The emphasis is on the analysis of security implications of various system architecture decisions.
In this regard, the recent development in the understanding of secure software development from the theoretical computer science point of view is presented. Computation and language theory can determine what kind of systems is possible to implement securely in terms of the malicious input data.
The presented aspects of secure software development are demonstrated in designing and implementing of a parser for a simple message format of a simple communication protocol.
|
