New Techniques to Improve Network Security

With current technologies it is practically impossible to claim that a distributed application is safe from potential malicious attacks. Vulnerabilities may lay at several levels (criptographic weaknesses, protocol design flaws, coding bugs both in the application and in the host operating system itself, to name a few) and can be extremely hard to find. Moreover, sometimes an attacker does not even need to find a software vulnerability, as authentication credentials might simply “leak” ouside from the network for several reasons. Luckily, literature proposes several approaches that can contain these problems and enforce security, but the applicability of these techniques is often greatly limited due to the high level of expertise required, or simply because of the cost of the required specialized hardware. Aim of this thesis is to focus on two security enforcment techniques, namely formal methods and data analysis, and to present some improvements to the state of the art enabling to reduce both the required expertise and the necessity of specialized hardware.