The quest for the correctness of computer systems: The program verification approach

Summary form only given, as follows. The complete presentation was not made available for publication as part of the conference proceedings. Parallel computer systems have become a critical part of the infrastructure of our society during the last decade. This is due to the emergence of modern platforms such as multicores and cloud technology. Nowadays, multicores are available on mobile phones, laptops, embedded systems, servers, and numerous other devices that are integral parts of our daily life. Furthermore, cloud technology is increasingly based on efficient storage and exchange of data across large-scale networks. Such networks include enterprise services (e.g., Google Cloud or Microsoft Azure), banking systems, online markets (e.g., Amazon), and social networks (e.g., Facebook). The ubiquity of parallel systems means that their efficiency and correctness are vital goals both from the security and economical points of view. However, the complexity of these systems implies that designers need to consider new challenges for which we currently lack satisfactory solutions. In this presentation, I will describe a particular approach to the analysis of the correctness of computer systems in general and parallel systems in particular, namely that of program verification. I will introduce the main concepts behind program verification, and the new challenges that the program verification community is facing due to the emergence of the system paradigms mentioned above. While the verification approach is highly technical both from the mathematical and the algorithmic point of views, my lecture will be self-content and should be understandable even to non-experts.