Case studies in symbolic model checking

The need to formally verify hardware and software systems before they are deployed the real world has been recognized for several decades now. This is especially true of concurrent systems that are even more difficult to debug than sequential systems. For example, many of the protocols that get employed in real-life systems often look deceptively simple at first glance, and yet often contain hidden errors. In many cases, these errors cannot be revealed through simulation alone. Tools that are powerful enough to verify real-life systems of significant sizes have, however, only recently become available, thanks to developments such as efficient Boolean reasoning methods. As a result, there is a real opportunity amongst practitioners of formal methods to apply these tools to real-life examples and to teach tomorrow's designers?today's students?the use of these tools for solving practically significant problems. This paper is about our efforts in this regard. Among formal verification tools that can verify concurrent systems, model checking tools have become quite popular. There are several reasons for this. First, their ability to carry out proofs with relatively very little human intervention makes it possible for designers to automatically carry out many of the proofs, thus freeing them up for more creative tasks. Second, most concurrent hardware and software systems are one-ofs, and hence it is very difficult to recoup human effort put into one project for use in another.