Today, concurrent programming has become mainstream, dictated by need for ever increasing performance which, having reached the end of Moore's law, can only be achieved by parallelism. Indeed, application software from areas such as medicine or natural sciences heavily relies on parallel hardware infrastructure like (GP)GPU accelerators, FPGAs, and multi- and many-core machines.

 

However, it is also today that we see concurrency faults coming up every day. The simple reason is that writing concurrent programs is difficult. Most programmers "think sequentially" and therefore make mistakes when writing concurrent software. Notorious programming errors include deadlock and violations of atomicity or order of operations, which are mainly caused by the wrong use of synchronization primitives like semaphores or locks. Even worse, the inherent non-deterministic behavior of concurrent software makes bugs difficult to reproduce. Addressing these challenges requires efforts from multiple related disciplines, involving concurrency bug detection, program testing and validation, and programming language design.

 

Another important area of computer science where concurrency naturally arises is that of reactive systems, which maintain an ongoing interaction with their environment. In contrast to sequential systems whose meaning is defined by the results of finite computations, the behavior of reactive systems is mainly determined by concurrent execution, communication, interaction, and mobility of non-terminating processes. Typical examples include operating systems, control systems for production lines, power plants, or vehicles. As many of such systems are safety critical, their development calls for rigorous formal techniques for design, implementation, and validation.

 

The aim of this course is to provide a basic understanding of modeling formalisms for concurrent systems. It will address two basic approaches, which are respectively called the interleaving and the true concurrency approach. The former is based on the idea to reduce the phenomenon of concurrency to well-known concepts, by interpreting parallel behavior as non-deterministic merging of sequential execution. It is represented by various process algebras, which provide a formal apparatus for reasoning about structure and behavior of systems in a compositional way. The true concurrency approach mainly comes in the form of Petri nets, which are well suited for explicitly modeling the concurrent behavior of distributed systems.

Basic knowledge in the following areas is expected:

The written exams are scheduled for Friday, 21.02.2014, 11:30-14:00 at AH 2 and Tuesday, 25.03.2014, 10:00-12:30 at AH 1. Registration is possible via Campus.

Average Grade: 3.1

Failure Rate: 16.7%

Average Grade: 3.325

Failure Rate: 25%

• Lecture
• Exercises

• Luca Aceto, Anna Ingólfsdóttir, Kim Guldstrand Larsen and Jiri Srba: Reactive Systems: Modelling, Specification and Verification. Cambridge University Press, 2007.
• Wolfgang Reisig: Understanding Petri Nets: Modeling Techniques, Analysis Methods, Case Studies. Springer Verlag, 2012.
• Maurice Herlihy and Nir Shavit: The Art of Multiprocessor Programming. Elsevier, 2008.
• Jan Bergstra, Alban Ponse and Scott Smolka (Eds.): Handbook of Process Algebra. Elsevier, 2001.