Network calculus is a modern system theory of (networks of) queuing systems. It is relatively young (established around 1990) and provides the theoretical framework for understanding the queuing behavior of networks and its relationship to scheduling disciplines. We distinguish deterministic and stochastic network calculus, where the latter has just been established a few years ago. Especially in the context of stochastic network calculus, a very active research community currently forms, which is on the one hand interested in extending and improving the theoretical basis, but on the other hand applying the theoretical principles to stochastic networks - such as wireless systems – with quite deep insights provided by these applications.
This course intends to introduce PhD students to the foundations and some of the applications of network calculus as established over the last twenty years. The course covers four blocks: An overview of “traditional” queuing theory, deterministic network calculus, effective bandwidth/capacity theory and stochastic network calculus. In each block the emphasis is put on the relation of theory to practical problems in communication systems and networks. Hence, after covering the theory we show the applicability of the theoretical framework in each block to real problems. The course furthermore consists of homework assignments and a final project in which students work on after the course lectures have been completed. The projects are intended to relate to the research area of each PhD student, connecting it with the tools from network calculus for system analysis. Thus, the course targets at enabling PhD students to apply network calculus tools to the analysis of deterministic and stochastic systems, while also providing students with the basics of network calculus and its biggest contributions to system (and network) understanding so far.
Experience from the first round of teaching the course (fall 2014 – spring 2015) shown that about 80 % of the course participants have brought their course projects to journal publications. While this does not need to be the ambition of the student, this is a clear possibility for everyone enrolled in the course.