Network Calculus - What It Can Do For You and Where It Needs Your Help

摘要

In this talk, the theory of network calculus is discussed from the point of view: (1) how powerful it is in applications as well as (2) what fundamental challenges it still faces before it can be considered a similarly mature theory as the well-known classical queueing theory. To this end, we briefly review the basic concepts and mathematical foundations of network calculus, before we then show its applicability for the example of wireless sensor networks. The latter work culminated in the so-called sensor network calculus which is a customized and adequately extended set of methods based on the fundamental network calculus results enriched by the requirements stemming from the specific nature of wireless sensor networks. The lessons learned from this exercise are: on the one hand, network calculus provides already a rich and useful set of methods, but on the other hand, substantial work was still necessary to apply it to the particular context of wireless sensor networks (and to some degree still remains). Hence, in the second part of this talk, we outline some of the fundamental open issues with respect to the foundations of network calculus. We discuss current efforts to address these, although some of them are still completely open and thus constitute good opportunities for future research.