Müller, T. (2019):
Evaluation of ZMTP in Constrained Environments
As the prevalence of constrained environments progressed over the past few years, the demand for distributed computing approaches that could potentially remedy the limited processing power of such networks emerged. Since constrained environments, such as the Internet of Things or Wireless Sensor Networks, are typically composed of nodes that exhibit severe restrictions on computational power, memory capacity and energy supply, resource scarcity is a fundamental characteristic of constrained environments. This thesis therefore introduces the ZeroMQ Message Transport Protocol (ZMTP), which enables distributed computing through peer-to-peer networking and is based on TCP. In order to assess the scalability of ZMTP in constrained environments, a protocol port tailored to the RIOT operating system is used in the course of this thesis. An experimental approach comprising a series of experiments with runtime measurements is conducted by means of two distributed applications that constitute representative use cases. The consequent analysis and evaluation of the measurement results reveal an undesired scaling behavior of ZMTP, that is an insufficient performance improvement mainly caused by the vast communication overhead, which limits the speedup through parallelization within both use cases. As a result, the protocol exhibits a rather weak scalability in general, so with respect to these two use cases, ZMTP is considered not suitable for the purpose of efficiently realizing distributed computing in constrained environments.