MA: Autocalibration of Resources in Co-Simulated Virtual Machine Networks
The development and research of heterogeneous distributed systems such as the Internet of Things (IoT) is dependent on accurate simulations and emulations of such networks. For this reason, our Distributed Systems Engineering Lab (diselab) has recevently developed testbed frameworks [1,2] that support network simulation/emulation and infrastructure simulators such as SUMO .
To ensure a certain degree of realism with regards to the compute power of emulated virtual machines (VMs), resources allocated to these VMs needs to be calibrated. As our frameworks aims to automatically set up and maintain emulated device networks, this needs to be done automatically. The goal of this thesis is, therefore, the investigation of resource management possibilities in different virtualization technologies as well as the implementation and evaluation of an autocalibrator for this purpose.
Interested students can develop a personal topic with regards to the following areas:
- Automated instruction profiling of applications in a black-box scenario
- Realistic IoT device emulation
- Adaptive resource management of virtual machines
A high degree of autonomy and some experience with scientific writing is expected.
If this sounds interesting to you, please send us an email containing some background information on your interests, a CV, and a current mark sheet so we can quickly identify a fitting thesis topic together.
 ieeexplore.ieee.org/abstract/document/8759362  and github.com/osmhpi/cohydra 
 dl.acm.org/doi/abs/10.1145/3368235.3368832  and github.com/citlab/hector_iot_testing_framework 
 www.eclipse.org/sumo/