Simulation and Software Technology
Area of research:
Diploma & Master Thesis
With the increasing software requirements in embedded systems, a single CPU might not have sufficient resources to execute all the required tasks. Additionally, in safety critical applications that can be found, for instance, in the aerospace domain, tough environment conditions like radiation necessitates the need for back-up CPUs. Also, it is an advantage if the system can be reconfigured in case of failure or change in mission phases. This requires the system to be a distributed system. The ScOSA Flight Experiment project is one example of a distributed system. The Scalable On-board Computing for Space Avionics (ScOSA) activity aims to develop a distributed, heterogeneous and reliable on-board computer (OBC), offering more computing performance than current OBCs and the option to reconfigure the entire OBC whenever needed.
In distributed systems it becomes important to analyze the data exchange over the network on different CPUs. Timing analysis helps to evaluate end-to-end delay between the sender and receiver and to verify if the expected message reached the desired CPUs or was lost somewhere in the network. A time log of all the data packets in a network can help in providing a proof of concept for a distributed system algorithm.
For instance, it can support the
analysis of correct sequence of data packets verification of the correct error reporting in event of failure acknowledgment of service request
The goal of this project is to evaluate the timing analysis of data packets for different distributed systems. These systems will differ on the basis of various network topologies and will evaluate a distributed system communication protocol.
The result of the analysis will contribute to:
find the favorable or unfavorable network topologies for a distributed system validate the developed network protocols before they can be deployed on an actual system reduce overall development costs as testing on actual systems can be more expensive than a simulated ones
The distributed system can be tested with Software in Loop (SiL) testing. SIL can be defined as a testing procedure in which the developed algorithms/programs are executed on a simulated model of an actual system for longer/infinite periods. This test strategy helps to improve the algorithm. SiL can be carried out on a model of the system that can be simulated in simulation tools like OMNeT++.
OMNeT++ is a collection of C++ simulation library and framework primarily aimed for simulating networks. It has been extensively used in the past for resource modeling, model internet protocols, switched LANs and more.
The system will consist of five nodes which will be simulated on OMNet++. These nodes will be connected via Ethernet or SpaceWire.
Your tasks can be briefly summarized as:
research/Analyze different strategies and topologies to model a distributed system model these strategies with OMNet++ obtain the timing results for these strategies by using SIL compare the results of the study, concluding the advantages and limitations of each prepare your thesis