Konzeption eines zentralisierten Echtzeitüberwachungssystems für verteilte Anwendungen: Ein skalierbarer und fehlertoleranterAnsatz

Abstract

This Master’s thesis addresses the design, analysis, and optimization of a centralized
real-time monitoring system for distributed applications, exemplified by the interactive
exhibition Talent Space. The focus lies on ensuring stability, scalability, and efficiency
in a highly connected environment where numerous stations must be coordinated and
monitored in real time.
Based on the existing system, benchmark scenarios were developed that systematically varied frequency, packet size, and the number of clients. Core metrics such as
round-trip time, throughput, error rates, and connection stability served as the foundation for a comprehensive evaluation.
The experimental results confirm the robustness of the current architecture. Even
under deliberately exaggerated stress tests with high message rates, large payloads,
and many concurrent clients, the system remained stable and reliable. Latency and
throughput stayed within the expected ranges without noticeable packet loss or critical
failures. This demonstrates that the existing setup already provides a solid foundation
for productive use and further expansion.
At the same time, the work identifies clear opportunities for the next development
stage: more precise standardization of status and error signals, extended observability
through well-defined SLI/SLO indicators, and optional efficiency gains through more
compact data formats or high-performance frameworks where appropriate. These measures do not alter the proven baseline system but strengthen its long-term scalability
and reliability.
Overall, this thesis validates the suitability of the existing architecture and outlines
a concrete roadmap for its evolution in Phase 2 of the Talent Space project.

Date of Award	2025
Original language	German (Austria)
Supervisor	Christoph Schaffer (Supervisor)