Year: 2025
Type: article
Abstract: Background. Monitoring destination reachability, path discovery, and latency analysis is essential for network performance, reliability, and security. The traceroute utility is widely used for path diagnostics, but its text-based output limits the ability to perform complex analysis of network capabilities. ICMP traffic filtering, rate limiting, and load balancing cause missing hops in route displays, non-monotonic delays, and artifacts that are difficult to detect in tabular form. Existing visualization methods rely on geolocation databases and emphasize geographical aspects rather than network topology. Therefore, developing a geolocation-independent graph-based visualization tool for traceroute is relevant. Materials and Methods. The proposed graph-based approach converts single and multiple traceroute results into a unified weighted directed graph, where vertices represent IP-defined nodes. Unknown nodes are handled using two strategies: explicit labeling or skipping with a direct connection between known nodes. The following edge-weighting criteria are introduced: occurrence frequency, used to identify critical links, and average round-trip time (RTT), used to detect latency segments. Graph visualization is performed using the Fruchterman-Reingold algorithm. The system is implemented in Java with JavaFX, Spring Core, and GraphStream, supports both archived results import and real-time tracing via a multithreaded producer-consumer model, ensuring graph integrity. Results and Discussion. Combining multiple traceroutes into a single graph enables quick identification of shared segments and critical nodes. Systematic RTT non-monotonicity, likely caused by rate limiting, supports using absolute rather than incremental RTT values. In the experiments conducted, the multithreaded mode accelerates graph construction fourfold with seven threads, but it also increases the number of timeouts. Using five threads achieves an optimal balance between speed and data quality. Conclusion. The proposed approach improves network analysis efficiency by constructing a geolocation-independent graph for diagnostics. The developed system reduces cognitive load compared to tabular logs analysis and speeds up detection of bottlenecks, common segments, critical nodes, and routing anomalies. Keywords : traceroute, graph visualization, network topology, network monitoring, weighted directed graphs
DOI: 10.30970/eli.32.1
Institution: Lviv Polytechnic National University
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7124334635