Measurements Analytics In SRv6 Segment

Abstract

The increasing reliance on latency-sensitive applications such as video conferencing, online gaming, and delay-based geolocation has made it crucial to ensure consistent and minimal latency in international backbone networks. This thesis focuses on the development of a tool to detect and analyze latency imbalances and other anomalies in SRv6 networks using network measurements. The primary sources of latency imbalances are identified as architectural issues in network design and hardware programming errors in routers. The research begins with a comprehensive study of the transport domain and the IPv6 protocol, detailing the structure of packets and the routing algorithms employed including IS-IS and Flex Algo. It will introduce as well the Dijkstra algorithm’s and ECMP for load balancing. An emulation tool is implemented to simulate network traffic and evaluate latency across different paths. The analysis reveals that improperly configured ECMP paths can lead to significant variations in latency, adversely affecting network performance. A novel algorithm for detecting hardware issues related to incorrect network interface pro- gramming is proposed. This algorithm helps identify routers that cause increased latency due to misprogrammed paths. The tool’s effectiveness is validated through extensive testing on emulated networks, demonstrating its capability to detect and address both architectural and hardware-related latency issues. Finally, a graphical interface is developed to provide a user-friendly platform for network operators to visualize and manage the detected anomalies. The interface integrates simulation, topology analysis, and hardware issue detection functionalities, enhancing the overall utility of the tool. This work contributes to improving the reliability and performance of SRv6 networks by providing a robust solution for latency imbalance detection and mitigation.