Master-Seminar Connected and Mobile Systems

Low Earth Orbit (LEO) satellite networks require frequent handovers because satellites move rapidly relative to users on the ground. Conventional handover mechanisms based on radio measurements may cause long access delays, signaling overhead, collisions between simultaneous handover requests, or short connection durations. Predictive handover mechanisms and coordinated connections to multiple satellites offer alternative approaches for improving service continuity.

Your task is to firstly provide an overview of handover mechanisms in LEO satellite networks. Then discuss in detail at least three handover approaches, including conventional measurement-based handover, predictive or learning-based handover, and multi-satellite connectivity. Finally, compare these approaches based on characteristics such as access delay, handover frequency, signaling overhead, etc. and compatibility with current satellite network architectures. Conclude by discussing the merits, limitations, and suitable application scenarios of each approach.

• Handover Protocol Learning for LEO Satellite Networks: Access Delay and Collision Minimization
• Distributed Massive MIMO for LEO Satellite Networks

Modern kernel-bypass frameworks such as DPDK, netmap, and AF_XDP promise multi-Gbps to multi-hundred-Gbps packet processing on commodity hardware, far beyond what the traditional Linux networking stack can deliver. This performance is not the result of a single trick, but of a handful of recurring mechanisms — e.g., kernel bypass, batched I/O, polling instead of interrupts, huge pages, zero-copy buffer management, and pinned/lock-free memory pools — that keep reappearing across different frameworks, often reinvented or refined over time. Many of these ideas did not originate with the framework that popularized them; for instance, netmap appears to be an early publicator of several mechanisms that were picked up by later, more widely deployed systems. Your task is to first identify and explain the key mechanisms underlying today's high-performance packet processing frameworks, tracing each one back to the publication that first introduced it and explaining in depth how the mechanism works and why it improves performance. Next, you will give an overview of the major frameworks (e.g., DPDK, netmap, AF_XDP) and, for each, qualitatively describe which of the identified mechanisms it uses and how it adapts or extends them compared to the original proposal. Finally, summarize your findings in a structured overview that maps mechanisms to their first publicators and frameworks to their adopted mechanisms. (Following publications can serve as a starting point for the literature survey.)

• netmap: A Novel Framework for Fast Packet I/O
• The eXpress Data Path: Fast Programmable Packet Processing in the Operating System Kernel
• User Space Network Drivers

The Canadian Traveller Problem describes the challenge of finding a shortest path from source to destination in a road network, where we do not know the current status of roads and junctions. The name references a real world problem in Canada in the 80s and 90s, where people did not know, if roads or junctions are blocked due to snow. Modern navigation systems include almost live information about roads, which is sufficient to find the shortest path for everyday travel. But Emergency Vehicles (EMS) need the most optimal path to their destination, where info from navigations systems might not be sufficient. In Vehicle-to-Everything (V2X) networks, so called Road-Side-Units (RSUs) may be deployed at intersections and could inform an EMS-vehicle about the current situation at the intersection. The EMS could then leverage this information to calculate its shortest path to the accident site.

Your task is to research and compare methods and heuristics that efficiently calculate the shortest/fastest path in a road network with limited information. Especially consider methods, that could be used by EMS on their way to an accident, which get some road network insights through a V2X network.

• V2X-assisted emergency vehicle transit in VANETs
• Shortest paths without a map
• The Canadian Traveller Problem