Delay Performance Analysis of Integrated Communication and Remote Sensing Satellite Networks Under EDF Scheduling

With the widespread deployment of integrated communication and remote sensing satellite networks, we can acquire remote sensing data and achieve near real-time transmission and processing of Earth observation information. However, in such network architectures, satellites must simultaneously handle multiple types of services, posing significant challenges for Quality of Service (QoS) guarantee. Given the limited ground stations and coverage constraints, the integrated communication and remote sensing satellite network naturally exhibits a sink-tree topology. We therefore model it as such to analyze delay performance under EDF scheduling, which is optimal for tree-structured networks. First, we establish a dual-layer network system model comprising Low Earth Orbit (LEO) and Geostationary Earth Orbit (GEO) satellites. Then we derive the end-to-end service curves for Earliest Deadline First (EDF) scheduling under sink-tree topology using network calculus theory. By introducing martingale theory, we develop a novel analytical framework that provides tight upper bounds for end-to-end delay violation probabilities. Simulation results validate our analytical approach, demonstrating that the proposed framework can accurately predict delay performance under various network conditions. This research provides valuable guidance for deadline allocation and resource scheduling in integrated communication and remote sensing satellite networks.