An availability modeling method of LEO satellite network considering dynamic effects of failure

Application availability modeling for low-Earth orbit satellite networks (LEOSNs) has become a significant research focus. However, existing methods lack support for capturing the effects of failure on applications. Due to the dynamic characteristics of LEOSN (such as time-varying and handover), these effects are amplified and coupled. To address this challenge, we conduct failure modes and their corresponding mechanisms from four perspectives to explore the amplified effect in LEOSN. We clarify that each mechanism makes the amplified effects more tractable to model. Based on failure modes and mechanisms, we propose an availability evaluation model, NEMFC, in which a specific evolution rule represents each mechanism. Compared to existing methods, the effects of failure reduce estimated availability by 1.859%–6.116%. Sensitivity analysis shows that failure parameters lead to availability variations of 3.656%–10.382%, significantly exceeding other parameters and underscoring the necessity of modeling these failure modes. Additionally, our experiments show that the same configuration effect has different impacts on the availability of constellations of different scales.