Virtualized Computing and Communication Service Provision in Large-Scale Software Defined Satellite Networks

Enhanced on-board computing and communication hardware have facilitated the integration of resource management software and virtualization technologies in space, transforming satellite networks (SNs) from mission-specific support to facilitators of diverse mobile applications via microservices. How to efficiently manage and allocate virtualized services within software-defined SNs is critical. This study investigates the provisioning of virtual computing and communication services for mobile applications in large-scale software-defined SNs. Using a service-aware temporal graph, we model the problem of minimizing latency and its constraints related to meeting stringent service requirements. Traditional optimization methods encounter impractical running times, and the k-shortest pathbased algorithm may yield suboptimal solutions. To address the problem effectively for real-world implementation, we propose an optimized service-aware service provision scheme. Validation through simulations of a software-defined Starlink constellation underscores its practical effectiveness.