Enabling Practical and Pervasive Content Delivery from Emerging LEO Mega-Constellations

Emerging Low Earth Orbit (LEO) mega-constellations face the challenge of limited bandwidth when providing global Internet services to users. Constructing Content Delivery Networks (CDNs) in LEO mega-constellations is viewed as a feasible solution to address this issue. However, deploying cache satellites, or satellites with CDN servers, for practical and pervasive content delivery is costly and challenging. To overcome this challenge, we consider LEO mega-constellations without inter-satellite links and formulate an integer linear programming problem called User Coverage-aware Cache Satellite Deployment, which aims to maximize the minimum user coverage among all time intervals with a given number of cache satellites. To efficiently solve this problem, a heuristic algorithm called SpaceCache is also proposed. Performance evaluations are conducted based on the Starlink mega-constellation with ground station locations and real-world CDN traces. Compared with benchmark algorithms, we show that the minimum user coverage performance can be improved by up to 41.82% and the cache hit ratio by up to 24.83%.