The experiment involved a network of four computers, four users, and eight distinct services. While only one physical computer currently in orbit ran the software, the remaining nodes functioned as simulations to assess how the system handles task distribution based on fluctuating compute availability and communication links. The test successfully generated operational plans directly from space, proving that the software can manage resource scheduling without constant ground-based intervention.
Founded in 2024 by former Google and SpaceX engineer Afullo, Satlyt aims to solve the persistent bottleneck of limited bandwidth and intermittent connectivity in space. By enabling satellites to process data—such as identifying wildfires or analyzing sensor inputs—onboard, the company hopes to reduce reliance on the slow process of downlinking raw information to Earth. Afullo noted that while the team encountered a technical snag regarding how the software time-referenced data delays, the core mission of executing tasks and returning outputs was achieved.
Significant hurdles remain before orbital computing becomes a standard utility. The team must still refine secure data transfers, permissions between different operators, and recovery protocols for when nodes drop offline. Satlyt plans to move beyond low Earth orbit testing, with the ultimate goal of creating a platform where developers can deploy applications for space without needing to rebuild underlying infrastructure for every individual mission.

Comments (0)
No comments yet. Be the first!