With the rapid growth of low earth orbit (LEO) satellites, enabling LEO AI inference becomes a fast-increasing trend. However, due to resource heterogeneity, scheduling complexity, and fast movement, how to decide the place of executing each AI inference task is nontrivial in LEO systems. In this paper, we propose LEOEdge, an edge-assitsed AI inference system for LEO satellites. We first introduce the adaptive modeling technologies that automatically generate the model for each satellite according to its computation resources. We then propose a layered scheduling optimization scheme to schedule the AI inference task in a distributed manner. LEOEdge also designs a seamless data transmission scheme to avoid transmission failure due to the LEO satellite movement. We conduct a series of simulation tests to validate the performance of the proposed LEOEdge, in terms of the neural network searching efficiency, average time execution latency, and delivery latency.
The cellular uplink is increasingly important for modern applications but poses significant challenges to congestion control algorithms (CCAs), due to the highly dynamic conditions and complex operations. Existing CCA paradigms, mainly for downloading data, are unable to react promptly to uplink fluctuations because they rely on ACKs to indicate congestion, introducing inevitable control loop latency. Furthermore, the cost associated with modifications to both the sender and receiver cannot be overlooked. We observe a fundamental opportunity in uplink scenarios: the sender's proximity to the wireless-side allows for fast and accurate control without relying on ACK-based feedback, and requires no modifications to either the receiver-side or the hardware layer. In this paper, we propose Aether, mainly consisting of: (1) a monitor deployed on the driver-layer that captures uplink dynamics, by leveraging our key finding that driver buffer dynamics serve as an indicator of uplink congestion; (2) a bottleneck detector that activates Aether when the uplink becomes the bottleneck; (3) a rate controller that steers transmission toward uplink capacity. We implement Aether on commercial devices and develop a trace-driven emulator for testing Aether. Extensive evaluations demonstrate that Aether outperforms state-of-the-art schemes, improving average throughput by 9.28% to 234.11% while consistently reducing latency by 5.34% to 56.94%.