Smaller is Better: Generative Models Can Power Short Video Preloading

TL;DR

PromptPream leverages local computation and generative models to prefetch short videos efficiently, significantly reducing bandwidth waste and playback stalls while enhancing user experience.

Contribution

This paper introduces PromptPream, a novel preloading paradigm that uses semantic prompts and generative models to optimize bandwidth and reduce stalls in short video streaming.

Findings

Reduces stalls and bandwidth waste by over 31%.

Improves Quality of Experience (QoE) by 45%.

Employs novel prompt inversion and scheduling algorithms.

Abstract

Preloading is widely used in short video platforms to minimize playback stalls by downloading future content in advance. However, existing strategies face a tradeoff. Aggressive preloading reduces stalls but wastes bandwidth, while conservative strategies save data but increase the risk of playback stalls. This paper presents PromptPream, a computation powered preloading paradigm that breaks this tradeoff by using local computation to reduce bandwidth demand. Instead of transmitting pixel level video chunks, PromptPream sends compact semantic prompts that are decoded into high quality frames using generative models such as Stable Diffusion. We propose three core techniques to enable this paradigm: (1) a gradient based prompt inversion method that compresses frames into small sets of compact token embeddings; (2) a computation aware scheduling strategy that jointly optimizes network and compute resource usage; and (3) a scalable searching algorithm that addresses the enlarged scheduling space introduced by scheduler. Evaluations show that PromptStream reduces both stalls and bandwidth waste by over 31%, and improves Quality of Experience (QoE) by 45%, compared to traditional strategies.