AppFlow: Memory Scheduling for Cold Launch of Large Apps on Mobile and Vehicle Systems

2026-03-17

AI Summary

This work addresses the severe cold-start latency experienced by large applications in multitasking scenarios, where frequent memory reclamation or process termination often leads to startup delays exceeding the one-second user-experience threshold. To tackle this issue, the paper proposes AppFlow—a system-level predictive scheduler that, for the first time, jointly optimizes file-access pattern prediction with memory scheduling. Without requiring any application modifications, AppFlow accelerates cold starts through selective preloading, adaptive memory reclamation, and context-aware process termination, seamlessly integrating across the Android framework and the Linux kernel. Experimental results demonstrate that AppFlow reduces average cold-start latency by 66.5%—for instance, from 2 seconds to 690 milliseconds—and maintains 95% of launches under one second during a hundred-day real-world evaluation.

Technology Category

• Planning, Routing, and Scheduling: Applications
• Multiagent Systems: Applications
• Machine Learning: Hardware-aware ML

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

GB-scale large apps like on-device LLMs and rich media editors are becoming the next-generation trend, but their heavy memory and I/O demands, especially during multitasking, cause devices to reclaim or kill processes, turning warm apps into cold launches. The challenge lies not in storing them, but in fast, accurate launching. For users, 1s is the usability cliff, yet our measurements show 86.6\% of GB-scale cold launches exceed it. Also, Android Vitals flags only $\geq$ 5s as slow, exposing a large satisfaction gap. Existing optimizations are designed in isolation and conflict. For example, preloading reduces I/O stalls but consumes scarce memory and is undone by reclamation, while reclamation and killing free memory but sacrifice background survivability, leading to repeated cold relaunches. Our key insight is that, although multitasking makes runtime behavior complex, each app's file access pattern remains predictable. The challenge lies in exploiting this predictability, i.e., preloading without exhausting memory, reclaiming without undoing gains, and killing selectively to preserve background survivability. We introduce AppFlow, a prediction-based system-wide scheduler that integrates a Selective File Preloader, an Adaptive Memory Reclaimer, and a Context-Aware Process Killer. Implemented across the Android framework and Linux kernel without app changes, AppFlow cuts GB-scale cold-launch latency by 66.5\% (e.g., 2s$\rightarrow$690ms) and sustains 95\% of launches within 1s over a 100-day test, significantly improving responsiveness and multitasking experience.

Problem

• cold launch
• memory scheduling
• large apps
• multitasking
• user experience

Innovation

• cold launch optimization
• memory scheduling
• selective preloading
• adaptive reclamation
• context-aware killing