Nov 2026· Proceedings of the 34th ACM International Conference on Multimedia· 0 citations· 13 references
Abstract
Modern Extended Reality (XR) systems are becoming increasingly heterogeneous, as latency-critical perception and rendering pipelines are now coupled with compute-intensive and bursty foundation-model-based agents. This trend makes edge-cloud offloading essential for sustaining interactive performance under dynamic workloads. However, existing offloading architectures are limited by a fundamental bottleneck: task migration in XR is often dominated by the transfer and restoration of heavyweight execution context, such as scene assets and model parameters, which incurs severe cold-start delay and undermines real-time responsiveness. We present DEXA, a task offloading framework that enables low-overhead migration through static-dynamic context decoupling. DEXA separates each task into a heavyweight static context and a lightweight dynamic context, pre-stages static assets on edge and cloud nodes, and reduces runtime migration to dynamic-state synchronization. On top of this design, DEXA employs a penalty-based online scheduler to jointly optimize service latency and migration cost. We implement DEXA on a prototype that integrates high-fidelity rendering and LLM-based agent serving. Results show that DEXA effectively mitigates cold-start overhead during migration, maintains rendering SLA compliance, and improves system throughput and responsiveness under dynamic load.
Serverless computing has emerged as a pivotal cloud deployment paradigm due to its efficient utilization of on-demand resources. However, the persistent cold start problem continues to hinder system performance. Current strategies often employ oversimplified preloading/unloading mechanisms that inadequately exploit inv...
Cheryl Lee, Zi-Shan Su, Zhouruixing Zhu et al.· 0 citations
Pervasive intelligent applications are increasingly deployed on mobile and Internet of Things (IoT) edge devices. Consequently, Large Language Models (LLMs) are increasingly used to support these applications. Yet, due to their high resource demands, LLMs are mostly deployed in the cloud. Layer-wise edge-cloud inferenc...
Jing-Po Xu, Paul Joe Maliakel, Ivona Brandić et al.· 0 citations
Sluice is introduced, an adaptive data stream processing system for hybrid clouds, compatible with dataflow streaming engines, and shows that Sluice seamlessly shifts computations between edge and cloud, while consistently achieving higher throughput than baseline systems.
Yuan-Li Wang, Prateek Jain, Lei Huang et al.· 0 citations
The widespread deployment of foundational Large Language Models (LLM) across cloud datacenters marks a fundamental paradigm shift from stateless deep learning toward distributed, memory-intensive generative workloads. However, serving massive foundation models under stringent Service Level Objectives (SLO) exposes crit...
Jannatun Noor, Athar Noor Mohammad Rafee, Md Sadman Haque et al.· 0 citations
GPUs have become indispensable infrastructure for cloud platforms, powering workloads from large-model inference to real-time rendering, and the transition to cloud-native GPU computing is accelerating. One of the key challenges in this transition is live migration, which is essential for enabling transparent workload...
Kai-Liang Xu, Sen-Hao Yu, Song-Tao Xue et al.· ACM Transactions on Architec...· 0 citations
Large-scale data-intensive workloads powered by iterative graph or machine learning processing have an insatiable demand for memory due to the continuous generation of new intermediate data. Although existing caching strategies strive to keep performance-critical data in memory, they often fail in practice because they...
Jongseop Lee, Min-Sui Sun, Xinyue Ma et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.