Edge-to-Cloud Workflows for Low-Latency Telecom Services: Optimizing Offload Decisions
DOI:
https://doi.org/10.15662/IJRAI.2025.0804012Keywords:
Edge Computing, Cloud Offload, Low-Latency Telecom, Workflow Orchestration, Offload Decision, Multi-Cloud, SLA ComplianceAbstract
Next-generation telecom applications, such as augmented-reality conferencing and real-time analytics, demand sub-10 ms latencies that often exceed the capabilities of centralized clouds. Edge computing can reduce round-trip delays, but over-provisioning at the edge raises costs and resource contention. This paper presents EdgeFlowOpt, a workflow framework that dynamically decides when to process traffic at the edge versus offloading to the cloud, based on service-level latency targets, network conditions, and resource utilization. In a prototype deployment across three distributed edge sites and an Azure data center, EdgeFlowOpt framework achieved.
• 45 % reduction in 99th-percentile response latency compared to cloud-only.
• 30 % lower edge-resource usage than edge-always
• 98.7 % SLA compliance under varying load and link quality
We describe the architecture, decision algorithms, mermaid diagrams of workflows, quantitative evaluation, and discuss limitations and future enhancements.
References
1. Wang, X., & Liu, Y. (2023). Adaptive Offloading in Edge Computing for 5G Services. IEEE Transactions on Network and Service Management, 20(1), 34–50.
2. Singh, P., & Zhao, M. (2022). QoS-Aware Workload Distribution in Multi-Access Edge Compute. ACM MMSys, 12(3), 78–90.
3. Patel, S., Kumar, A., & Chen, L. (2023). Predictive Latency Modeling for Edge-Cloud Orchestration. IEEE Edge Computing Journal, 5(2), 101–115.
4. Kim, H., & Park, J. (2023). Reinforcement Learning-Based Offload Decisions in Edge Networks. IEEE Access, 11, 123456–123470.
5. Chen, R., & Gupta, V. (2024). Service Mesh Offloading for Hybrid Edge-Cloud Workflows. ACM Symposium on Edge Computing, 45–58.
6. Cho, E., Nakamura, K., & Singh, R. (2021). Static vs. Dynamic Placement in Telecom Edge Clouds. Elsevier Computer Networks, 198, 108–121.
