SLA-Aware Hierarchical Joint Scheduling for Hybrid Tasks and Service Function Chains in Converged Cloud-Edge-End Networks
Year: 2026
Type: conference-paper
Abstract: This paper studies hybrid-request scheduling in converged cloud-edge-end networks, where ordinary tasks and service function chain (SFC) requests compete for shared compute and bandwidth. We propose SLA-aware Hierarchical Joint Scheduling (SHJS), which unifies admission, initial placement, compute-bandwidth allocation, and local repair in one framework. SHJS first screens requests with an SLA risk score, then performs weighted joint allocation, and finally repairs overload through rerouting, migration, and selective rollback. Experiments cover four scales, twenty paired seeds, and two topology profiles. On the bottleneck topology, SHJS keeps the objective value within 0.189717-0.191477 from Small to XLarge, improves the objective by 38.62% over Seq-TaskFirst and 54.29% over Greedy-Latency on average, and raises admitted SFC deadline satisfaction by 37.99 percentage points over the no-admission ablation. Holm-corrected Wilcoxon tests confirm the significance of the gains. The results show that moderate admission control can produce better admitted-SLA than highthroughput heuristics under strong network pressure.
Authors: Yiran Zhao, Chenwei Wang, Zihan Shen, Songlin Sun
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7214101091