CoRaiS: Lightweight Real-Time Scheduler for Multi-Edge Cooperative Computing

TL;DR

CoRaiS tackles real-time scheduling in heterogeneous, multi-edge cooperative computing by introducing a system-level state evaluation model to abstract hardware diversity, formulating a MILP baseline for edge-dispatch, and proposing a lightweight RL-based scheduler with a matching-on-demand architecture. The approach learns to minimize the maximum per-edge completion time via edge and request embeddings processed by a context decoder, enabling real-time, near-optimal decisions across varying workloads and scales. Simulation and prototype evaluations show CoRaiS achieving near-optimal performance with decision times around $0.02s$, while generalizing to larger instances and handling heterogeneous edge configurations. The work advances scalable, real-time edge coordination and provides a foundation for hierarchical interoperability in future multi-edge ecosystems.

Abstract

Multi-edge cooperative computing that combines constrained resources of multiple edges into a powerful resource pool has the potential to deliver great benefits, such as a tremendous computing power, improved response time, more diversified services. However, the mass heterogeneous resources composition and lack of scheduling strategies make the modeling and cooperating of multi-edge computing system particularly complicated. This paper first proposes a system-level state evaluation model to shield the complex hardware configurations and redefine the different service capabilities at heterogeneous edges. Secondly, an integer linear programming model is designed to cater for optimally dispatching the distributed arriving requests. Finally, a learning-based lightweight real-time scheduler, CoRaiS, is proposed. CoRaiS embeds the real-time states of multi-edge system and requests information, and combines the embeddings with a policy network to schedule the requests, so that the response time of all requests can be minimized. Evaluation results verify that CoRaiS can make a high-quality scheduling decision in real time, and can be generalized to other multi-edge computing system, regardless of system scales. Characteristic validation also demonstrates that CoRaiS successfully learns to balance loads, perceive real-time state and recognize heterogeneity while scheduling.

Figures (11)

• Figure 1: The illustration of unbalanced workloads and resource utilization of edge computing.
• Figure 2: An illustration of the motivating use case.
• Figure 3: An illustration of the motivating use case with mobility.
• Figure 4: Multi-edge cooperative computing system modeling. Part A: The edges are physically interconnected. Each edge has its dedicated client region and the services deployed on each edge operate independently. Part B: The multi-edge cooperative computing system is decomposed into multiple independent service-oriented subsystems, where the edges with same service are grouped.
• Figure 5: Territorial schedulers are introduced into the service-oriented multi-edge cooperative computing system to facilitate effective and real-time scheduling.