A novel approach for dynamic task scheduling for IOT in fog-cloud environment

Area of Science:

• Computer Science
• Distributed Systems
• Artificial Intelligence

Background:

• The proliferation of real-time Internet-of-Things (IoT) applications challenges centralized cloud computing.
• Distributed computation across IoT, fog, and cloud layers introduces complex scheduling demands.
• Existing schedulers struggle with heterogeneous resources, dynamic task arrivals, and competing objectives like latency, energy, and cost.

Purpose of the Study:

• To introduce a novel scheduling framework, the Quantum-inspired Biased Dynamic Scheduler (QBDS), for the IoT-fog-cloud continuum.
• To optimize a configurable Composite Objective Function (COF) that integrates makespan, energy consumption, cost, load balance, resource utilization, and temporal metrics.
• To address the limitations of current scheduling approaches in dynamic and resource-constrained edge computing environments.

Main Methods:

• Developed a priority-aware task ranking mechanism that adaptively weights task characteristics (deadline slack, execution length, memory, data size).
• Implemented a sinusoidal, quantum-inspired biasing technique to perturb task and node metrics, aiding exploration and escaping local optima.
• Utilized a penalty-aware multi-objective cost evaluator for informed task-to-node assignment decisions.

Main Results:

• QBDS consistently improved key performance indicators including makespan, energy consumption, cost, and resource utilization.
• Demonstrated robust performance scaling under heavy computational loads and diverse network topologies.
• Ablation studies confirmed the effectiveness of individual components within the QBDS framework.

Conclusions:

• QBDS offers a superior scheduling solution for latency-sensitive IoT applications in hierarchical computing environments.
• The quantum-inspired approach effectively balances multiple, often conflicting, optimization objectives.
• QBDS provides a scalable and efficient method for managing resources in the evolving IoT-fog-cloud landscape.