A fault-tolerant aware scheduling method for fog-cloud environments

Area of Science:

• Computer Science
• Distributed Systems
• Cloud Computing

Background:

• Fog computing extends cloud capabilities to the network edge, supporting Internet of Things (IoT) devices.
• The dynamic and heterogeneous nature of edge devices leads to frequent failures, necessitating robust fault tolerance.
• Existing scheduling and fault-tolerant methods often neglect time-sensitive requests, causing unacceptable latency in IoT applications.

Purpose of the Study:

• To propose a fault-tolerant scheduling method (FTSM) for fog-cloud IoT environments.
• To minimize service latency and overhead while enhancing cloud reliability and capacity.
• To address the limitations of previous approaches concerning time-sensitive requests.

Main Methods:

• Categorizing devices and requests into three classes: time-sensitive, time-tolerant, and core.
• Mapping time-sensitive requests directly to edge devices.
• Assigning time-tolerant requests to edge or cloud core devices, and core requests to the cloud core.
• Dynamically selecting the most suitable fault-tolerance technique (replication, checkpointing, resubmission) for each request.

Main Results:

• The proposed FTSM effectively reduces average service time and operation costs.
• It enhances throughput and success rate for service requests.
• The method improves the overall capacity utilization of the fog-cloud infrastructure.

Conclusions:

• The FTSM provides a superior approach to scheduling and fault tolerance in fog-cloud IoT systems.
• It significantly mitigates latency issues for time-sensitive requests.
• The method offers a flexible and effective solution for managing heterogeneous IoT environments.