Task offloading for multi-server edge computing in industrial Internet with joint load balance and fuzzy security

Area of Science:

• Industrial Internet of Things (IIoT)
• Edge Computing
• Task Offloading

Background:

• Industrial terminals face challenges with latency-sensitive and computation-intensive tasks in the Industrial Internet.
• Edge computing (EC) offers a solution by offloading tasks to adjacent edge servers to augment computational capacity.

Purpose of the Study:

• To address the challenge of developing accurate offloading strategies for EC in the Industrial Internet.
• To study the latency-aware multi-server partial EC task offloading problem, incorporating load balancing and security.

Main Methods:

• Developed a task offloading model supporting partial offloading, multi-server distribution, load balancing, and fuzzy risk assessment.
• Formulated the model as a constrained optimization problem, proving its NP-hardness.
• Proposed a bi-layer offloading algorithm using adaptive genetic algorithm and simulated annealing particle swarm optimization for joint load balance and fuzzy security.

Main Results:

• The established model effectively reduces the objective value, decreasing it by 27% compared to full edge execution and 46% compared to local execution.
• The proposed bi-layer offloading algorithm demonstrates superior solution accuracy over existing methods.

Conclusions:

• The proposed latency-aware partial offloading model and algorithm are effective for the Industrial Internet.
• The approach successfully balances latency reduction, load balancing, and security considerations in task offloading.