D2D assisted cooperative computational offloading strategy in edge cloud computing networks.

Area of Science:

• Edge Cloud Computing
• Distributed Systems
• Task Scheduling

Background:

• Existing edge cloud computing (ECC) offloading strategies often neglect task waiting delay, which is critical for latency-sensitive applications.
• Optimizing user cost in ECC typically focuses on delay and energy consumption, overlooking significant waiting times.

Purpose of the Study:

• To propose a D2D-assisted collaborative computational offloading strategy (D-CCO) that incorporates task waiting delay into user cost optimization.
• To enhance offloading decisions in ECC by considering a comprehensive user cost model.

Main Methods:

• Developed a task queue system involving local devices, peer devices, and edge processors.
• Utilized stochastic optimization and back-pressure algorithms to create the D-CCO strategy.
• Analyzed system stability theoretically by deriving queue upper bounds.

Main Results:

• The D-CCO strategy effectively reduces overall user cost, including task delay, energy consumption, and waiting delay.
• Theoretical analysis confirmed the stability of the proposed algorithm.
• Simulation results demonstrated the superiority of D-CCO over alternative offloading strategies.

Conclusions:

• The D-CCO strategy offers a more holistic approach to computational offloading in ECC by including waiting delay.
• The proposed method ensures system stability while minimizing computing costs.
• D-CCO significantly improves upon existing strategies in terms of user cost reduction.