Auction-Based Microservice Placement in Fog Computing: Algorithms, Analysis, and Performance Evaluation
D. Suresh Kumar, Utkarsh Kumar, Abhigyan Prakash, Anjan Bandyopadhyay, Sujata Swain · 2024
This research paper introduces a distributed fog service placement (DFSP) algorithm for efficient microservice allocation in the field of fog computing environments, aiming to reduce power usage and data transfer costs. The algorithm utilizes an iterative combinatorial auction approach, where fog applications bid for resources on fog nodes to host their microservices. Two versions of the algorithm are proposed: DFSP-GF, which relies on global feedback, and DFSP-NF, which uses neighbor feedback. Both versions exhibit fast convergence, maintain privacy, and meet resource constraints. Simulation results showcase the efficacy of the DFSP algorithms in terms of convergence speed, resource utilization, allocation distribution, and total cost. DFSP-NF provides a scalable and practical solution for resource-constrained fog environments, while DFSP-GF offers further cost optimization at the cost of higher communication overhead. The DFSP algorithms present a distributed and privacy-preserving approach to solving the fog service placement problem, considering the challenges posed by the distributed nature of fog resources. The choice between DFSP-NF and DFSP-GF depends on specific requirements, such as scalability, communication efficiency, and the trade-off between global optimization and local decision-making..