A Survey on Services Placement Algorithms in Integrated Cloud-Fog / Edge Computing
Résumé
The evolution of computing paradigms, such as Fog and Edge, has led to the emergence of new applications that require a placement of services close to the end users. Optimal placement of such applications over network nodes is therefore an important issue in integrated Cloud-Fog/Edge computing environments. The service placement problem is challenging due to the complexity of such distributed systems, limited computing resources, placement constraints, global efficiency of the application and quality requirements for end users. This survey attempts to provide a comprehensive and structured presentation of services placement problem research in integrated Cloud-Fog/Edge. We propose a new classification of services placement problem approaches by adopting an algorithmic viewpoint and a breakdown into three categories: graph-based, heuristics and machine-learning approaches. We also provide an analysis and some discussions of the methods, and identify some future challenges and issues in the case of microservices. We hope that this survey will provide a better understanding of the solutions and directions taken by research on this important topic.