Container-Based Microservice Scheduling Using Reinforcement Learning in Distributed Cloud Computing

Full text
Author(s):	Marques Matos, Gabriel Henrique ; Carvalho, Marcos ; Macedo, Daniel F. Total Authors: 3
Document type:	Journal article
Source:	2024 IEEE LATIN-AMERICAN CONFERENCE ON COMMUNICATIONS, LATINCOM; v. N/A, p. 6-pg., 2024-01-01.
Abstract
Containers provide significant advantages for application deployment in cloud computing environments. Applications are now divided into multiple small services (microservices) that communicate among themselves to ensure effective application operation. However, this paradigm shift introduces challenges in distributed cloud computing, where microservices are spread across different zones that may be hundreds of kilometers apart, resulting in latency issues in microservice communication. This paper proposes microservice scheduling based on Reinforcement Learning (RL) by considering microservice interdependencies to optimize their placement. Experimental results indicate that our RL-based scheduler reduces average latency by 22.51% compared to the default Kubernetes scheduler. (AU)

FAPESP's process:	20/05182-3 - PORVIR-5G: programability, orchestration and virtualization in 5G networks
Grantee:	José Marcos Silva Nogueira
Support Opportunities:	Research Projects - Thematic Grants


FAPESP's process:	18/23097-3 - SFI2: slicing future internet infrastructures
Grantee:	Tereza Cristina Melo de Brito Carvalho
Support Opportunities:	Research Projects - Thematic Grants