|
Title:
|
PERFORMANCE COMPARISON BETWEEN SLURM BARE METAL AND SLURM ON KUBERNETES IN HPC WORKLOADS |
|
Author(s):
|
Lucas Borges, Wellington Oliveira, Tafarel Souza, Phillip Furtado and Erick Bezerra |
|
ISBN:
|
978-989-8704-71-9 |
|
Editors:
|
Paula Miranda and Pedro Isaías |
|
Year:
|
2025 |
|
Edition:
|
Single |
|
Keywords:
|
Kubernetes, High-Performance Computing, Bare Metal, Simple Linux Utility for Resource Management, Benchmarks,
DevOps |
|
Type:
|
Full Paper |
|
First Page:
|
166 |
|
Last Page:
|
173 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
if you are a member please login 
|
|
Paper Abstract:
|
The increasing adoption of Kubernetes in High-Performance Computing environments has sparked significant debates
about its impact on performance compared to bare metal Simple Linux Utility for Resource Management. This study
provides a comprehensive comparison between the two paradigms, using rigorous benchmarks to evaluate critical metrics
such as scalability, latency, resource utilization, and fault tolerance. The results show that Kubernetes, when integrated
with Simple Linux Utility for Resource Management, offers enhanced flexibility and scalability, particularly in scenarios
requiring container orchestration and process automation. This combination allows for better adaptation to dynamic and
heterogeneous environments, while also facilitating the implementation of DevOps practices in High-Performance
Computing. On the other hand, Simple Linux Utility for Resource Management bare metal maintains its efficiency in direct
resource management and performance, making it particularly advantageous in scenarios where the priority is maximizing
hardware resource utilization and minimizing management overhead. The analysis identifies the advantages and limitations
of each approach, highlighting how the choice between these technologies should consider the specific application needs,
such as workload nature, environmental complexity, and fault tolerance requirements. Additionally, the study addresses
the evolution of High-Performance Computing practices, demonstrating how technologies like Kubernetes can drive
infrastructure and processes. This work aims to contribute to the scientific and technological community by providing
valuable insights for optimizing High-Performance Computing cluster infrastructure and informing technology adoption
decisions. |
|
|
|
|
|
|
