Optimizing Kubernetes Performance by Handling Resource Contention with Custom Scheduler

Mulubagilu Nagaraj, Akshatha (2020) Optimizing Kubernetes Performance by Handling Resource Contention with Custom Scheduler. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (944kB) | Preview

Abstract

In virtualized environment, multiple instances use same resources of the environment, implying that chances of requesting similar resources for finishing jobs at same point in time is critical. Furthermore, resource contention is faced if requested resources exceed the accessibility of resources. Most cases enter a waiting state, when more than one job requests for same set of resources in completion of the task with only few requests being attended to execute. As a result of such delays, there is overall performance degradation. It is a peculiar issue which resurfaces in Kubernetes container management system. Activities like placement location and resource provision is the main aim of Kubernetes scheduler. Container placement, dependent on CPU and memory parameters, is the default and conventional rule in Kubernetes. Although it is very well known that these two factors are not the only resources in a shared environment leading to resource contention issues. Careful and exhaustive resource usage is taken into consideration for placement of the container through the implemented scheduler. To reduce this issue, the implemented solution discusses about how two containers intensely utilizing the same resources could be placed in two distinct and separate Kubernetes pods. So, this proposed Kubernetes scheduler handles resource contention problem resulting into improved performance.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Cloud computing
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Dan English
Date Deposited:	28 Jan 2021 16:56
Last Modified:	28 Jan 2021 16:56
URI:	https://norma.ncirl.ie/id/eprint/4543

Actions (login required)

View Item