OpenStack Implementation using Multinode Deployment Method for Private Cloud Computing Infrastructure

The worldwide market for cloud computing alone is expected to reach \mathrm{US}{\}947.3 billion by 2026. OpenStack technology creates opportunities to build private clouds. Department of Information Technology ITS creates its own private cloud using OpenStack and employs the multinode deployment mec...

Full description

Saved in:
Bibliographic Details
Published in2023 International Seminar on Intelligent Technology and Its Applications (ISITIA) pp. 12 - 17
Main Authors Ciptaningtyas, Henning Titi, Hariadi, Ridho Rahman, Husni, Muchammad, Ghozali, Khakim, Sholikah, Rizka Wakhidatus, Setyadharma, I Made Dindra
Format Conference Proceeding
LanguageEnglish
Published IEEE 26.07.2023
Subjects
Cloud computing
Computer architecture
Correlation
Cyber Range
Information technology
KYPO
Openstack
Private Cloud
Seminars
Training
Online AccessGet full text

Cover

More Information
Summary:The worldwide market for cloud computing alone is expected to reach \mathrm{US}{\}947.3 billion by 2026. OpenStack technology creates opportunities to build private clouds. Department of Information Technology ITS creates its own private cloud using OpenStack and employs the multinode deployment mechanism developed by Kolla Ansible. The OpenStack implementation's results were evaluated by using the Phoronix Test Suite, Ethr, and Rally-Openstack tools to identify the correlation between instance computing performance and the number of running instances, as well as OpenStack service performance while generating instances and changing flavors. The results show that the CPU performance of each instance will decrease with an average decrease of 0.68 GFLOPS or 1.41% as the number of instances increases. The memory performance of each instance will decrease with an average reduction of 0.37 GBps or 4.46%. The duration of the instance creation is unaffected by the number of concurrent requests, with an increase of only 0.3 seconds. The number of simultaneous requests also affects the time of flavor changes. The duration of flavor changes can be up to 2 times (in 4 concurrent requests) from one concurrent request but experiences instability in 5 or more concurrent requests. The KYPO Cyber Range Platform successfully implemented a linear training demo by allocating one sandbox pool to run the training using Openstack resources of 10 vCPUs and 14.3 GB of memory.
ISSN:2769-5492
DOI:10.1109/ISITIA59021.2023.10221042