ANALYSIS OF CEPH-BASED SOFTWARE DEFINED STORAGE SYSTEM PARAMETERS EFFECT ON DATA STORAGE SYSTEM PERFORMANCE

• Main Author: Hilmi, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/39477
• Institution: Institut Teknologi Bandung

The result of data are varied. The growth has implications for the increasing need for data storage systems. The problem of data storage systems at enterprise level is the limitation in adding capacity easily and efficiently due to dependence on developing storage systems to storage service provider vendors. In this Final Project, research and implementation of Ceph-based Defined Storage software are carried out. Ceph is a software, able to overcome the problem of data storage systems at the enterprise level. The software can be built using all types of data storage devices without being tied to any brand and vendor. On the other hand, it could become a challenge in building data storage clusters with Ceph-based. The high of the flexibility causes no universal configuration that can be applied to Ceph storage systems. However, there are critical parameters in designing a Ceph cluster. The main focus of this book is discussing the parameters that have a major influence on Ceph storage performance. The parameters tested are the network capacity used in the Ceph cluster, the type of data storage device, and the data storage backend. The testing process is carried out by determining the critical parameters, then the parameters the test method that will be able to represent the performance of the Ceph storage cluster will be derived. These parameters are very important to consider when it would build a Ceph storage cluster. If there is an error in configuring the parameters, the result can result in non-optimal performance, even though it has used high performance devices.