Analysis of object-based storage implementation on storage data center

Abstract

The data center is the backbone of various services offered by internet giant companies  like  Google,  Amazon,  and  others.  Increasingly  complex,  varied,  and  the rapid  growth  of  services  offered,  the  data  center  will  continue  to  grow  and  develop which  makes  data  center  management  and  planning  more  challenging.  One  of  the most  important  parts of  data  centers  that  require  good  planning  is  network  storage. This network storage determines how data is stored in the data center so that it can be accessed  at  any  time  by  service  users  effectively.  Network  storage  is  a  term  used  to refer  to  a  network-based  storage.  Based  on  the  results  of  a  survey  conducted  in Australia, human error was the biggest cause of sudden data center malfunction with a percentage  of  28%.  The  next  biggest  causes  are  electrical  problems,  hardware  and software which each have a percentage of 25%, 21% and 11%. While natural disasters such  as  earthquakes  only  have  a  percentage  of  8%.  This  study  applies  object  based storage  to  data  center  storage.  The  storage  object  used  is  openstack  object  storage. Measurement  of  network  performance using  COSbench  software.  Measurement  of network performance is carried out under normal conditions which include read, write and  delete  activities. The file  size  operated  when  measuring  network  performance  is divided into four clusters, tiny, small, medium and large.