International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 181 - Number 28 |
Year of Publication: 2018 |
Authors: Saleh A. Khawatreh, Enas N. Al-Zubi |
10.5120/ijca2018918084 |
Saleh A. Khawatreh, Enas N. Al-Zubi . Survey on a Google File System (GFS). International Journal of Computer Applications. 181, 28 ( Nov 2018), 9-16. DOI=10.5120/ijca2018918084
The Google File System (GFS), a proprietary scalable distributed file system sophisticated by Google to be used for its huge distributed data applications, primarily the search engine. It improves efficiency, reliability, scalability, transparency, security and fault tolerance despite of using cheap commodity computers and serving a large number of clients. GFS is similar with the previous distributed file systems in the goals, it divides the files into chunks of data each one is 64 MB in size and generate several copies of each data chunk. Then mounts these data chunks on several servers which could be based on global geographical locations, and that will enhance the reliability of retrieving files online and overcome any limited access to one or more servers. It will also enhance the concurrency access and control over files online due to the fact that several copies of each file do exist at the same time. GFS has successfully met Google’s storage needs, for generation and data processing such as research and development issues that need huge data sets. The biggest cluster provides hundreds of TBs of storage distributed in thousands of disks on thousand machines, and it provides concurrency access by hundreds of users.