Loading...

Evaluation of Deduplication Technuques in Data Storage Systems

Bazazzadegan, Mohammad Hossein |

573 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: English
  3. Document No: 44312 (52)
  4. University: Sharif University of Technology, International Campus, Kish Island
  5. Department: Science and Engineering
  6. Advisor(s): Asadi, Hossein
  7. Abstract:
  8. Deduplication is a data reduction technique, which eliminates redundant data by storing only a single copy of each file or block, along with a reference to the unique copy of data. Deduplication reduces the storage space and bandwidth requirements of data storage systems, and becomes more effective when applied across multiple users. All deduplication techniques use cryptographic algorithms to detect duplication instead of byte-by-byte comparison between input data and all previously stored data. In this manner, the data stream is divided into non-overlapping chunks of data. Then, the hash values of non-overlapping chunks create an exclusive criterion to identify duplicates. In this thesis, we have showed the storage efficiency of deduplication storage systems in Virtual Machine (VM) workloads. Further, we have proposed hybrid storage architecture for deduplication, which uses Hard Disk Drives (HDDs) and Solid-State Drives (SSDs). This has been achieved by characterizing data and metadata into hot and cold data and storing them in SSDs and HDDs, respectively. Afterward, we have showed that the proposed technique improves the overall performance of deduplication storage systems. To evaluate performance improvement achieved by the proposed deduplication storage architecture, we have used DiskSim toolset to model the proposed storage architecture. Finally, we have represented the overall performance improvement about 3.7x which has been achieved by the proposed storage architecture in a deduplication storage system
  9. Keywords:
  10. Solid State Disk Drive ; Hard Disk Drives ; Deduplication ; Data Chunk ; Metadata

 Digital Object List

 Bookmark

No TOC