Analysis of Object Storage Using Perfaccel
Loading...
Files
Date
2017-09-28T13:15:03Z
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Object storage is popular in the data storage market due to its ability to scale and store large volumes of data. Environments with massive amount of data require some insights to further optimize and organize their storage needs.
Object Storage System Characteristics
• Data is stored as independent objects with a unique identifier
• Flat storage design that facilitates greater scalability
• Could be scaled-out to 100s of Petabytes
• Doesn't use RAID but instead implements Replication
- At PBs scale RAID could have implications such as long rebuild times
• Access Over RESTful API over HTTP
- Amazon S3, OpenStack Swift, Ceph
The Object Storage vendors provide CLI Utility to interact with their storage backend, these tools are simple and easy to use, which are intended for performing basic operations. Attempting to use these vendor provided tools for analysis is inappropriate as they are not designed specifically for this purpose. A more dedicated analytics tool is desired which provides much deeper insights based on which the resource could be allocated ensuring efficient storage utilization.
This project aims at providing sufficient analytical data using PerfAccel which is a product of Datagres. PerfAccel is a profound grid scale management and acceleration solution providing deep storage analytics and I/O caching. It uses high speed devices
such as ssd to cache and accelerate performance, through which the underlying object storage can be improved. It provides various analytical data which can be used to optimize object storage and aid storage administrators in allocating resources.
Description
Keywords
1NZ14MCA28, Analysis of Object Storage Using Perfaccel, Shaik Mohammed Ali