Performance Analysis of Distributed File System Based on RAID Storage for Tapeless Storage

Tape storage is used as a long-term data storage solution in institutions and research centers around the world, contributing to the advancement of business and scientific research. Despite the existence of other storage media, tape storage has gained popularity owing to its low cost per storage capacity, making it an affordable option for initial implementation. However, recently, the strengths of tape storage have become ambiguous, and their drawbacks have become more evident. In the current storage system, tape storage requires complex management compared to other storage media, which incurs additional human resource costs, making it less cost effective. On the other side, other storage media continue to become cheaper and have larger storage capacities. Therefore, companies and research centers are no longer required to continuously purchase tape storage. With this trend, distributed file systems can be an alternative to tape storage. Distributed file systems combine storage nodes into a network through Reliable Array of Independent Nodes (RAIN). Erasure coding is used to store data in this file system, which makes the system cost-effective and highly reliable. This allows the construction of a safe and efficient storage system. In this study, we constructed and evaluated a distributed file system based on Redundant Array of Independent Disks (RAID) storage. We configured and evaluated distributed file systems that use Reed-Solomon in the same environment, providing the performance characteristics of each distributed file system.

Automated summary

Tape storage is widely used for long-term data storage, but it requires complex management and is less cost-effective compared to other storage media. Distributed file systems have become an alternative option due to their cost-effectiveness and reliability.