RNCDDS - Random Network coded distributed data system
Hiroshi Nishida, Thinh P. Nguyen · 2017
Many Random Network Coding based storage systems have been proposed to increase robustness in terms of data preservation and storage efficiency. However, there are two major practical issues: (1) slow encoding and decoding speeds and (2) difficult access to the data. In this paper, RNCDDS solves problem (1) by introducing a new and efficient Galois Field arithmetic library and problem (2) by employing a new JavaScript program. As a result, RNCDDS is not only theoretically more robust and storage efficient than major distributed data systems such as Hadoop (HDFS) and GlusterFS, but also outperforms them in all speed measurements. Also, the proposed JavaScript program enables easy data fetch through a web browser including watching a video with an HTML5 video player and has potential for drastically reducing data storage amount in a cloud system or Content Delivery Network.