International Journal of Computer and Communication Technology
Abstract
Distributed data storage systems are used to store data reliably over a distributed collection of storage locations, called peers. Coding schemes are used to store a portion of the data in the peers ensuring the complete retrieval of data, during peer failures. This has applications in various areas like Wireless Networks, Sensor Networks etc. In this framework we consider a large file to be stored in a distributed manner over few peers of limited capacity. Each peer stores a portion of the coded data, without the knowledge of the contents of other peers. Random Coding is one of the coding schemes used for this. In [1] coding coefficients are chosen randomly from a finite field to encode the data. The encoding is basically a linear combination of file pieces (pieces are elements of finite fields). The data downloader downloads these coded data from several peers and decodes to get the original data. The decoding is basically solving a system of linear equations over a finite field, which is the most time consuming step in the whole process. We give a simple C++ implementation of the schemes in [1] and plot the results. We are trying to find a scheme where coding vectors can be chosen such that the decoding complexity is reduced significantly. Also in a dynamic setting where nodes enter and leave system intermittently, are discussed.
Recommended Citation
Tabatabaeian, Saman; Lal, Rajendra P.; and Naik, Wilson
(2014)
"Coding Schemes for Distributed Storage Systems: Implementation and Improvements,"
International Journal of Computer and Communication Technology: Vol. 5:
Iss.
1, Article 8.
DOI: 10.47893/IJCCT.2014.1218
Available at:
https://www.interscience.in/ijcct/vol5/iss1/8
DOI
10.47893/IJCCT.2014.1218
