UDC 681.3

Authors

N.I. Chervyakov, V.V. Berezhnoy, M.A. Deryabin, A.S. Nazarov

Abstract

This article discusses fault tolerant storage systems using various RAID technologies. These technologies, for the most part, enable hot-swapping of failed disks without loss of performance, for the most part, combine parity with data duplication, which requires at least 50% of the hardware overhead.
To ensure the fault tolerance of data storage systems, it is proposed to use an excess system of residual classes (ISOK), in which data is divided into separate low-bit blocks on the basis of ISOK, recorded on separate disks. This presentation of information, combined with the equality of blocks of information and the ability to correct distorted data during crashes and failures of individual drives on the fly, allows you to provide the fault tolerance feature of a storage system with less hardware than RAID technologies.

Keywords

data storage systems, fault tolerance, redundancy, a system of residual classes, hardware costs.