An Erasure Code Coding and Data Reconstruction Method for Reducing Data Update Cost

Abstract

The invention discloses an erasure code encoding and a data reconstruction method that reduces data update cost, adds an additional diagonal check digit to each coded strip, and designs encoding rules for each diagonal check digit , Eliminate the dependency between the check bits in the array erasure code, reduce the number of check bits associated with each original data bit, and reduce the related check bit update caused by the array erasure code when the original data is updated Cost: Reconstruct the data in the case of single disk failure or double disk failure, and complete the recovery of invalid data when there are less than two data or verification disk failures, so as to realize the function of data fault tolerance. By increasing the storage overhead of a small amount of parity bits, the present invention has higher data encoding and data reconstruction speeds, and reduces the additional update cost of the parity disk when the erasure code system is updating data, which is beneficial to slow down the speed of new NVM Wear and tear of high-speed storage devices such as / SSD.

Description

technical field [0001] The invention belongs to the technical field of computer storage systems, and in particular relates to an erasure code encoding and a data reconstruction method for reducing data updating costs. Background technique [0002] Erasure codes such as RS and RDP can tolerate multiple disk failures at the same time, and are usually used to ensure the reliability and durability of systems such as RAID and distributed storage. Unlike the traditional copy fault-tolerant technology that saves multiple mirror copies of the original data, the erasure code coding technology encodes the original data to obtain multiple redundant check data and saves it together with the original data. When the original data is lost, read the remaining healthy original data and check data, and use the corresponding data reconstruction method to decode and recover the invalid data. [0003] When encoding erasure codes, the original data is first divided into stripes. Each stripe cont...

AI Technical Summary

Problems solved by technology

The high data update cost will not only affect the performance improvement of some write-intensive storage systems, but also reduce the service life of SSD / NVM and other storage devices with limited erasing times due to frequent update of the check disk, which is not conducive to high-performance SSD / Extensive use of NVM in storage systems

Images