With the usage of high-capacity, multi-terabyte hard drives, storage administrators using RAID face the potential prospect of waiting days to restore data in case a drive fails. But according to W. Curtis Preston, founder of Truth in IT and BackupCentral.com, erasure coding is a possible alternative. In this video, he explains how erasure coding provides data protection while maintaining performance. Some organizations dealing with huge data sets rely on erasure coding to as an alternative or complement to traditional backup.
Erasure coding breaks data into segments, expanded and encoded with redundant data pieces and stored across a set of different locations, such as disks, storage nodes or geographic locations, according to Whatis.com. Preston said this protects against data loss in case of media failure.
"I can put in a replacement array, and I'm not rebuilding; I'm copying data. That's the beauty of an object-based storage system," said Preston. "I'm not building parity; I'm just copying the stuff from different places, which is very different than the concept of using parity to rebuild data. That's the concept of an erasure coding system."
Preston gave an example of an erasure-coded setup in which data is broken up into nine segments to be distributed throughout an array.
"We're going to calculate three parity segments for those nine segments … Part of the logic is to distribute these segments as far out as we can. And then I can read that object as long as I have nine -- any nine -- of those 12 objects. And it's the same performance … whether I'm using a parity object, or an 'original' object, if you can use that term. And that means I can survive three simultaneous failures of media," said Preston. "The idea is that I took a piece of data, I chopped it up into n segments -- that number is definable … I chose nine and then I find n number of parity segments, and that got me nine or 12 chunks all together, and I can throw those out anywhere in any 12 pieces of storage."