Everyone in the storage industry is familiar with RAID data storage technology, but many would be surprised to learn that the term "RAID" goes back to a University of California, Berkeley paper published in 1987.  Over the years, a number of RAID levels have become standardized, each with their own feature set, and there have been a number of additions by various software and hardware vendors in order to differentiate their offerings from others.

What many of these RAID levels have in common is a scheme called parity which has been around since the first days of ferrite-core magnetic memory.  I dare say, if it hadn't been for parity, we would never have had the IBM 360.  The other common element of RAID is that all but one of the data protection schemes remove usable storage space from the pool that might otherwise be used to store data.  Depending on the RAID level, 50% of the usable storage could be used for data protection.

In today's environment where storage capacity is exploding, data integrity is paramount and IT budgets are stuck at 2010 levels, new methodologies that achieve nine nines (99.99999%) or more of data protection while minimizing overhead redundancy are being embraced by the industry and customers alike.

One such data protection technology is called Erasure Codes.  Erasure codes, or more appropriately, forward error correction code, efficiently stores and protects data.  Erasure codes divide a data object (blocks or files), typically through a linear equation, into "m" fragments and recode them them into "n" fragments where n > m.  The relationship between m and n is called the encoding rate.  The key property of erasure codes is that the original object can be reconstructed from a subset of the n fragments.

For the really geeky, hard-core storage architects out there who wish to read more about Erasure coding, we at storageFOUNDRY recommend a white paper called Erasure Coding vs. Replication: A Quantitative Comparison, from Hakim Weatherspoon and John D. Kubiatowicz of the Computer Science Division, University of California, Berkeley.  It can be found at the following link: http://www.cs.rice.edu/Conferences/IPTPS02/170.pdf

Suffice to say, we here at storageFOUNDRY do not see an immediate end to RAID.  It has proven itself for decades but as storage expands beyond a few terabytes to tens and hundreds of petabytes, RAID and the corresponding downside to RAID rebuilds is no longer the appropriate choice for data protection.  Erasure coding is the future.