Storage: The Amazon Solution to Long Term Data Archival

Amazon is targeting a new market with the release of their Amazon Glacier service: long-term data archival.  The new service could seeks to revolutionize the archival market.  Glacier archival costs are significantly cheaper than traditional archival methods, and it is offered as a cloud service, it’s use would eliminate the need for on-site archival infrastructure.  It’s called Amazon ‘Glacier’ because it targets the ‘cold storage‘ market.

Glacier is cheap, costin just a penney per stored gigabyte per month, that’s about one tenth the cost of storage with Amazon’s S3.  The drawback though is that storage retrieval times can be on the order of three to five hours, not to mention the time then needed to download that data.  And if you access more than 5 percent of your stored data, access fees begin to kick in.  Once the data has been retrieved, it will be kept available for download for 24 hours and protected by 256-bit Advanced Encryption Standard (AES-256).

TechNewsWorld quotes Altimeter Group analyst Chris Silvia describing Amazon Glacier by saying that  “This is deep freeze, cold data.  This is data that is stored for long-time backups. Not something you need to … access a lot. It is cheap, it is slow, but it removes an operational cost of running your own on-site backup including tape systems.”

Amazon’s Glacier is particularly well suited for the archival of items that require long-term retention and that will only require infrequent access.

Alyssa Henry, Vice President of AWS Storage Services, said that “Today, most businesses rely on expensive, brittle, and inflexible tape for their archiving solution.  This approach requires expensive upfront payments, is difficult to operate and maintain, and leads to wasted capacity and money.  Amazon Glacier changes the game for companies requiring archiving and backup solutions because you pay nothing upfront, pay a very low price for storage, are able to scale up and down whenever needed, and AWS handles all of the operational heavy lifting required to do data retention well.”

Amazon hasn’t described the exact technology being used by Glacier.  Some speculate that the technology is based on tape backup.  But Jack Clark at ZDNet came to these conclusions after some snooping and talking with Amazon engineers:

• Glacier uses commodity low-RPM hard drives
• Amazon is able to only spin up a percentage of all drives at one time, which causes the retrieval delay time
• Once retrieved from these system, it’s moved to a holding area from which it can then be downloaded