Amazon launches Glacier cloud storage, hopes enterprise will go cold on tape use

Amazon launches Glacier cloud storage, hopes enterprise will go cold on tape use

Summary: The tapeless Glacier service sees Amazon Web Services target on-premise tape systems with a redundant cloud storage technology, though to win business it will have to battle enterprise concerns about the stability of its cloud.

TOPICS: Cloud, Amazon, Storage

Amazon has released Glacier, a data archive service for enterprises and small businesses that encourages companies to give up their tape systems and move to the cloud.

AWS Glacier
Amazon Web Services has launched Glacier, its new tapeless storage service. Image credit: AWS

The cloud-based technology, launched on Tuesday, sees Amazon take on tape, the leader in enterprise storage. Files stored via Glacier have an annual durability of 99.999999999 percent, according to Amazon; this means that if a company uses it to hold 100 billion objects, it can expect to lose one each year. There is no limit to the amount of data businesses can place in Glacier.

"Using Amazon Glacier... unlimited archival storage is available to [AWS customers] with a familiar pay-as-you-go model," Werner Vogels, the company's chief technology officer, wrote in a blog post on Tuesday. "The service redundantly stores data in multiple facilities and on multiple devices within each facility."

The service aims to compete with on-premise storage systems on data redundancy and price. The starting cost of Glacier is $0.01 per gigabyte per month. The archiving of media assets, research and scientific data, and enterprise information, as well as magnetic tape replacement, are ideal use cases for Glacier, according to Amazon.

A replacement for tape

Asked what IT equipment Glacier uses, Amazon told ZDNet it does not run on tape. "Essentially you can see this as a replacement for tape," a company spokesman said via email.

Instead, Glacier runs on "inexpensive commodity hardware components", he said, noting that the service is designed to be hardware-agnostic. This suggests the system will be based on very large storage arrays consisting of a multitude of high-capacity low-cost discs.

But to convince businesses of the value of the technology, Amazon needs to show that the service is resilient and accessible. This is not a sure thing, and time will tell. Each time Amazon has a major outage, it makes it easier for an organisation to turn away from the cloud, as the chance, however slight, of critical data loss can hurt businesses.

There are other factors that might crimp the service's appeal for enterprises. Data stored in Glacier takes three to five hours to access. Furthermore, though customers can retrieve five percent of that data for free each month, after that they are charged a retrieval fee. This charge comes with additional data-transfer fees — it costs nothing to load data into Glacier, apart from the bandwidth a business pays their own ISP. However, to take out anything over a gigabyte per month costs money, starting at $0.120 per gigabyte.

Amazon is encouraging customers to use Glacier for archival, rather than short-term, storage. If information is deleted within three months of being uploaded, there is an "early deletion fee", it said.

Boon for enterprises

Aaron Levie, chief executive of enterprise data storage and collaboration service Box, reacted favourably to the announcement.

"[Glacier] seems like a huge boon for enterprises and the cloud," Levie told ZDNet. "Amazon is continuing to show it's going to be the most disruptive player in cloud infrastructure."

Meanwhile, commenters on popular developer message board Hacker News praised the technology; user 'kdsudac' noted that Glacier is an order of magnitude cheaper than Amazon's mainstay storage service S3, while buro9 praised it for its redundancy.

In the future, we can expect Glacier to gain integration with Amazon S3 to let customers "seamlessly move data between Amazon S3 and Amazon Glacier based on data lifecycle policies", Vogels said.

Glacier is available now from datacentre clusters located in Europe (Ireland), the US (North Virginia, Oregon, North California) and Asia (Tokyo).

Topics: Cloud, Amazon, Storage

Jack Clark

About Jack Clark

Currently a reporter for ZDNet UK, I previously worked as a technology researcher and reporter for a London-based news agency.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Cost forecast for Glacier vs S3

    We ran an experiment with Glacier on and found that if you start with 100GB then add 10GB/month, it would cost $102.60 after 3 years on AWS Glacier vs $1,282.50 on AWS S3! Very cool!
  • Interesting separation from S3

    You could already use S3 for exactly the same purpose, of course, but more expensive - and the "Reduced Redundancy Storage" option in S3 was presumably aimed at a similar use case. I wonder why they've introduced this third variant as a new product, rather than a third storage option within the existing S3 setup - getting the benefit of all the existing knowledge, tools and partnerships?

    Some of the online storage companies, like Dropbox, use S3 behind the scenes: I imagine they'll be very keen to use this system instead for at least some data.

    We actually had a similar setup at Cambridge University years ago named Pelican, intended for exactly this: cheap, slow, bulk storage for personal backups and archives. A tape-disk hybrid I think, available internally over SSH/SCP/rsync, plus FTP. Very useful in its day.
    • Slower and certain archive-targeted restrictions

      There is evidence they've implemented it differently from S3. From the customer POV, there are restrictions that are unacceptable for S3 (even RRS) but totally fine for medium to long-term archiving:

      "You can retrieve up to 5% of your average monthly storage (pro-rated daily) for free each month. If you choose to retrieve more than this amount of data in a month, you are charged a retrieval fee starting at $0.01 per gigabyte. In addition, there is a pro-rated charge of $0.03 per gigabyte for items deleted prior to 90 days."

      Also, it can take "3-5 hours" to retrieve data.

      So the different branding makes sense, but clearly there will be demand for integration with S3.
  • Glacier?

    That's a funny choice of branding. Doesn't "glacier" imply really, really slow?
    • Exactly!

      Glacier is the perfect word to describe it.

      Glaciers grow through slow accumulation, and can "freeze" things for many years. Glaciers are also really big.

      Speed is not what Glacier is offering, since it takes hours to retrieve your backup.
  • It's still the cloud

    This “Glacier” is still subject to all the faults of any cloud service: reliability issues, security, hackers, government agencies taking the data, etc. Plus the fact that a business will be paying possibly considerable bandwidth costs to upload large volumes of data.

    "inexpensive commodity hardware components" That certainly sounds like cheap hard disks. And we all know that hard disks never fail.

    Let’s see now, you pay bandwidth to upload your data, then pay them per gigabyte per month, then you pay to download your own data (over one gig per month), plus you pay a per gigabyte fee if you don’t keep your data with them for at least three months.

    So you pay all these fees to store your important company data off somewhere in the cloud, just to save the cost of storing it on tape in your own facility? Gee, tapes must be expensive. I’d much rather pay a monthly fee to put my data someplace that just possibly I might not be able to get it back from, and where it will be available to third parties, than to pay one time for tapes that can be used repeatedly and have a long shelf life. (
    • Looking into this...

      @Doc.Savage Funnily enough I have some concerns about the price as well. However, rather than doing a back-of-napkin calculation I'm trying to get some solid figures from tape storage vendors so we can do a fair comparison. Will be doing a separate post when/if this information trickles through. Thanks for reading and commenting.
      Jack Clark
      • Looking a bit further...

        The Glacier service has a 3-5 hour delay in retrieval. I am pretty sure it is using tape, but tape is completely transparent to the end user. Compare this same feature to Infinidisc, where you not only get an intelligent cache appliance at the customer site, you get storage as low as $.04/GB/mo stored 2 copies. No delay in data retrieval, no deletion penalty, bandwidth purchased based on dedicated needs basis. There are good alternatives out there, ultimately customer requirements will dictate the solution.
    • There have been services like this for some time

      Now that they attach the phrase Cloud to it as a way to make it timely, it's all of a sudden in the same category as utility computing. I vote we kill the term Cloud. I say enough of the madness. I say call it what it is, be it utility computing, or offsite storage, or web hosting.
  • Dear Mr. Clark,

    It is not just the cost of tape vs. cloud. There are many other factors about the cloud that make it an undesirable choice.

    • So you're saying

      He has clouded judgement?
      • Sorry...

        Could not help it. I have poster tourrets...
        • Fogged Up Decisions

          From a distance, a cloud looks pretty. Up close, its just fog.

          For a business that already has a tape backup in place - what benefits would make you rip out your functional tapes and tape drives, and install / upgrade your Internet network connections ?