Strong opinions, weakly held

Better backups using S3

Today, Paul Boutin has a piece in Slate explaining how easy it is to back up your data using a Maxtor external hard drive. Here’s the case in favor of it:

My solution: Maxtor’s OneTouch III Turbo Edition, which comes with a minimum of 1 terabyte—1,000 gigabytes—of storage. OK, it costs $500, but if you’ve got loads of videos and music you don’t want to lose, that amount is worth it. Maxtor has a reputation for reliable drives at reasonable, if not rock-bottom, prices. The thing certainly looks reassuring—it resembles a metal and plastic vault and has the heft of an oversize brick.

The OneTouch III’s guts are heartening too. It contains two separate hard drives, configured to run in parallel. This is called a RAID configuration, something that’s common in professional data centers but not on desktop computers. The dual drives offer an extra layer of security. By default they’re interleaved as one super-fast device that reads and writes data from both drives at the same time, sort of like drinking soda with two straws. But you can also configure them as mirror copies of each other. That way, if one copy gets damaged, the other will work as if nothing happened. If you’re really worried about your BitTorrent trove, this is the way to go. (No, Maxtor won’t give you a pile of cash if the OneTouch loses your backups. But they do warranty the drive, and “go with the Maxtor, end of discussion” was a typical response when I asked my IT pals for advice.)

I was going to write a blog post comparing the advantages of backing up to your own storage device to those of backing up to Amazon S3, but the more I think about it, the more I think that there are very few advantages to buying an external hard drive these days, assuming you can find the right software to enable you to back up your data to S3 easily. For the $500 you spend on your hard drive, you can use a huge amount of the Amazon S3 service, and you don’t have to worry about one more device hanging around on your desk, paying for its electricity, or replacing it when it breaks or becomes obsolete. Plus with S3, you only pay for what you need. So if you only really need to back up your iTunes library, you pay for a few gigs of storage, not for one full terabyte of storage in the form of a hard drive.

I am notoriously bad about backups, so I’m using this as my excuse to improve myself. Here are the main things I need to back up:

  • My iTunes library (on a Mac)
  • My digital photos (on a PC)
  • My Keychain data (on a Mac)
  • My Password Safe data (on a PC)
  • My blog data (on a server running Linux)
  • My wife’s iTunes library and her email (on her Mac)
  • My wife’s photos (on her Mac)

There’s probably some other stuff as well, but once I have backups working, it should be easy to add more data into the mix.

I’m going to try to create a series of blog posts describing my progress through setting up S3 for backups.

Signing up for S3 was easy. I already have an Amazon account and an Amazon Web Services account. I clicked on the signup button, told Amazon to use the credit card and address I already have on file, and I was done.

Once I had S3 set up, I wanted to access the storage space just to get the lay of the land. I learned about S3Fox, a Firefox addon that provides access to S3 via Coding Horror. I gave that a try, and it seemed to work fine. I also downloaded a standalone application, S3 Browser, which appears to work as well.

I decided to start by backing up my Keychains. Apple provides documentation on where to find them, and uploading them to S3 via S3 Browser was simple enough. Eventually I want to set things up so that I can periodically synchronize that folder with S3. For now, just having a backup of the Keychain is good enough.

Next installment: Backing up my iTunes Library.

Update: Jason Levine on the hidden cost of online backups.


  1. Yikes! We currently back up our home systems nightly and share our pictures, music, and video on our 1TB Buffalo Terastation. If I’m interpreting the S3 pricing right, and using the following assumptions:

    • 3 * 60GB full backups, weekly
    • 3 * 1GB differential backups, daily
    • 200 GB of existing multimedia
    • 10 GB of new multimedia, monthly

    That works out to:

    • 3 * 60 * 4 * $0.20 == $144.00
    • 30 * 1 * $0.20 == $6.00
    • 10 * $0.20 == $2.00

    or a total of $152.00 in monthly transfer costs, plus a one-time cost of 200 * $0.20 == $40, and

    • 3 * 60 * 4 * $0.15 == $108.00
    • 30 * 1 * $0.15 == $4.50
    • 200 * $0.15 == $30.00
    • 10 * $0.15 == $1.50

    or a total of monthly storage costs of $144.00, bringing the total monthly to $296.00 plus a set-up cost of $40. Considering that storage like mine costs just $600 with prices dropping fast, running your own device pays for itself in just two months. Not to mention the world of difference in speed between backing up over a broadband connection versus a LAN and the costs you’d probably have to swallow upgrading your WAN connection. What am I missing?

  2. But if you actually do have 500GB of data to backup, S3 will run you $75/month just for the storage and $100 for the bandwidth to get it up there and another $100 when you need to actually pull it down later when (god forbid) you have to recover from it. That sounds a lot more expensive to me.

    S3 is appealing as an alternative to building and maintaining your own high-availability data center, but for personal backup on the 500GB range, it looks more expensive than its worth.

  3. Can you discuss the security tradeoffs with using S3? Why would I trust Amazon to securely store my personal data?

  4. Yes, there’s definitely an inflection point where it becomes more economical to buy your own hard drive. I’d imagine that if you’re regularly doing full backups of more than a few gigs, the speed of S3 alone would be a dealbreaker.

  5. Another problem is that a lot of DSL providers throttle upload speeds to just a little faster than dial-up. My download speed with my current DSL provider is around 4-5MBps, but the upload speed usually throttles at 60kbps. Insane. With cable, my upload speed was throttled at 128kbps. So downloading is ridiculously fast, but uploads are so slow, it’s not worth it for backups.

  6. John,

    Can you discuss the security tradeoffs with using S3? Why would I trust Amazon to securely store my personal data?

    With utilities like Jungle Disk, you can encrypt the data before you send it to Amazon.

  7. I’m pretty happy with my Mac backup setup, which is similar to your environment. I use SuperDuper to mirror by drive every night to an external hard drive. Not only do I have a backup of my data, but my OS as well.

  8. Hi Rafe, I am also trying to set-up the perfect backup system. To do that I am following the advice of my dad who is a very experienced nerd I can trust.

    His solution for backuping Mac stuff is to use: – JungleDisk for the S3 side – and ChronoSync for the Mac side

    ChronoSync manages backup schedules between disks. And JungleDisk makes your S3 bucket work as a local hard drive thanks to a WebDav implementation. This way you can use ChronoSync to automatically copy stuff to S3. Cool isn’t it?

Leave a Reply

Your email address will not be published.


© 2019 rc3.org

Theme by Anders NorenUp ↑