Getting serious about backing up (Part II)
March 25th, 2008 | by Tim Wilson |Earlier in the month I posted about the strategy I’ve been employing to back up my 24″ home computer. Using SuperDuper! and a 500-GB external drive attached to my Airport Extreme works great, but it doesn’t really solve the my-house-burned-down-and-now-I-lost-all-of-the-embarrassing-pictures-of-my-kids problem. To address that glaring deficiency I needed a way to move my backups offsite.
Enter Amazon S3. In their own words:
Amazon S3 provides a simple web services interface that can be used to store and retrieve any amount of data, at any time, from anywhere on the web. It gives any developer access to the same highly scalable, reliable, fast, inexpensive data storage infrastructure that Amazon uses to run its own global network of web sites. The service aims to maximize benefits of scale and to pass those benefits on to developers.
In other words, when I utilize S3 I rent a tiny little slice of Amazon’s massive infrastructure. The price is certainly right too. The storage is unlimited and costs $0.15/GB per month. Transferring the files to S3 costs $0.10/GB and from S3 $0.18/GB for the first 10 TB. My monthly cost is about $4.
As extensive as it is, S3 is designed to be technically simple. Unfortunately, technically simple doesn’t mean simple for the end user. To really use S3 most people are going to want a front-end tool. I chose Jungle Disk. Jungle Disk runs on OS X, Linux and Windows and costs only $20. When I start the Jungle Disk software, it appears just like any other drive on my OS X desktop. I can browse files and move things back and forth just like any other disk. I can also install Jungle Disk on all of my computers with one license which makes it ideal for storing files that you might need to access from work and home. If you’re concerned about Amazon snooping through your files, Jungle Disk will encrypt all of your data before it gets sent to S3.
I decided to spring for Jungle Disk Plus for another $1/month because I wanted to take advantage of block-level file updates and resumable uploads. Both of these features reduce the total amount of traffic that gets transferred.
All in all, I feel pretty safe at this point. I’ve got regular full-system backups that can be used to restore my system from scratch if I have a major hard drive crash, and I’ve got some insurance for my irreplaceable files. The next step will be increasing the amount of storage available at home. That 500-GB drive is pretty much full. I’m thinking Drobo.
Tags: amazon, backup, drobo, jungle disk, s3
3 Responses to “Getting serious about backing up (Part II)”
By Tomasz Stasiuk on Apr 20, 2008 | Reply
That is pretty much the same strategy I use: nightly SuperDuper! imaging + hourly backups of critical files to S3.
The only difference is I also added an external HD at the office for TimeMachine. So TM at office, hourly s3 with versioning in the cloud, and nightly images.
I may add another hd for SuperDuper! imaging at the office, just to have two full, bootable, images in two separate locations. But… that may be overkill for most people.
By Nathaniel Lindley on May 14, 2008 | Reply
I’ve added one more layer to my backup at home. I use TimeMachine with an External FW for everything on the computer. Then I use Jungle Disk more for a remote storage, to transfer between computers. I could use it to back up Pictures, Music and Movies, but that monthly cost could add up-even at Amazon’s low prices.
My current 3rd strategy is to use CrashPlan. This is a $20 app that you put on the computer you want to backup. Then find another computer at your house as a destination (for free) to backup to. Pick the folders you want backedup and it runs in the background. Here is the nice part for the house-burned-down scenario. Get a friend to install CrashPlan on one of his computers and allocate you some space (or provide your own external storage) and it will back up over the internet to their house. There is no cost to be a CrashPlan destination, and you can backup to as many destinations as you want. Couple great features about CrashPlan:
• all the data is encrypted from the moment it leaves your computer
• It updates itself as they add features and fix bugs
• you can customize what gets backed up when and where
• it does all the network negotiation and such between your computer and the destiation
• their customer service has been really good so far. (Local MPLS company)
• you don’t pay for storage or bandwidth (except your own hard drive cost)
Right now I only have paid for it to backup one machine. My friend has two computers backing up to me. CrashPlan runs on my Mac but the critical Quicken files on my PC just save to a shared folder on the Mac and that gets backed up from there. I have control over the restore process, too. One thing that might be a drawback is that if Comcast actually starts putting a limit on your bandwidth usage you might need to change your habits.
That was much more than I expeceted but I have been really pleased with it so far.
By Tim Wilson on May 14, 2008 | Reply
Crashplan sounds cool. My neighbor and I have been tempted to run some Cat5 between our houses to facilitate some “WAN” backups, but we haven’t gotten around to it yet. Joel, if you’re reading this let me know when you’ve got that conduit in the ground.