Jungle Disk & Amazon S3

I’ve been looking at Jungle Disk for online storage and backup and so far I like what I see. Jungle disk is a good example of a software developer that has built a service on the infrastructure provided by Amazon Web Services. In this case it’s the Amazon Simple Storage Service. Amazon S3 for short.

Jungle Disk runs under OS X, Windows and Linux. I’ve only used the Mac version and everything in this article is related to the Mac version.

Jungle Disk isn’t typical backup software and they rightfully refer to themselves as “online storage”. Still, I think it has it’s use if you can live within it’s limitations.
Some things to be aware of when using Jungle Disk:

  • Files are never deleted when using the automated backup. You must delete them manually or run the “Backup Cleanup” process. This can actually be a good thing from an archiving perspective. It’s a bad thing if you don’t know and are paying for gigabytes of files you no longer need.
  • The Jungle Disk drive is mounted and can be accessed the same way any drive can. Such as in Finder or Explorer. This makes it easy to copy to and from the Jungle Disk instead of using the automated backups.
  • Out of the box you can only have one Jungle Disk bucket on Amazon S3 for Jungle Disk. There are workarounds but they require reconfiguring Jungle Disk to use different ports and directories.
  • File size is limited to 5GB (an Amazon S3 limitation). Many “files” on a Mac are actually “packages” of many files and the limitation applies to the individual files inside the package. For example, an Aperture or iPhoto library may be 6GB, but the largest file in the library is only 1GB so there won’t be a problem.
  • Jungle Disk does not support block-level backups. So if any part of a file changes the entire file gets backed up. Example: I added a single 250KB photo to iPhoto and this alone generated a 560MB backup since the thumbnail and other library information is kept in large files. Since we pay for bandwidth nightly backups could be expensive (well, more expensive than it has to be). Even ignoring cost, upload speed is a limiting factor for me and most people so the 560MB backup took hours, all to get a 250KB change. I could easily run into a situation where a few changes take longer than a night to backup.

A Problem – Or Not

A backup usually caused my cable modem to lock up sometime during the night when the backup ran, which required me to power cycle the modem in the morning. Any backup requiring more than a couple hours has had a problem. This is the issue I’ve mentioned in previous OS Quest Logs. Eventually I found the problem wasn’t unique to JD, although it occurs more frequently with JD than other options.

Pricing

The use of Amazon S3 makes for a slightly unconventional pricing model.

The Jungle Disk software is $20 for a license with free lifetime updates. You can use the software for 30 days as a free trial. The trial is full featured. The license covers all your computers and all OS’s.

Amazon S3 has a rather complicated pricing structure and there’s no free trial. It may be complicated but it’s cheap. You pay $0.15 per GB per month for storage. This is pro-rated for both size and time. For example, I have 19GB currently on S3. My current bill for that this month is $0.79 because I’ve used an amount equal to 5.229GB-month.

In addition to storage you also pay $0.10-GB for bandwidth in and $0.18-GB for transfer out (less if you hit the 10 or 40TB plateaus)

The above charges are fairly easy to figure. But then there’s charges for upload and download requests. A penny per 1,000 PUT and LIST requests (think upload) and another penny per 10,000 GET requests (think download). I’ve uploaded thousands of files and downloaded hundreds while testing. I’ve paid $1.57 for 156K upload requests and $0.19 for 188K download. Just based on volume it appears GET requests are used for listing directories and by the Jungle Disk software.

Due to bandwidth and “request” charges it’s difficult to estimate exact costs, but it appears that 30GB of my normal usage would be about $5/mth which is what Mozy charges for an “unlimited” account. If I was to just leave the data out there I could keep 33GB for $5/mth. If your data changes frequently and your backups are large you’ll pay more for bandwidth.

Amazon S3 does have a cost calculator available.

Bottom Line

Amazon S3 is a great option for online backups and I really like Jungle Disk. The fact that it’s cross-platform is a huge plus and the ability to run it off a USB thumb drive is a big benefit.

It requires a slightly more technical user than other all-in-one backup services. The documentation at the Jungle Disk website provides clear setup instructions.

While Amazon S3 is very low cost you do need to monitor usage to avoid any surprises, unlike other services with a fixed charge.