Since data loss is never a good thing, I have a fairly robust backup system in place that involves backing up my most important data daily to external hard drives and then weekly to optical media. But being a bit paranoid, I’ve also always wanted some sort of way to back up my data to an online service so I could have a more robust, off-site backup solution as well.
There are a lot of companies in the online backup business these days, such as Mozy. But I have two sets of concerns about these services. First, as some users have complained, sites like Mozy are often not so great when it comes time to actually retrieve the data you’ve backed up. Second, Mozy and others use the sort of model that web hosts use — charging very low monthly prices betting that most users will only us a fraction of the service which will offset the folks who have lots of data to back up. Mozy, for example, advertises unlimited backups for $4.95/month.
I can see using those sorts of services for small backup sets, but I have about 300gb of data to backup.
Since October, I’ve been using Jungle Disk for on-line backup. Jungle Disk is basically just a backup front end for Amazon’s S3 storage utility. So all of my files are actually residing on Amazon’s servers, and the Jungle Disk application takes care of uploading all of my files and keeping my backup up-to-date.
The application itself costs $20. There is an optional Jungle Disk Plus service which basically uploads your files to a separate server before transferring to Amazon which allows things that S3 doesn’t support yet, such as large file resumes or block-level updates. Thats $1/month.
Storage charges, bandwidth and other charges are set and billed by Amazon. Amazon currently charges 15 cents/month/gigabyte for storage and 10 cents per gigabyte of bandwidth used in transferring files. So that 300gb of data I’ve currently got sitting on their servers is going to cost me $45/month. I consider that cheap given the peace of mind it gives me, but your mileage may vary.
All files are encrypted using 256 bit AES. I created a mind boggling long passphrase. The Jungle Disk Plus service does offer some optional features such as web-based access to your files that, IMO, reduce the level of security too much so I simply don’t use them.
One of the big issues, at least in the United States, is upload speeds since most of us do not have anywhere near the bandwidth leaving our homes that we do for downloads. My cable company recently offered 10mbs down/1mbs up, and I signed up for that before starting to backup all my files using Jungle Disk.
Connecting directly to Amazon’s S3, I typically got 720kbs up — or about 324 megabytes per hour. After Jungle Disk launched their Jungle Disk Plus offering, most of the time my connection maxed out so I was getting 990-1024kbs up or about 461 megabytes per hour. When you’ve got 300gb of data to upload, that’s still a very long upload. In all, it took about two months for my backup job to finally upload everything. Now, Jungle Disk simply checks at 2 a.m. to see what new/changed files have been added and takes care of those automatically.
There are a lot of things I really like about Jungle Disk, but probably the biggest feature as opposed to all of the other services I’ve seen out there is how easy it is to retrieve files. Forget using some special software or going through some convoluted process. Jungle Disk simply maps your S3 storage as a network drive. From Windows Explorer I can navigate through my stored files, find the one I’m looking for, and simply copy and paste to retrieve it. The software does come with a more robust feature for retrieving large amounts of data, say in the case of a catastrophic local data loss, but it is nice to simply be able to retrieve data through a standard file manager interface.
Jungle Disk also does a nice job of archiving previous versions of changed files and deleted files. Archiving older versions of files is completely optional, however, and I suspect many people will disable it given the potential costs.
Overall, my experience with Jungle Disk/Amazon S3 has been outstanding. The only thing keeping from uploading all 20TB of data I’ve got archived in one form or another is a) bandwidth and b) price. But over time, the bandwidth available to the home is only going to increase, while storage costs are only going to decline. For now, this is a great solution for off-site backup of data that would, at least in my case, be impossible to recreate.