Jungle Disk – online backup using Amazon S3

I’ve mentioned Jungle Disk in a few tweets before, but now I’ve got my first few bills from them, I thought I’d write a longer post about what they’re like.

For the uninitiated, Jungle Disk is an application that lets you use Amazon’s S3 storage as an online file store and backup service. You can set the app up to point at a directory of your hard-drive and forget about it – letting the client backup your files in the background. And you can get the client for Windows, Mac and Linux.

I started using it last December for digital photos. I had over 20 gigs of photos and video clips – photos from Uni, photos of my daughter’s first few years, photos from my wedding… and tons more.

And I wanted to put them somewhere safe.

There are alternative options, but none of them appealed:

  • Backing up to DVDs was an option, but DVD burning is still new enough that I don’t completely trust that they will last. I have piles of CDs that I burned while at Uni and none of them are readable any more. For DVDs to be a good long-term backup medium, you’ve got to trust them. And I don’t.
  • Backing up to USB hard-drives was an option, but I’ve read that the failure rates on USB hard-drives over time is fairly high (Jungle Disk point at figures like a 15% failure rate over 5 years) – so again, not sure that I trust it for that.
  • Backing up to network-attached storage was an option – and one we nearly went with. But it’s still on-site. And at the risk of sounding a little morbid, if my house was to burn to the ground, my photos would be lost. Most things are insured and/or replaceable, but my photos aren’t.

And we’re entering the age of cloud computing, right? So where better to use as a backup store? I can store my photos on Amazon’s servers, and let them worry about how to store them safely.

So what does it cost?

That was another factor – Amazon S3 compares well to the costs of buying your own USB or network-attached hard-drives if you were to replace your drives every two or three years to reduce risk of drive failure.

You pay a one-off fee to Jungle Disk of $20, then ongoing costs to Amazon of:

  • $0.15 per gigabyte-month of storage used
  • $0.10 per gigabyte of data uploaded
  • $0.18 per gigabyte of data downloaded
  • $0.01 per 1,000 upload requests
  • $0.01 per 10,000 download requests

As abstract figures they look pretty good, but until you’ve tried it you don’t really know how it will translate into real-life costs. (E.g. how efficient is their backup client in terms of the number of directory listing requests it makes when deciding what needs to be backed up)

This is how their charges translated into actual costs for me. Note that all charges quoted are in US dollars.

December 2007
I signed up with Jungle Disk in the last few days of December – after Christmas in the runup to New Year. So this bit hardly counts – I’d only uploaded a few gigs of photos by the end of December.

It was a little surprising to get my first bill a few days after signing up, but this was just a fluke of timing. Bills all come at the start of each month for the previous month’s usage.

$0.15 per GB-Month of storage used 0.146 GB-Mo 0.02
$0.10 per GB – all data transfer in 3.262 GB 0.33
$0.18 per GB – data transfer out 0.009 GB 0.01
$0.01 per 1,000 PUT or LIST requests 25293 Requests 0.25
$0.01 per 10,000 GET requests 15617 Requests 0.02
Total 0.63
tax 0.10
TOTAL $0.73

January 2008
I uploaded the rest of my backlog of photos over the first couple of weeks of January. It seemed to take forever! A couple of weeks of leaving my computer constantly. Some of this is probably a factor of the upload rate from my broadband provider, though.

The bill was a bit high, but about half of it is the data upload cost for a backlog of years of digital photos, so it’s probably more of a setup than an ongoing cost.

$0.15 per GB-Month of storage used 15.789 GB-Mo 2.37
$0.10 per GB – all data transfer in 19.812 GB 1.98
$0.18 per GB – data transfer out 0.191 GB 0.03
$0.01 per 1,000 PUT or LIST requests 35627 Requests 0.36
$0.01 per 10,000 GET requests 26061 Requests 0.03
Total 4.77
tax 0.84
TOTAL $5.61

February 2008
The first month where I only uploaded new photos taken that month. This is the sort of bill I can start to expect each month from now on, increasing by 15 cents each month for each GB of new photos and videos that I add.

$0.15 per GB-Month of storage used 23.087 GB-Mo 3.46
$0.10 per GB – all data transfer in 0.144 GB 0.01
$0.18 per GB – data transfer out 0.017 GB 0.01
$0.01 per 1,000 PUT or LIST requests 1233 Requests 0.01
$0.01 per 10,000 GET requests 1144 Requests 0.01
Total 3.50
tax 0.61
TOTAL $4.11

There are other good bits worth mentioning – I signed up to Jungle Disk Plus early enough that I got it for free for 12 months. Normally $1 a month, this gets you web access to your Amazon S3 storage. If I wanted to be able to download one of my backed-up photos from any Internet-connected machine, I can. I’ve not used this yet, but it’s a nice idea.

Problems? I’ve not had many. Migrating your settings to a new version of the Jungle Disk client when they come out with a new version is a bit clunky – seems that you have to re-enter your settings from scratch. It’s enough of a pain that I’m still a couple of versions behind on one of my machines. It’s not a major issue, but it’d be nice if they made the migration path a little easier. If you use the USB version of the client, you have to make sure you don’t overwrite jungledisk-backup.xml and jungledisk-settings.ini with the empty copies in the zip file for each new version. Thanks to JungleDave for pointing this out in the comments below!

And Andy mentioned that he has had some problems when uploading files from a Mac but downloading from Windows. I’ve used the Jungle Disk clients on Windows and Linux, but for separate uses only and not tried mixing them – so I’ve not seen this myself.

All in all, I’ve been very happy with it. I’m certainly happier knowing that my photos are somewhere safe. And for two or three quid a month, this feels like good value.

Tags: , , ,

18 Responses to “Jungle Disk – online backup using Amazon S3”

  1. Andy Piper says:

    So I think the problem is that JungleDisk is just another storage device to the OS, and the Mac stores files slightly differently so the directory listing has a lot more metadata in it when you view it on a Windows machine… or, maybe, JungleDisk is storing it’s own metadata. Actually the latter is more likely. I tried to use it one time to store a copy of an Eclipse plugin in a zip file for a colleague to download on-site… so I copied the file up via JungleDisk and sent him the link to the file in my S3 space, but when it came down he couldn’t unzip it. So I’m not sure what went on there.

    I have been using S3 Bucket Explorer which is a bit of a clunky GUI but seems to work nicely between platforms and results in “cleaner” directory / bucket listings on S3 via the web interface. I can see the attraction of JungleDisk, though. It would be cool to have a well-supported/integrated version of an S3 module for FUSE, too – for Mac and Linux users anyway 😛

  2. dale says:

    Thanks for that, Andy – it’s interesting to hear.

    Did you raise it on their forum? Looks like they are fairly responsive to questions and comments.

  3. JungleDave says:

    Thanks for the informative review.
    Regarding losing your configuration settings when upgrading – that shouldn’t happen. Are you using the “USB” version perhaps? The USB version ships with empty configuration files in the ZIP, so if you just overwrite your entire folder with the new version you’ll overwrite your configuration as well. In the normal installers this doesn’t occur as the config files are stored under your user profile.

    Regarding the Mac metadata – that is the finder info and resource forks, which are stored in the Apple Double format, the same way Finder does it for any other non-HFS network volumes.

  4. dale says:

    JungleDave – I do use the USB version, and that was the answer. I’ve updated the post accordingly. Thanks very much for your help!

  5. m_s says:

    Thanks for this – have you considered Mozy? I’ve used this for the last long while, but recently had to reinstall my operating system (OS X Leopard), and found that because I did this from an external drive backup instead of the Mozy version (because downloading would have taken ages…), Mozy now thinks that the files on my machine are different to the ones they have (apparently caused by a difference in data signatures, whatever that means), and so I’m facing another two weeks of uploading everything from scratch! Mozy costs $5/month for unlimited storage, so it looks like it’d cost me about the same to use JungleDisk… Any opinions?

  6. dale says:

    Hi – nope, I’d not heard of mozy before.

    It looks interesting – in terms of price alone, the flat fee is certainly competitive. I’m getting near to the point where I’ll be spending $5 a month with JungleDisk soon.

    I’d be interested to find out more about it.

  7. Harry Gantley says:

    $5 a month is nothing. I ran with a number of free services for a while to test the service. My worst experience (and I wont mention the provider here) involved nearly unlimited upload. This went swimmingly well till I decided to do a large test restore and I found out that while my upload allowance was large I had to “upgrade” to be permitted to download any usable quantity of data. Since then I have found numberous serious issues with all of the the free service offerings with support being the principal problem. Support in a data outage scenario is essential so I dont think free or nearly free advertising based services offer a future for commercial online backup. I am currently running four trials with comercial online backup providers. Thus far http://www.backupanytime.com and http://www.clunkclick.net are on top.


  8. Francis says:

    Hi Harry
    Good post. I am with backupanytime.com so its nice to see that despite not being a techie I made the right choice!
    All the best

  9. Leem Armitage says:

    You have made an interesting point Harry. Everybody talks about the backup, how much it costs, how much I can upload etc, but very few thinks about the retore process which to me is the most important part. Afterall there is no point backing up your data if you cannot get it back quickly.

    For this very reason alone I decided to use a company called PerfectBackup. The main difference is they offer a facility called instant restore. If you have lost all your data, give then a call and they will send it to you on a suitable hard drive within 24 hours free of charge. On top of that I think they are the only online backup provider who offers a ?1,000,000 data restore guarantee. They have an interesting/informative page at http://www.perfectbackup.co.uk/online-backup

  10. Liam Naughton says:

    1,000,000 data restore guarantee is about as low as it gets. In essence this means that if perfect backup fails, all of the clients will have to share the 1,000,000
    If they only have 1000 clients that would mean a paltry 1000 compensation average per client for losing all their data. I dont know why perfect backup traverse the forums boasting this. My provider (US based and that is where to keep your data due to hippa complaince) has a 10 million guarantee and when I queried it they explained it costs less that 10 grand due to their policy and procedure adherence. By the same token it looks like perfect backup are spending about a grand a year to insure client data. My personal car insurance costs nearly as much.
    By the way, 24 hours is not instant restore.

  11. Dear Mr Naughton, Dan here from PerfectBackup.co.uk. If you check our insurance terms and conditions you will see that each and every client is covered to the value of 1,000,000 sterling this figure is not shared between clients. Even clients who use our free of charge service are covered by the same insurance. To use your statement, if we only have 1000 clients and we lose all client data you will then finding our insurance company paying 1,000,000 x 1000 compensation. I am not at liberty to state how much our insurance costs , however I feel sure you will now understand our insurance cost more than a 1,000 per year.

    Regarding our instant restore facility, I am afraid Mr Armitage has our solutions a little mixed up. Our software can create a local backup as well an offsite backup. If data is lost from a server then it can be instantly recovered from a local backup without the need for download. This is not to be confused with our data delivery service. Should all local data be lost then it will be delivered back to you within 24 hours on a media of your choice. e.g. If you need to restore 100GB from your American provider then the download will take between 1 and 2 weeks with a good internet connection. To make the restore process as quick as possible we will send the data you.

  12. Thanks for excellent review. I have been thinking of using online backup service for some time now. But was not sure of the cost, and how much hassle it is to maintain your backups. Your review has given detailed breakup of initial and ongoing costs of using Amazon S3.

    About Jungle Disk Plus service, I found that apart from web based access, it provides file download resume service, and block transfer for only changed portion of files. Both these will be very useful if your backup contains large files. It will save bandwidth costs too.

  13. Robert Lindeman says:

    I read through these posts. I know its an old story but what I am interested in is the perfectbackup guys 1000000 per client insurance claim. I had a look at the site and couldnt find the page explaining it. I am in the business. Where do you get insurance in the billions for online backup? It doesnt make any financial possibility when compared to my premium. I wonder if Dan from Perfect backup is wrong. I am not for a second saying he is intentionally misleading us. I am with a hoskins agent and to get 1mil per client cover would make the busieness non viable. To add to the anomoly I charge more than perfect backup so have more income per client to devote to insurance. So, are the rest of us being ripped off by insurance companies or does Dan need to read his policy? I dont know, more detail Dan please. if the outcome of this is that you are correct and I can find out where to get that level of cover for my clients I will be very happy.

  14. […] blogeissa (blog 1 ja blog 2) kerrotaan hyvin eri (DVD, USB kovalevy, koti linux backup serveri) vaihtoehtojen hyva ja huonot […]

  15. Has any one used sparebackup.com?

  16. David says:

    I used sparebackup.com some time ago, they seamed ok, I ended up moving to http://www.blockbackup.com based on price.

  17. Kirti says:

    You can also try User Interface , which is called Bucket Explorer. Technically, its probably the best solution out there to transfer files to S3 and when you want to be in control, as it has easy to use UI, which is built on top of the robust JetS3t API.

    You can use Bucket Explorer as a simple FTP tool or a backup tool for Amazon S3 or you can use it to:

    1. Browse buckets and the files stored at Amazon S3.
    2. Upload and download files in multiparts, to and from Amazon S3 buckets.
    and many more..

    Bucket Explorer works on every OS where Java is supported. It uses Amazon’s ETag and its own SHA-1 hash combination to make sure that a file is never transferred again to Amazon S3 if is it not changed, to save the bandwidth costs and time.