Opinions: Battle of the Backups

Part 1 - Cloud backups & a Chat With CloudBerry

For about a year I've been using Jungle Disk to back up the essentials stored on my file server. Jungle Disk is a division of Rackspace, but they support storage on Amazon S3. When I first begun using their services, you were required to sign up to Amazon's Web Services, create an account for S3 and establish a "bucket" (sub directory/storage folder), and then connect that account to the Jungle Disk client software. I liked this because I thought it would mean I could switch to a competitor's software client easily and continue to access my data.

More recently, as well as beginning to support the Rackspace Cloud and re-branding "buckets" as "cloud files", Jungle Disk has started selling S3 storage as a part of the package so that there is no longer any interaction with Amazon required - they manage everything for you. This can be seen as removing signup hurdles, but it also means that a small degree of flexibility has been removed and that potentially any accounts created through Jungle Disk are in a proprietary format. Data could prove difficult to get at through other software services - how do we know the S3 account details to get at the bucket? Additionally, is it wrong to assume that Jungle Disk will slowly move away from great support of S3 and focus on the Rackspace Cloud? Should I even be worrying about this?

Ubuntu's next release "Karmic Koala" will be pushing applications to Amazon EC2, and Ubuntu One is already harnessing a proprietary cloud for data synchronisation. I'd like to be able to use Ubuntu community developed packages in the future for accessing data on my cloud(s), but I would also value being able to get at my data through any other platform's software. I would hardly describe the cloud backup software market as saturated, so I decided to consider some different software products. Twitter proved to be a great place to start my search for alternative S3 backup software. A couple of tweets between some friends about Amazon S3 command line tools, and I soon found myself being followed by a handful of companies offering Amazon S3 related services.

I got in touch with the team from CloudBerryLab.com. They have produced an S3 File Explorer which is available as Freeware or a Pro version, a Twitter tool for IE and Firefox that allows users to highlight text from a web page and tweet it, and they are currently working on the beta of CloudBerry Online Backup which will be a backup manager for S3. I've been asking some questions about the CloudBerry Online Backup client and how it will compare with Jungle Disk once out of beta.

First on my checklist was cross platform support as their beta release is for Windows. Fingers crossed user feedback will encourage a Mac/Linux version to appear on their roadmap eventually, but for the moment they want to focus on achieving the best possible solution for Microsoft users. I'm happy with that answer, strategy means saying no. I've only explored Jungle Disk on Windows, whereas I've had greater exposure to alternatives on Ubuntu such as S3FS, so their product needed to be tried.

CloudBerry's software will eventually upload to whichever cloud vendor takes your fancy. Every aspect is still in private beta, but support for Microsoft Azure and Sun's Cloud storage is coming. The team are keen to point out that any file that CloudBerry Drive put on your S3 cloud will be accessible from any other S3 browsing, syncing or backup software, so Windows users are free to choose how they manage their data. That's quite sensible - you pick a cloud, you invest time and bandwidth putting your data on it, and then if your backup needs change you can switch client software to manage your data differently.

If you've got a large volume of files to shift to Amazon S3, CloudBerry will help you create the manifest file needed for data that you snail mail on a hard disk to Amazon as a part of the AWS Import/Export programme. They are also working on a web portal - cloudshare.me - which will allow you to browse any publically readable files in an S3 bucket. I use the MyJungleDisk web portal to access my private data and upload to my S3 bucket from other machines that haven't got the software installed, all over https, so I hope the browser based tools CloudBerry make will ultimately support that type of functionality too.

So I fired up XP and installed the Cloud Berry Online backup client. The backup wizard guided me through connecting the software to a bucket, and then let me dictate precisely which folders and file types to watch. I was impressed by the ability to compress files, especially the choice to opt-out of compressing certain file types. There was a section for enabling encryption, dictating your own added security for files with a drop-down selection of algorithms - AES, DES, RC2 and 3DES. Finally, the wizard let me extend the option to purge old versions beyond the default of one day and establish a schedule for the backup operation.

The tabbed interface made light work of checking up on backup plans, and the history tab showed exactly what was sent up to the cloud and when. It also meant there was a lot of blank space under most of the tabs once I maximised the window, but I hope this means that I'll soon be able to manage more than one backup plan (potentially on different cloud vendors) without the view becoming cluttered and unreadable.

I then took a look at the S3 Explorer Pro, and was pleasantly surprised to find that it was also an FTP client. It was extremely simple to set up, just enter the bucket details or FTP location and you're connected instantly. I found it a bit odd that double clicking on a local file actually launched it rather than copying it to the remote location. Everything appeared standard, represented in the traditional WS_FTP / FileZilla two pane view, except that I was able to open new tabs for connecting between local and remote locations in different configurations. It made me chuckle that I could copy files from my webserver (connected via FTP) to my S3 bucket, so I decided to let it store a complete backup copy of my blog on S3, just because I could.

If you're a Windows user, take a look at CloudBerryLab.com