Review: personal backups to Amazon S3

With the announcement of CloudBerry Online Backup, there are now two major commercially supported backup programs for consumers that offer the option of storing your files on Amazon’s S3: CloudBerry and Jungle Disk. In this review I’ll have a closer look at both options; but I’ll start with a quick intro about why you would like backup to S3 and the various options you have for making these backups.

Jump to: [Intro] – [CloudBerry Online Backup] – [Jungle Drive] – [Summary/Conclusion]

Why choose Amazon S3 for backups

There are several reasons to use Amazons Simple Storage Service for hosting your backups; one of the most important features is that it’s an open system. Amazon is a company that will be around for the next couple of years, and you can use a variety of methods to access the files stored there. Jungle Drive even provides a decryption library that you can use to access the files stored without using their software.

In addition to that, both products in this review have a very simple billing scheme. You know exactly how much money the company producing the software is making; all bills for the software you’re using depend on how much space you have in use. You are completely in control of your backup costs, without having to conform to “acceptable use” policies, or having to upgrade to a more expensive subscription once you go over a certain storage level.

Jungle Disk has some more advantages listed on their “Why It’s Better” page.

CloudBerry

The CloudBerry online backup software is developed by CloudBerry Lab; it’s a small company with lots of experience producing utilities that make working with Amazon’s webservices easier. The product is supposed to have a one-time cost of $29; after that, all you’ll pay are the storage fees for Amazon’s S3.

The Online Backup solution from CloudBerry is currently in beta; the major backup features such as scheduling, defining multiple backup sets and file versioning are all there. I’ve tested Beta 3.

Jungle Disk

Jungle Disk is the established player in this field; they have been offering backup software for S3 for years, and were recently bought by Rackspace. Their software now offers two cloud storage providers; Amazon S3 and Rackspace Cloud Files.

In addition to the backup features, they also let you mount your backup disk as a normal hard drive; this lets you easily copy files to and from S3 and Cloud Files storage pools.

Contrary to the CloudBerry software, you pay a monthly $2 fee for using Jungle Disk. This makes them more expensive to use after the first year.

Not reviewed, but interesting: Maluke S3 Backup

A third alternative might become available in the future; Maluke Co offers a beta of a free S3 backup utility that appears to offer the same backup features as the aforementioned products. Unfortunately, the current beta of a year old, and I couldn’t get it to work on my test machine (see below). The developer is working on it again, so we might see a third option for S3 backups appear later this year.

Test setup

For the tests, I used a virtual machine running Windows Vista (32-bit), with 2 GB of RAM, and a 2 GHz CPU. Disk storage was on a RAID-10 SATA array, while the network connection was a dedicated 100 Mbps link to make sure the network was no bottleneck.

I used a storage bucket on S3 Europe (my test machine is in Europe, so this ensures that latency and bandwidth to Amazon’s servers won’t influence the test results).

To test the performance of the backup utilities with different types of files, I created the following test files:

• A set of 10 files, 100MB in size each, containing just null characters. That means this backup set contains a gigabyte of files that compress very easily.
• A set of 10 files, 100MB each, that contain totally randomized data. These are used to test the performance of both products with large files that can’t be compressed.
• A set of 1000 files, 1 kilobyte each, with random data; these are used to test the performance with backup sets that contain large amounts of mostly small files.

All test files were generated on a Linux system, with the following commands:

for i in {1..10}; do
  dd if=/dev/urandom of=random/random_100M_$i bs=1k count=102400;
done

for i in {1..10}; do
  dd if=/dev/zero of=empty/empty_100M_$i bs=1k count=102400;
done

for i in {1..1000}; do
  dd if=/dev/urandom of=smallfiles/small_$i bs=1k count=1;
done

Jump to: [Intro] – [CloudBerry Online Backup] – [Jungle Drive] – [Summary/Conclusion]

Related posts:

1. CloudBerry offers beta of new backup utility
2. New KeepVault release gets Windows 7 support
3. Free full system image backup software
4. McAfee offering branded Mozy online backup client
5. Backup for dummies