Yes and No.
Or to put it another way: It Depends!
Another backup bites the dust
Barrientos Island, Antarctica
EOS 5DmkII, 100-400mm (A2_007594)
Today there are a variety of network backup services available, giving you the option to make backups of your data onto network servers via your Internet connection. Probably the biggest name is Amazon’s S3, although there are others. S3’s good in that it encrypts your data and distributes it across many servers, providing a redundant/reliable service. You’re charged a monthly fee based on the amount of data you have in the cloud, and usually also upload and download fees.
Having a copy of your data stored offsite is always sensible (e.g. if your house/office is damaged by fire) and having the ability to access it online can be very convenient. It can be a useful component in your backup scheme, but it’s not a panacea. In this article I explore some of the shortcomings.
Having your files available via a web interface from any Internet-connected machine is good in terms of convenience, and possibly bad in terms of security. You’d better be sure that there’s not just an easily-stolen password protecting your data…
How much do you want to backup? Or restore?
Today Internet connections are generally faster than in the past, and for many of us it seems less of an issue to up/download megabytes (even gigabytes) of data. For those of us still stuck using dialup it’s almost a moot point of course.
As a starting point, which data do you need to backup? If you’ve followed my earlier advice about separating your data into “sets”, you should be able to identify which data to back up where. As photographers we tend to have more (and larger) files than many users (except for some video collections I suppose). But with laptops having 320 GB and even 500 GB internal drives, it’s very easy for all of us to build up large collections of files. The decision of whether or not they’re important files and need to be protected is up to you of course…
Testing the backups
EOS 5DmkII, 100-400mm (A2_007968)
How long is it going to take to make the initial backup?
Consider for example that I have about a terabyte of files I want to back up. I may have a backup copy on external disks already, but maybe I decide I want another copy “out in the cloud”.
My own Internet connection is via an ADSL2+ link that gets approximately 16 Mb/s downstream and 1 Mb/s upstream. So in a perfect situation it’s capable of 1.6 MB/s download, and 100 kB/s upload (big ‘B’ is for Byte, little ‘b’ is for Bit). For most of us this is not a slow connection. If I manage to keep the 1Mb/s upstream link saturated 24/7, it’s going to take me about 115 days to get my terabyte uploaded. Of course, I’ll never be able to keep the link’s saturated (maybe 80% is a better estimate: closer to 150 days). It’s slow and gradual, but it’ll get done.
Thereafter for each 2 GB of new (or modified) data I add to the system, it will take at least 5.5 hours to update the network backup. 30GB would take at least 3.5 days. That could be manageable, depending on how much you shoot (and how large your files are). Today the RAW files from my main camera weigh in at 25 MB each, and derived TIFF/PSD files usually start off at 120 MB before adding any Photoshop layers. Space is getting used up faster each year.
What have I created?
Petermann Island, Antarctica
EOS 40D, 100-400mm (A2_013936)
How long to restore?
But what if I need to recover the files from the cloud? With my ADSL connection I can download files a lot faster than I can upload them: downloading that 1 terabyte of data would take me a bit over 7 days (again assuming that I could keep the line saturated and didn’t want to use the Internet for anything else).
However… even if I had the most expensive plan my ISP offers (and it is one of the better ISPs around) I would be limited to 140 GB per month. Once I go over that I’m throttled back to 64 kb/s. So in fact it would take me more than 7 months to download all my data! Currently I’m on a 65 GB/month plan (which is more than enough for our normal traffic) so it would take me at least 16 months. My ISP only counts downloaded data, so at least the uploads aren’t an issue. Some ISPs count both the uploads and downloads against your monthly allowance (and they tend to max out at 60 GB/month plans).
In some parts of the world there are Internet connections available with no traffic limitations, but the rest of us have to count our gigabytes.
Even if I could download as much as I wanted, it would still take me more than a week to get my 1 TB of data. Compare that with the day or so to get the backup disks that are sitting on the other side of the city and hook them up to my computers.
So for large sets, restoring everything from the cloud is rarely going to be a feasible option. For smaller sets of data yes, and for restoring a few files from a large set.
Avoiding the upload/download problem
Many providers recognise this, and provide services such as Amazon’s AWS Import/Export. You ship them a drive with files on it, and they copy that data into the cloud for you. Or you ship them an empty drive and they ship it back to you with your data. Of course, packaging and shipping your drive(s) to/from Amazon in the U.S. securely does have its own complications of cost and time.
Contemplating a new life
Barrientos Island, Antarctica
EOS 5DmkII, 100-400mm (A2_007655)
Cloud backups can work, but you need to have done your research before you start. For example:
- How much will it cost to store your data?
- How long will it take and cost for the initial upload?
- How long will it take and cost to recover your files after a disaster?
In my own environment I’ve decided to not use network backups. I have large volumes of data to backup, and having established local disk backups (with off-site copies on the other side of the city) I have a system that works.
If a natural disaster strikes that takes out my whole city, I’m guessing I’ll have other things to worry about. Luckily we’re not on top of a major tectonic plate boundary…