Backing up your photos - a guide for photographers

backup Oct 18, 2009

One of the marvels of digital photography is the fact that taking photos no longer requires an investment of cash every single time you want to see how your photos turned out. One thing I tell my students all the time: TAKE PHOTOS! With digital, there is no recurring cost to practice.

This will lead you to taking seventy-seven bazillion photos and having to store them all somewhere.

WTD121.jpg

Now, there are entire books dedicated to the subject of digital asset management, but in this article, I am going to talk about the most potentially expensive one: backups.Back in the film-only days, asset management meant having a good secure (fireproof) filing cabinet and some form of catalog to keep track of what photo was on what strip of film. The advent of personal computers allowed for the digitization of this information (via film scanners) and backups onto some form of durable media.

Nowadays, we usually skip the whole film scanning step (unless that is your thing) and go straight to storing images on our computers and backing up from there. The problem is that we are usually pretty bad about keeping up this sort of discipline unless you are

  • massively Type-A
  • or

  • were raised by Shaolin monks.

We get about as far as loading them on the computer and hope that the drive won’t fail.

I have bad news for you champ, your drive will fail. It’s not a question of “if,” but “when”. In fact, it’s easier to predict than you think. All hard drive manufacturers list the MTBF (mean time before failure) value for every drive they make. MTBF is usually expressed in hours of operation. All you have to do is open a spreadsheet and plug in the date you installed the drive, the MTBF value and the following formula:

=(install date cell) + ((mtbf cell)/24)

This will give you the estimated date the drive will fail if used constantly (24/7/365). Remember, this is an estimate! The drive could fail sooner (bad drive or clumsy owner) or later (charmed life — buy lotto tickets and share with me) but it will fail.

Hence the need for backups, which are in our best interest and thus we are loath to do it. To this end, OS makers have included utilities in the latest and greatest versions that automatically backup our data in the background while we check e-mail, surf, or play games online.

I love it when technology can be used to overcome an inherent flaw in human nature. It gives me hope that one day, we will at last find a cure for stupidity.

I, being a paranoid bloke, am not happy unless I have at least three backups of my important stuff and one of them has to be offsite.

Thus my workflow includes the use of a program like ImageIngester Pro on the Mac/PC or Downloader Pro for the PC. Both of these apps do one thing: read photos off of cards and get them into your computer in an organized fashion. They both have the capability to read image metadata and build folder structures based off of it. Most importantly, they allow to make a backup at the moment of ingestion making sure you have a copy squirreled away somewhere else.

I add to this a networked storage device, allowing me to keep that backup copy stored somewhere else. I currently use an HP MediaSmart Server ( a Windows Home Server based device – read my review here ), but a NAS or a DroboPro would work just as well. One advantage that the MSS and DroboPro share is the ability to expand storage without requiring matched drives. They also maintain duplicate copies of the data for backup purposes. The DroboPro also has the advantage of being an iSCSI target as well as being VMWare ESX certified if you really want to geek out your server systems.

This, however, is a secondary backup. My first backup is handled via Time Machine on my Mac to a separate 1.5TB drive for that express purpose. I tried to backup via TM to the HP MSS (which you can) but the MSS only sports 100MB ethernet and that was not enough to handle the massive amounts of data TM was throwing at it in my case. YMMV. You can also use ChronoSync or SuperDuper as a means to backup your data.

I also recommend a tertiary backup to an offsite location. This can be as simple as using an external drive and storing it in a safety deposit box, but I am looking to have the computer do all of the work.

To that end, I started investigating online storage solutions.

At first, I tried JungleDisk which leverages Amazon’s S3 Storage Service to back your data up. It also has a client for WHS which makes the whole process transparent. The thing to be aware of here is that this solution has two charges involved: the JungleDisk monthly rate (flat cost) and the S3 storage charges (variable, based on amount stored and transferred). I discovered that these variable charges are the deal-breaker here for me as a professional photographer. I had a big job last December and when I got the bill in January I nearly had a coronary. The price jumped up by 8x! So long, JungleDisk.

My search continued. Mozy Pro (server version) looked promising, but payment structure was oriented towards megabytes of storage, not terabytes. Scratch Mozy Pro.

In July, a friend told me about Carbonite backup. I checked them out and they looked very promising. Flat annual fee, unlimited storage, and native clients for multiple platforms. The FAQ on the site states that the initial backup could take a while but it should average about 4 GB/day upload. I had around 600GB of data to backup. At first, this promise was kept.

After a while, I started noticing that amount being transferred was dropping steadily. Being a systems programmer by day, I decided to to do a little log analysis and one spreadsheet later I had proof of what I had suspected: Carbonite throttled my backup to a tiny fraction per day. By my calculations, the initial backup that I started on July 20, 2009 would complete on April 5, 2011. I sent this information to Carbonite Customer Service and the drones there confirmed that large backups are throttled back in favor of smaller backups. Goodbye Carbonite.

I then tried BackBlaze based on the favorable reviews by Scott Bourne and Chris Pirillo. By this time my backup needs had increased to 752GB.

Price: $5/month/machine – check.
Transfer limits: unlimited – check.
Storage limits: unlimited – check.
Native clients: YES – check.

So far BackBlaze looks very promising. The client software is Snow Leopard compatible and monitors the backup giving you detailed info regarding its status. A link is provided to a web page that estimates you initial backup completion date. The throttle slider is a great feature for the bandwidth challenged. I have mine opened up all the way to get the initial backup done quickly and BB estimates that it will complete in 24-25 days. My own calculations are leaning towards 66 days, but that is a damn sight better than St. Swithin’s Day 2011. I’ll be able to refine my estimate once I have a few more days worth of data. They even offer restore data on optical or USB media which is nice in case the network is down.

So, in a nutshell:

  • Tier 1 backup – Time Machine (Windows 7 has an automated backup app too – it’s just not a slick);
  • Tier 2 backup – Network Storage (NAS (good), HP MediaSmart Server (better) or a DroboPro (best));
  • Tier 3 backup – online/offsite backup (BackBlaze current top contender)
Great! You've successfully subscribed.
Great! Next, complete checkout for full access.
Welcome back! You've successfully signed in.
Success! Your account is fully activated, you now have access to all content.