Photo Storage and Backup Strategy (Updated)
When a single RAW image is 20MB and short HD videos are 200MB, it doesn’t take long to figure out that the internal drive of your computer is NOT the right place for your precious DSLR files.
Once they’re not on your main hard drive (which you backup, right?) – then what is your new backup strategy?
This summer, I knew I had to find a better solution. My criteria was simple:
- Must work with my existing computer and photo management software
- Must be easy to maintain in the months and years to come
- Must be fast and easy to backup
- Must be able to keep a full copy outside my home (just in case my place burns down or I get robbed)
- Must be affordable
Yes, I do my serious post-processing in Lightroom and Photoshop but that’s on my Macbook Pro. The output of that work does find itself back in the main iPhoto library (eventually). The only reason I don’t use Lightroom for everything is that it’s not very family-friendly. This dual-setup isn’t perfect, but it works well enough for now.
To solve my storage and backup needs, I considered three options…
Option 1: Fancy Multi-Drive RAID Enclosure
I considered getting a fancy drive enclosure such as a Drobo or a more generic version (I shopped OWC for those). These enclosures protect your data via RAID technology which ensures you can survive a drive failure. Consider it a real-time backup.
There were two main issues with this option:
- Fancy enclosure + drives = very expensive (Drobo is US$350 and you still have to buy 4 drives to fill it)
- RAID protection is great but a bunch of connected disks doesn’t give me one complete copy on a single disk that I can take off-site
Option 2: One USB Drive + Network Backup
My dream scenario was actually to get a single, cheap USB external hard drive and protect it with a network backup solution. I liked this idea as a network backup creates an off-site copy all in one step.
There are a variety of specialized “cloud-based” storage providers like Amazon S3 (which I use to backup this blog), but they’re not viable for 100’s of GBs of data coming from your home.
- Backup over the Internet is very slow (initial backup of 300GB would take many weeks and it would be worse with a DSL connection that is very poor upload performance)
- Amazon S3 is cheap but if you add up bandwidth and transfer charges, it costs more than option 1 after only a year or two.
Amazon S3 costs $0.140 per GB-month. Transferring data is $0.100 per GB. A competitor, Rackspace Cloud Files, is similarly priced at $0.15 GB-month (first 5GB free and free data transfer).
With about 300GB to store, you’re paying no less than US$500/year and that climbs to US$2500/year if you need 1.5TB.
There is actually a network backup solution provider that touts US$5/month for unlimited personal backups (yes, storage included). It’s called Mozy Home and it’s owned by EMC, a reputable company in the storage space.
UPDATE 1: Another online service to consider is Backblaze. it’s $5/month as well. They say that a typical home connections can backup online 2-4 GB per day.
UPDATE 2 (Feb 4, 2011): Mozy is increasing their prices to $6/50GB. Bye bye unlimited plan. Perhaps Backblaze will follow. More about this change can be read on TechCrunch.
UPDATE 3 (Apr 25, 2012): Google Drive has been announced with prices that undercut the market (effectively $0.60/GB per year). I’d trust Google with my data long term as well.
With Google Drive, you can get 100GB for $5/month, 200GB for $10/month, 400GB for $20/month, 1 TB for $50/month, and it goes up to 16 TB. This option has become much more viable for many people at these price levels, especially with lower storage needs (again, 100GB is only $60/year).
For me, I’d need the 400GB package which totals $240/year – that’s not a horrible price to pay for automated peace of mind. That said, it could take MONTHS to do that initial synchronization and that would risk an upload limit surcharge by my home Internet provider if there is one, I’ll have to check.
UPDATE 4 (Aug 21, 2012): Amazon Glacier has been announced with a price that’s 1/5th those of Google Drive and is only $0.01/GB per month (12 cents/yr per GB). This is like cold storage for data files as they’re not immediately accessible once archived this way. You effectively request an “archive” or “vault” (that’s one file or collection of files) and must wait 3-5 hours for it to be available again. Once available, it’s online for 24 hours for you to use/download. They brag that the annual durability of these files is 99.999999999%. That’s pretty good ;). They will soon launch an automated system so that files you store in Amazon S3 can be archived into Amazon Glacier automatically (e.g. based on age).
UPDATE (Feb 1, 2013): HayStack Arq (US$29) is a cloud backup utility for Mac that supports Amazon S3 and Glacier. It also has a nice iOS app called ArqView. Thank you AJ Sethi for the tip. Windows users can try CloudBerry (US$29).
UPDATE 5 (Mar 14, 2014): Google Drive just dropped their pricing in a major way. 100GB is now only $2/month, 1TB is $10/month and 10TB is $99/month! That undercuts most of the competition quite a bit. It’s still twice as expensive as Amazon Glacier at $0.02/GB per month but this is active storage you can readily access and work with. As soon as you graduate to 1 TB ($10/month), it’s on-part with Amazon Glacier!
They say most computers can upload 4-6 GB per day for that initial backup, which means 300GB would take no less than 50 days to upload before you even started the incremental backups! Still, it should work in theory. My experience with Mozy Home’s free 2GB service had been mediocre so I couldn’t imagine it working well at 300 GB and eventually 1TB+.
Option 3: Three Inexpensive USB Drives
I finally landed on the right solution for me. It was quite simple and it made use of competitively priced, stand-alone external USB hard drives.
I bought three (3) Seagate FreeAgent 1.5TB USB drives (7200 RPM) at Costco for C$90 each. It was a great price so I didn’t hesitate. I got these in August 2010 and I still don’t see 1.5TB drives going for that price today (but it’s getting close now).
Whenever I outgrow 1.5TB, I’ll probably be able to replace them with drives twice as large for the same price. Upgrading them would also free up my current drives for use elsewhere. You can’t do that if the drive enclosure and drives are separate.
There is indeed a software component to this solution and a few manual steps to follow. I couldn’t use Mac’s Time Machine for backup as that’s only for my computer’s internal drive. My photos live exclusively on these external drives.
I named the main drive “MASTER” and iPhoto only ever connects to that one drive to locate its libraries. I named the other two “MIRROR A” and “MIRROR B”. I consider those read-only copies and never make active use of them.
The software I use to copy one drive to another is called SuperDuper! It’s only C$30 (one-time purchase) and it handles both the initial and incremental backups very well. You can even setup an automated schedule (e.g. daily, weekly). I don’t use the schedule as I keep the mirror drives unplugged to maximize their longevity.
So why have the third drive? That’s the off-site copy.
I keep that third drive in my office drawer and bring it home every few weeks. I re-sync it to get any missing photos and videos (a single step with SuperDuper!) and bring it back to work the next morning.
I bought 4.5TB of capacity to get 1.5TB of usable space. That’s unfortunately the cost of a reliable backup strategy. At C$270, it’s easy enough to justify. If my house burnt down, it’s the photos of my kids growing up that I’d miss the most.
If you do the math, it works out to be $0.18/GB. If those drives last me 2 years, that’s only $0.09/GB-year. If they last longer, it only gets better.
UPDATE (Apr 25, 2012): As mentioned above, Google Drive has been announced with prices that undercut the market at only $0.60/GB per year. It might be worth considering given the automation. You still have to buy a single USB drive for your data unless your internal disk is already big enough. Factor that in and any bandwidth costs for that initial upload (if any).
UPDATE (Aug 21, 2012): If you don’t need immediate access to your photo archive, Amazon Glacier is 1/5th the cost of Google Drive at $0.12/GB per year. That’s a great and inexpensive long term solution for sure. Amazon offers tools for bulk uploading content as well. Learn more by reading the Amazon Glacier FAQ. Mac users can try HayStack Arq (US$29). Windows users can try CloudBerry (US$29).
UPDATE (Mar 14, 2014): Google Drive just joined Amazon Glacier with some amazing pricing. At $2 a month, it’s easy to start and at $10/month for $0.01GB, it’s the best price in town. Moreover, it’s readily accessible storage whereas Amazon is more archival in nature. At these prices I may have to backup purely online… I’ll have to wait until LTE upload speeds are common for residential internet connections.
An entry level RAID enclosure + drives would have cost more and and still wouldn’t have provided me the off-site copy. The network backup solution would have been slow and cost no less than $55/year (not including potential charges by my ISP for massive uploading).
Like I mentioned earlier, this solution provides a nice upgrade path. Once I outgrow 1.5TB, I’ll probably get three more massive drives for the same price. The next drives will likely be USB 3.0 compatible which will be much faster too.
I’m curious to know what other people are doing to protect their valuable photos and videos. Please share your thoughts in the comments of this post!
If you’re buying USB drives now, checkout B&H.
Also, be sure to read my (newer) related post: Ensuring Your Digital Photo Archive Survives for Generations (i.e. to DNG or not DNG)