With the explosion of Digital cameras, digital video recorders, iTunes, MP3 players, Youtube and more, storage of digital media is going to be very important, if not downright critical. How many people do you know who really backup what digital media files they have? It doesn’t need to be photos or videos. I know this entry is filed under Photography but digital media can mean documents and spreadsheets as much as photos and video. It can also be your digital audio collection.
Primarily, the backup we’re talking about is just a second copy of your information somewhere else. Where? Well, anywhere other than your PC is good. I don’t agree much with same PC backup (and generally, many users can’t tell the difference between physical drives and partitions, etc). So the question is where and how to store the backup.
WHERE?
The most simple location is probably very ubiquitous, the USB drive (aka thumb drive, usb stick, etc, etc). With drive capacities at 32Gb and beyond, and at affordable prices, it makes is ideal for users with limited amount of information. My parents for one can do with a 4gb USB drive. If you have more stuff to backup, get a portable or external hard drive. Here are some that I know of that users have had good experiences with: Western Digital MyBook Essentials or Seagate FreeAgent Go if you want something portable. For larger disk capacity, Western Digital Essential, Home or Studio or Seagate’s FreeAgent Xtreme make good buys. For heavier disk storage and on a networked environment (like me), you can go with a Network Storage Appliance (NAS). I recommend QNAP’s TS-209 or Synology’s DS-207. Those on a tighter budget can opt for DLink’s DS-323 or single drive versions of the QNAP and Synology models I mention previously. However, those require a bit more know-how to setup and manage. Me, I run a custom built NAS which is actually a dual-core box (speed doesn’t matter) with 8-SATA disk ports (these matter, RAID is not required) and 4GB RAM (this also matters). Read the full entry to see how it is setup.
HOW
This is as important as where because if you don’t store your backup properly, you’d find yourself in trouble when you need the backup. Generally, most people can get by with weekly or even monthly backups. We aren’t large corporations needing daily or hourly backups so a week or so is good for 99% of home users. Some probably can get by with once a month. If using a portable hard drive, make sure you keep that in a safe place. I’m not going into backup software but because I don’t use them myself — primarily, my goal is to mirror (copy all) the data across to another disk. I use Acronis TruImage to backup the system drive but applications can be re-installed, data can’t so my focus here is on data. Firstly, ensure your data is stored logically and in an organized drive. This makes backing up easy. My favorite tool is robocopy. It works very reliably, and the ability to monitor folders for changes is invaluable. I use robocopy to copy from my network PCs directly to the backup server regularly. If you want a prettier utility, there’s Allways Sync.
My Backup
I use an old 1.8Ghz dual-core machine with 4GB RAM as my Network Attached Storage. While the NAS products from Synology and QNAP are good, pricewise, a small custom built PC will outperform and be more expandable than one of these systems while still costing less (4 or more disk models only). Of course you have to factor in the price of the Operating system (You can use Linux) and power consumption. Anyways, that’s a topic by itself. Basically, I’ve 7 hard drives on this server. One drive holds the operating system. I use Windows XP because sometimes I need to offload tasks to it such as BitTorrent (QNAP and Synology NASes can do this too), Photoshop batch jobs (mostly resizing but occasionally a large RAW conversion) and running my Nikon LS-40 slide scanner (the Nikon software likes to hog the CPU when the scan is underway for some unknown reason). The other disks are purchased and installed in pairs. I’ve a 500GB pair, 750GB pair and a 1TB Pair.
The disks are in pairs because I use robocopy to watch for 500 changes or 45 minutes (whichever comes first) to mirror them. That way, if one disk should die suddenly (which happened before) my data is still safe. I’ve dabbled with consumer RAID available on some systems and find them very unreliable. If you want RAID, I suggest going with Areca which some people I know have had up to 4TB (12 disks, RAID6) running reliably. It’s not cheap and disks are cheap enough nowadays that a robocopy mirror will provide all you need (we’re talking about 2 or 3 users max. RAID may offer better performance if you have larger number of concurrent users). So my 6 disks end up providing a total of 2.2TB’s of space. RAID is nice because it shows you a single drive. Yes, but with XP onwards (Linux is already there), you do have NTFS junctions which are symbolic links. They allow you to create reparse points to physical disks. Basically I can create a folder called 2007 which can point to drive 2. Of course all this means a bit more management and careful setup but it’s a lot cheaper that a good RAID setup plus easier to recover should any disk decide to buy the farm.
For example, a photos folder will have various individual symbolic links for the individual year’s worth of photos (organized subsequently by day and shoot) but I do not have to to where the photos are physically. \\storage\share\photos\2007 can point to the storage server’s 500Gb disk or 1TB disk but it doesn’t matter. Of course, in my case, it does sequentially with newer folders typically being created on newer disks. With a 20-month refresh (new pair of disks every 20-months) you probably have almost no problems with disk failure plus you get to buy larger disks at economically priced points. I aim to keep 6 as the maximum number of disks (power, heat, etc) so the next disk acquisition will be 1.5TB disks and they’ll replace the 500GB disks. Access speed is roughly 25-35Mb/sec over a gigabit network (tested with Canon CR2 raw files averaging 12Mb). Nothing fantastic and some tuning should improve the results but since concurrent access is not critical (I’ve only 3 users max), I can live with it. Faster that shuffling DVDs and still be able to backup everything on our network PCs.