Wednesday, January 19, 2005

Backing up your data

I've been researching, and thinking about how on earth to backup all of my data in a cost effective way. The problem started when I got a powerbook and a digital video camera about 2 years ago. Couple that with iMovie and suddenly I was generating MASSIVE amounts of data. I ripped 5 years worth of analog video tapes onto my LaCie firewire external drives (200GB =~ 186GB) when I ran out of space. Worse yet, I ran out of space and didn't have any of it backed up. So I spent HOURS ripping the data and I could lose it with the death of a disk.

I considered buying additional firewire drives, but that's pretty expensive and you start to run out of power outlets pretty quickly. Then I considered DVD backup, but they only hold 4-8 GB and I have about 400GB so you're looking at multiple days of burn time to get one copy (at least with 1x burner). I also considered tape devices, but the low-end ones are still pretty expensive and only hold 40GB of data. If you consider higher capacity autoloading drives you're looking at $3-4k minimum.

What I wanted was huge amount of storage (~1 TB) that was redundant and online. So I setout to understand RAID. The previous article does an excellent job explaining RAID-5 and is exactly what I wanted "redundant, online, data that scales to massive sizes". I started looking for software solutions for my mac (thinking I could just use my current drives plus a couple more) but there aren't any. SoftRAID and Apple have solutions for RAID-0 and RAID-1, but not RAID-5 (though SoftRAID people say they're looking at RAID-5). I relented and went with a hardware solution, which are expensive but you can use commodity disks with them.

I bought the Arena NAS4 from Mac Ally and 5 200GB Maxtor SATA drives. The beauty is that I now will have 800GB of redundant, network accessible (10BaseT, 100BaseT, or Gigabit) storage. It wasn't "cheap" (~$2000) but doing software based RAID-1 with 8 of these is roughtly the same cost and you don't get network accessibility (for free at least), it isn't as fast (no striping and software RAID is slower), and you've got to have 8 power outlets. I considered sticking with LaCie's RAID solution, but it's not available until "Spring 2005" and it didn't have the NAS feature which was an added bonus.

A few final thoughts. Remember, it's not at all clear in any of the product literature that I've seen, but the self-contained hardware RAID solutions almost never include disks so you need to factor in that cost when considering the solution. Also, yes I know that I could get commodity hardware, a RAID card, and run Linux. However, doing that, I'd have to trade the nice compactness of my new solution and the hot-swapability. I also doubt that it'd be all that much cheaper when you're talking about hosting 4+ drives. Finally, FYI... I found great prices on

[Update] Michael asks "Why not Mirra?" My answer...

  1. I didn't know about it ;)

  2. PC only?!?

  3. Maximum data size is 250GB; I wanted at least 500GB.

  4. It's not redundant so even if you only wanted 250GB of space you still have to have 250GB elsewhere

  5. Poor product specs. Can you tell if it has Gigabit ethernet ports?

  6. It seems like one of their key strengths is ubiquitous network access; toss the thing on the net and you can get at your data (and keep it synced) from wherever you want to access your data)

Michael also mentioned that he's heard of cases where RAID controllers went bonkers and corrupted data... Hmmm... Not sure what to do about that one.

No comments:

Post a Comment