Like the op, I have been researching the different options for myself. I looked at Ubuntu, free nas, unRaid, etc.
I had a few requirements for whatever I chose:
-Redundancy through parity preferable as duplication takes too much room
-Nice to have capability to withstand 2 drive failures would be a bonus
-Not have to spin up all the disk for every access (for energy, noise, and drive wear)
-Able to serve 3 blu-ray mkv HD streams at once (worst case scenario for me)
-Easy to setup and use
-Works with Windows since that is where most of my apps are
I didn't want a typical RAID solution that keeps ALL the drives spinning even if I write/read a small file off of it. This produces more noise, heat, and drive wear. Plus, if the raid controller goes down I can't take a drive out and insert it into another computer and read off the disk.
I can live with the speed of a single hard drive (which can max GigE anyways) since I am not a datacenter with tons of users on it. I just need casual access to files in my data pool. I like the JBOD/WHS DE/Spanning approach in that it looks like one large drive. JBOD and Spanning are too risky if a drive goes your screwed. I like WHS DE but duplication takes too much space. In the new WHS Vail they took a MAJOR plus away from the first version. In WHS you can pull a drive and read it like any other NTFS drive on another computer. The new DE eliminates this as it writes in 1gb blocks to the hard drives.
After doing some research I found 1 solution to my needs at least. I feel it ties in perfectly to a home media XBMC style approach. The one that really caught my eye
was FlexRaid.
I am not trying to pimp flexraid out but just give a heads up since it is pretty new. Flexraid works on Linux AND Windows (even WHS). You can assign whole drives, folders, or even single files to create an array. You can assign multiple parity drives all setup however you like. It uses a Raid 4 or Raid 6 setup. It is super easy to use with the web gui or cli. With FlexRaid view it presents itself as one large data pool. You can create virtual views and even virtual folders for files sorted off their metadata if you want (like a separate folder view for all your music sorted by years or genre) -soon to be implemented.
You can create/delete the array (even with existing data on the drives) and recreate it without losing your data (the array config is stored in a separate file). You can pull a drive from the array and read it from any other computer as it's a standard NTFS setup. Drives not being used are powered down (following your OS settings). You can manage the array through the web gui across the network if you wish.
It took awhile to get my head around the way the array works since I am used to a more typical raid approach. Anyways, try it out and see if you like it. If you don't, just uninstall it. Your data won't be touched. I am using it right now and I am highly impressed!
NOTES:
The software is relatively new so documentation is lacking. For info you need to go to the forums. Also, for 64 bit Windows users, there is a new build coming in a few days as soon as the developer get's his signed driver certificate in. The build is done just waiting to package for the cert. I would wait for it to try it out. The new Linux build is done.
The parity is "snapshot" based. You set a schedule via the gui on when to RSynch the parity. This makes it perfect for data that is not changed often like movies, shows, and music (media). There are verify functions as well to make sure the data is intact. The developer is making a "live" version that will not be snapshot based and also developing a "Nas" version. He is also going to implement a "self healing" feature to prevent failing hard drive data corruption (I think this is from ZFS to protect against bit rot correct?).
Sorry to all if that was a bit long.