Why Windows Home Server Does Not Use RAID

Qblogger over at the Windows Home Server Team Blog recently wrote an article discussing the use of, or non-use of, RAID technology in WHS.  Can you use RAID in WHS?  Sure you can.  Is it more secure, or faster?  Perhaps.  But do you need to?  No, according to Microsoft.  From the article, here is why:

Windows Home Server storage system design requirements

• Must be extremely simple to use. Must not add any new concepts or terminology average consumers would not understand. Simple operations should be simple and there should not be any complex operations.
• Must be infinitely & transparently extendible. Users should be able to just plug in more hard drives and the amount of storage available should just grow accordingly. There should be no arbitrary limits to the kinds of hard drives used. Users should be able to plug in any number of drives.  Different brands, sizes, and technologies should be able to be mixed without the user having to worry about details.
• All storage must be accessible using a single namespace. In other words, no drive letters.  Drive letters are a 1970′s anachronism and must be squashed out of existence!
• The storage namespace must be prescriptive. In other words, our research told us that consumers want guidance on where to store stuff. Our storage system needs to be able to tell users where photos go. Where music goes. Etc…
• Must be redundant & reliable. There are two components in every modern computer that are guaranteed to fail: fans and hard drives. Because they have moving parts,  Windows Home Server must be resilient to the failure of one or more hard drives.
• Must be compatible. Compatible with existing software, devices, disk drives, etc…
• Must have great performance.
• Must be secure.
• Must enable future innovation. Both the amount of storage consumers are using, and capacity/$ are growing at Moore’s Law like rates (while nothing else really is). This creates a discontinuity in the industry and an opportunity for innovation. The storage system must operate at a higher level of abstraction to enable rich software innovation (file level vs. block level).

There have been many discussions in the forums regarding the use of RAID for a WHS computer.  The one concern that I seen over and over again in these posts is backup for the WHS.  Before I can discuss backup for the backup server, we must discuss what the purpose of a WHS is.  When WHS was first released, Microsoft advertised it as a backup solution for client computers.  If a computer hard drive dies, one simply pops in a new drive, loads the restore disk and restores that computer to the last saved backup.  WHS also created a a folder called “shares”.  In this folder, the WHS user could copy files to this folder, which all machines in the WHS network could access.  I can’t say definitively, but I do not believe Microsoft expected the end user to simply copy files over and delete the originals.  I do not believe that WHS’s original purpose was to act as a file server, without the benefit of that data being backed up elsewhere.  Anyone who “moved” data in this manner was asking for trouble.  And quite often they got it.

Microsoft’s original answer to this was the “folder duplication” option.  This option would create 2 copies of any shared folder that had this option enabled.  If a hard drive failed, other than the system drive, simply remove it and replace.  Although I have never seen any discussion regarding how this is done, WHS somehow recreates this duplication information on the new drive.  This is Microsoft’s alternate redundancy solution to a RAID 1 array for the shared folder data.  If the WHS user is cognizant enough to turn it on.

The “real” backup solution for the “shares” folder was introduced in Power Pack 1 which allows the user to create a data backup drive.  From the PP1 release documentation:

You can can now back up and restore your home server shared folders to an external hard drive.

When you add a hard drive to your home server, you are given an option to use the hard drive to back up your home server shared folders. When you select this option, your home server appears in the Computers & Backup tab. You can then use Backup Now to back up your shared folders to the hard drive.

You can view and restore your home server shared folder backups on the Computers & Backup tab.

What this allows the WHS user to do is create a shared folder backup and take it offsite.  The ultimate and only backup strategy in the case of a total disaster, such as a home being destroyed in a fire.

Many users were upset with the release of PP1 in that there was no option to backup the client backups or the System drive.  Along came the WHS BDBB add-in that allowed the WHS user to backup the client backup database.  As for the System drive, the only way to backup this drive up, to my knowledge, is with an imaging software such as Acronis True Image or Norton Ghost.

So, to recap the above, these are the backup/redundancy strategies that are available to a WHS user:

• Folder duplication (redundancy).  Does the same thing as a RAID 1 array, and takes up the same amount of space as a RAID 1 array.  The disadvantage is that this is only for folders beneath the “shares” folder that are turned on the create this duplication.  Neither solution, however, is a true backup.
• WHS’s external backup of the “shares” folder.  The only true backup solution for data on a WHS machine.  This would be available to a JBOD or RAID solution, as this is copied to a drive external to the data array.
• Client backup database.  The ability to create this backup is through the use of an add-in solution called WHS BDBB.  This is a true backup.
• System Drive backup.  The only option I am aware of to backup the System drive is through a 3rd party program such as Acronis True Image or Norton Ghost.  There are many potential disadvantages here.  How well do they work with a WHS OS?  How well do they restore the System drive in the event of a failure?  How easy is it to reintegrate the client backups and shares folder?  I do not know the answers to these questions.

All of this seems like a rather piece-meal solution.  Much of the above discussion leads one to suggest that a RAID array would be a better solution.  Perhaps for total system redundancy, but only for redundancy.

If we look once again at the three software sections of a WHS machine, what is the importance that can attached to having a backup to each section?  To me:

• “Shares” folder backup.  This is a no brainer.  Backup of this section of a WHS machine is of the highest importance.  If a user were to “move” over an entire family photo collection, and then lose it, that is a catastrophe.  Which is why WHS allows a duplication option for these folders.
• Client backups.  By definition, this is a backup.  The data contained here is an image of the client machine.  You can restore a complete machine or portions of it.  Do you need a 3rd copy?  Perhaps, perhaps not.  Losing a WHS machine and a client machine at the same time would be an unusual occurrence.  Not unheard of, but unusual.  At least one has the option to do so with WHS BDBB.
• System backup.  I do believe that this has a high priority.  That is what system restores are for.

In the above discussion, I have tossed around 2 key terms.  Redundancy and Backup.  Whether it is Microsoft’s folder duplication option, or a RAID solution, neither is a backup solution.  Both are redundancy solutions to protect the user against the loss of a single hard drive.  Lose the system disk in a Microsoft solution, and you may end up losing some of that data.  Lose a RAID array, you’ve lost everything.  The only true backup solution is to physically create a copy of your “shares” folder data on a separate drive that can then be removed from the machine, which just so happens to be the option that Microsoft created in PP1.  If needed, you can also create a copy of the client backups.

Just remember, redundancy is not a backup.

So, in the end, the major advantage of a RAID solution over the Microsoft JBOD duplication solution is that is it provides redundancy for the entire OS and the data on that computer.  The major advantage of Microsoft’s JBOD duplication option is the ease of adding additional space to the data drive.  Both have redundancy of a partial (native WHS) or complete system (RAID).  Just do not expect either to be a backup solution.