I've been researching this for my organization, so I think I can give some good insight.
One of the biggest advantages of virtualization is that the virtual machines are hardware agnostic - any hardware capable of running your chosen VM software (VMware, Hyper-V, etc.) can run the virtual machines.
Since the virtual machine is always presented with the same hardware, if your server fails, you can replace it with different hardware and not have to worry about driver issues on the virtual machine, or even having to reinstall the OS from scratch because the hardware is too different. This is especially true for any windows OS, which tends to have more problems changing hardware in my experience.
Another advantage is that recovering the system is just a matter of copying the virtual machine's files over to a new system and starting it. It's usually simpler and faster than restoring from regular backups.
In looking at a disaster recovery plan, you really need to be specific about what kind of disasters you are anticipating, and how much data loss is acceptable. For example, are you planning for:
- Your server failing?
- The datacenter failing or losing its network connectivity?
- The entire building or city being destroyed or inaccessible? (ie. natural disasters)
Virtualization on it's own doesn't solve any of these. What it does do is make solving them easier, and make recovering take less time, and give you more options as to where you can restore your systems.
Another thing that might be an advantage to you is simply the ability to run multiple operating systems on one physical server. In your case, you could seperate your SQL server from the web servers into their own OS. By doing that, if there were problems with the operating system, you'd only have to repair or restore from backups that service, not the entire system. If would also mean that if you outgrow your current server, you can expand by just moving the SQL server to another physical system without having to chance anything else.