|chris30 you are mixing clustering with virtualization.|
A cluster requires a SAN since there needs to be shared storage so if one cpu goes down the other still provides the path to the data.
Virtualization is where you host other OS's/Apps on the same box.
You can combine the two but at that point you need to go to blade servers. If the company can only afford you they can't afford blades.
There are a number of standard strategies for system failover. What you choose lands somewhere between what is acceptable downtime and cost of failover.
First level is mirrored/raided drives with hot spares
Then there is a regular backup that is tested after each backup.
Next is spare system replacement hardware or a vendor maintenance contract that replaces failed hardware in 4 hours.
After these basics you get into all sorts of different scenerios. Even virtualization requires those basics to be met.
So what business is the company in what is an acceptable level of downtime?