Adding redundancy to vCenter on a physical server

I implemented vCenter on a physical Windows server in order to manage a vSphere v.5 infrastructure based on 2 ESXi v.5 servers.

How can I make the vCenter redundant and fault tolerant, in order to protect it in case of failure of the Windows server?

• Can I install a second vCenter on a VM and keep it synchronized with the main vCenter?
  
• Can I add a second vCenter as a Virtual Appliance and keep it synchronized to the main vCenter?
  

As I am quite sure I am not the first asking this question, is there any document containing the best practice to make the vCenter redundant and fault tolerant?

Regards

marius