All these servers could be replaced by one, well-specified, system requiring only one network and one keyboard, video and mouse connection.  More importantly to your bottom line, only one power supply and one hardware warranty is required.  We would always recommend that your virtual infrastructure is managed from a separate, low-spec server which can also act as a domain controller for your network. 

 

Obviously in the real world you may feel uncomfortable putting all your eggs in one basket and there are two ways to address this single point of failure:

 

Hardware resiliency - although the latest servers are well-built and largely reliable, it is a fact of life that individual components can fail.  Hardware resiliency can be improved by:

• use of redundant power supplies to minimise power loss through a power supply unit failure
• use of RAID disk technologies to use multiple, hot-swappable, disk drives
• use of multiple network ports to provide failover and bandwidth management.

VMware functionality -

VMware virtual infrastructure centre provides functionality to reduce the risk of a server failure causing the loss of productivity.  By using a Storage Area Network (SAN) or Network Attached Storage (NAS) and a second VMware server, additional functionality can used to provide High Availability solutions, DRS and the use of V-motion.