In Virtualization, a Hypervisor is a software that servers as a mode of partitioning a physical server or host server into multiple OS based Virtual Servers or guest servers. A Hypervisor is responsible to allocate resources (CPU, RAM, HDD) of the physical server or a Dedicated Server among various Virtual Servers or Virtual Machines.

A Hypervisor provides a virtual platform to run multiple Operating System instances at once keeping each one of them unknown from one another. As said a Hypervisor partitions a physical server, here the Hypervisor have two types: 1) Native or bare metal or Hardware based, and 2) Hosted or Software based.

1.  Native or bare metal or hardware based Hypervisor: This type of Hypervisor runs directly on the hardware of the physical server and have full control on that hardware so as to provide separate OS instances to the guest servers. An example of this type of Hypervisor is VMware, XEN and Microsoft's Hyper-V. With so created Virtual Server is separated from other VMs that they are unknown with the fact that they are on the same 'parent server'. They are also called semi-dedicated servers and whatever the allocation is, in terms of hardware, its 'Dedicated' allocation.
2. Hosted or Software based Hypervisor: With this type of Hypervisor a pre-installed OS is required on the physical and so the control of the server is with the OS and not with the Hypervisor. It includes VMware workstatoin, Microsoft Virtual PC and Parallels Workstation are hosted hypervisors. 

The most used Hypervisors are either VMware or Hyper-V, though VMware is suggested and opted by many users and companies as VMware provides good features and  high level of scalability. With the new VMware Vsphere (vCloud) the cloud computing have become more common now a days as the user can scale his server in the run-time without affecting up-time.