HP DL360 Server virtualization technologies for x86-based HP BladeSystem and H - Page 4

than having the hypervisor layer trap and translate all privileged instructions. As a result of

Get HP DL360 - ProLiant - G3 PDF manuals and user guides
UPC - 613326948835
View all HP DL360 manuals
Add to My Manuals
Save this manual to your list of manuals

Page 4 highlights

Figure 2. In virtual machine technology, the software abstraction layer (hypervisor) abstracts the physical-layer resources. In x86 environments, there are several variations2 within software-layer abstraction of the server hardware, including these general categories: CPU binary translation. The hypervisor traps selected CPU instructions, such as CPU-privileged instructions3 and performs necessary translations to make the guest OS think it has complete control over the server hardware. Hosted OS, application-layer abstraction. The software-layer abstraction resides as an application on top of a host OS. Most use binary translation and do not require a modified guest OS. This type of machine abstraction is typically used in smaller departmental environments rather than large data centers. Hardware-assisted virtualization (full virtualization). New processor hardware, such as AMD V or Intel VT x processor virtualization technologies that extend the x86 instruction set, assists the hypervisor. These new instructions move some supervisory control for the workload container into the CPU hardware layer. Paravirtualization. The guest OS is modified to include paravirtualized I/O drivers. The modified guest OS makes calls directly to the virtualized I/O services and other privileged operations, rather than having the hypervisor layer trap and translate all privileged instructions. As a result of the guest OS modification, the guest OS is now capable of being aware of its virtualized host (hypervisor). Hosted OS, kernel-layer abstraction. This method differs from the others because the abstraction technology is built directly into the OS kernel rather than having a separate hypervisor layer. As a result of the tight integration of the machine abstraction into the OS kernel and the assistance of AMD-V and Intel VT-x technologies, any host OS applications appear to run as peers to the virtual machine guest OS. Appendices A and B provide additional descriptions of these techniques. Table 1 lists some of the most common vendor implementations and the primary hypervisor techniques used. Vendor 2 Terminology varies when referring to software-based hypervisor technology. The definitions in this document may differ from other documents. 3 Appendix A contains additional information about processor virtualization and privileged instruction calls. 4

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
4
Figure 2.
In virtual machine technology, the software abstraction layer (hypervisor) abstracts the physical-layer
resources.
In x86 environments, there are several variations
2
within software-layer abstraction of the server
hardware, including these general categories:
CPU binary translation. The hypervisor traps selected CPU instructions, such as CPU-privileged
instructions
3
and performs necessary translations to make the guest OS think it has complete control
over the server hardware.
Hosted OS, application-layer abstraction. The software-layer abstraction resides as an application
on top of a host OS. Most use binary translation and do not require a modified guest OS. This type
of machine abstraction is typically used in smaller departmental environments rather than large data
centers.
Hardware-assisted virtualization (full virtualization). New processor hardware, such as AMD V or
Intel VT x processor virtualization technologies that extend the x86 instruction set, assists the
hypervisor. These new instructions move some supervisory control for the workload container into
the CPU hardware layer.
Paravirtualization. The guest OS is modified to include paravirtualized I/O drivers. The modified
guest OS makes calls directly to the virtualized I/O services and other privileged operations, rather
than having the hypervisor layer trap and translate all privileged instructions. As a result of the
guest OS modification, the guest OS is now capable of being aware of its virtualized host
(hypervisor).
Hosted OS, kernel-layer abstraction. This method differs from the others because the abstraction
technology is built directly into the OS kernel rather than having a separate hypervisor layer. As a
result of the tight integration of the machine abstraction into the OS kernel and the assistance of
AMD-V and Intel VT-x technologies, any host OS applications appear to run as peers to the virtual
machine guest OS.
Appendices A and B provide additional descriptions of these techniques. Table 1 lists some of the
most common vendor implementations and the primary hypervisor techniques used. Vendor
2
Terminology varies when referring to software-based hypervisor technology. The definitions in this document may differ from
other documents.
3
Appendix A contains additional information about processor virtualization and privileged instruction calls.