 |
Community |
 |
 |
 |
|
Hide | |
Announcements
How-tos
Documentation
General Q&A
Best Practices
General Virtual...
Virtualization ...
Suggestion Box
Virtual Applian...
Virtual Applian...
How-tos
Documentation
General Q&A
Best Practices
General Virtual...
Virtualization ...
Suggestion Box
Virtual Applian...
Virtual Applian...
|
Topic Title: why does native virtualization outperform full virtualization? Topic Summary: Created On: 02/05/2007 12:48 PM |
Linear : Threading : Single : Branch |
 |
- Liang | - 02/05/2007 12:48 PM |
|
- cbarclay | - 02/05/2007 01:28 PM |
|
- Liang | - 02/05/2007 06:33 PM |
|
- cbarclay | - 02/06/2007 02:21 PM |
|
- Liang | - 02/09/2007 12:50 PM |
|
- cbarclay | - 02/09/2007 12:56 PM |
|
- Liang | - 02/09/2007 01:26 PM |
|
- cbarclay | - 02/09/2007 01:31 PM |
 Topic Tools
|
02/05/2007 12:48 PM
|
|
|
Based on the definition of native virtualization, I would appreciate if Virtual Iron people could post the reason why native virtualization can outperform the full virtualization (e.g. by adding IOAT?).
|
|
|
|
|
|
02/05/2007 01:28 PM
|
|
|
We recently had a blog entry where we discussed performance. Performance results depend on a number of factors. Hypervisors that run directly on hardware without a host operating system generally scale better and have lower overhead than alternative approaches.
|
|
|
|
|
|
02/05/2007 06:33 PM
|
|
|
But you still did not explain why native full virtualization mode is faster than full virtualization mode? Especially with the new vt-d specification, guest domains could access HW directly. So please give me some details about the advantages of native full virtualization mode? I have read all the available VI documents, I can not find any of them.
|
|
|
|
|
|
02/06/2007 02:21 PM
|
|
|
VT-d is an interesting technology that may be applicable to some usage scenarios. The downside of VT-d is that it does not abstract virtual servers from hardware, and thus can make it difficult to perform LiveMigrations and virtual server configuration. It is definitely a technology that we are closely monitoring and intend to incorporate into the product.
Native virtualization and full virtualization are two approaches to virtualization. Each has its pluses and minuses and must be evaluated relative to the task at hand. We expect that hardware advances will continue to follow Moore's Law, putting further distance between software-only approaches. |
|
|
|
|
|
02/09/2007 12:50 PM
|
|
|
You can do full virtualization with either SW (VMWare) or HW approach (Xen). So what is the most advantages of native virtualization vs. HW full virtualization? Or are they just the same?
|
|
|
|
|
|
02/09/2007 12:56 PM
|
|
|
Native virtualization requires hardware assisted virtualization. Hardware virtualization cannot be used alone -- it lacks memory and IO virtualization. Native virtualization is hardware virtualization plus memory and IO virtualization.
|
|
|
|
|
|
02/09/2007 01:26 PM
|
|
|
My question is what the difference between the native virtualization (the term invented by VI) and the hardware full virtualization used by Xen (which also includes memory and I/O virtualization) is.
|
|
|
|
|
|
02/09/2007 01:31 PM
|
|
|
Virtual Iron's native virtualization includes a number of benefits not available in Xen:
Edited: 02/09/2007 at 01:32 PM by cbarclay |
|
|
|
|
FuseTalk Standard Edition - © 1999-2007 FuseTalk Inc. All rights reserved.
Copyright © 2003-2007 Virtual Iron Software, Inc. | Privacy Statement | Terms of Use | Site Map