 |
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: Performance comparions between VI and ESX? Topic Summary: Created On: 01/09/2007 03:14 AM |
Linear : Threading : Single : Branch |
 Topic Tools
|
01/09/2007 03:14 AM
|
|
|
Guys wondering what performance comparisions have been done between VI and VMWare ESX? I'm assuming that VI is more a comparision at the high end rather than against the VMware server?
I'd be interested in some thoughts on this - I saw in one of the VMware blogs some mention of a new vendor neutral benchmark under development I suppose just for this purpose. Is VI's approach faster than VMwares? In my experience customers are interested both in cost of ownership and also in performance. If they virtualise, they're concerned that their users shouldn't necessarily be able to tell. With VI's approach am I right in understanding there is no need to run a "vmware tools" type software model inside the client Virtual Machine? Also what mechanisms are there to migrate customers who may already been running VMware or MS virtual s/w? Cheers BJ |
|
|
|
|
|
01/09/2007 08:53 AM
|
|
|
Virtual Iron is comparable to VMware ESX -- it is "bare metal" hypervisor virtualization. Virtualization overhead depends on the application characteristics, such as CPU, disk, and ethernet IO. There are a two benchmarks being developed by the industry, one by VMware and one by Intel. Both are being submitted to Spec.org and will be in committee for some time.
It's pretty straightforward to migrate from VMware or Microsoft Virtual Server to Virtual Iron. You can use a P2V tool like PlateSpin PowerConvert, or you can migrate the virtual hard disks. We wrote a blog on how to do this. Edited: 01/09/2007 at 08:55 AM by cbarclay |
|
|
|
|
|
01/09/2007 08:25 PM
|
|
|
So what about the requirement for a sw tools layer inside the vm iself like vmware tools - does VI need/require this?
Would you say that VI's approach would be then pretty much of the same order of magnitude performance wise as ESX? Shouldn't necessarily be any faster or slower? What is your take on the underlying file system and how important a factor it is (ie: the underlying filesystem that will hold the VHD files in VI's case). Vmware obv tout their VMFS as being optimised for this purpose - what are your thoughts on this? One other query I had related to your ability to restart a workload on a differerent VI server if one of the VM workload should become unavailable due to say underlying hardware failure on the primary VI host. Doe this require some special snapshot type feature at the SAN level to faciliate this? Thanks for the fast response. |
|
|
|
|
|
01/09/2007 09:36 PM
|
|
|
Virtual Iron provides VSTools, similar to VMware's tools, to provide high performance IO in virtual servers. I would say that both approaches are similar, with the main difference that Virtual Iron is the first to take complete advantage of hardware assisted virtualization which promises to dramatically outpace software assisted virtualization.
The file system for virtual hard disks does not appear to be a significant factor in overall virtual server performance. Virtual Iron does not use a clustered file system (VMFS is), so performance penalties are minimal. Virtual Iron provides a capability called LiveRecovery that provides capabilities such as you request. Edited: 01/09/2007 at 09:39 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