Why was Novell not included in the testing since the company has a mature Xen offering ? Did Novell decline to be included or were the authors unaware of the products in the marketplace?
|
Does Verizon's Voyager stack up to the iPhone? |
|
|
5 IT skills that won't boost your salary 
[1,407]
Women 4 times more likely than men to cough up personal info [589]
Japan's 10 funniest tech-related commercials [Videos] [407]
Throwing away a promo CD is "unauthorized distribution"? [1,265]
Adults too quick to dismiss educational video games [682]
Attack of the iPhone clones [Slideshow] [578]
10 things IT needs to know about AJAX [1,258]
This Year's 25 Geekiest 25th Anniversaries [Slideshow] [409]
|
|
Hardware architecture not included as a factor?
It seems to me that this test should also take into account performance of each hypervisor based on different server architecture, to ensure that performance degradation of each hypervisor wasn't hardware related.
Consistency
We chose to do the tests on a single mainstream server for consistency sake. Knowing that there are new hardware-enabled hypervisor routines, we chose to use a server that only complied with the vendor's 'approved' or supported hardware list but without hypervisor-enabled technologies, as these are inconsistent in terms of long-term support.
Novell SUSE Xen
We are indeed aware of SUSE Xen and have reviewed it in the context of previous tests of SLES versions 9 and 10.
At this time we see Novell's implementation of Xen to be surprisingly similar to Citrix XenServer. At the time of testing, like Citrix XenServer, SUSE Xen didn't support Windows 2008 Enterprise Server, which was a requirement for this particular test.
Self-Fulfilling Prophecy
Having a test which is artfully constructed to exclude one of the three main competitors is really no test at all.
It's OK to mention limitations of the various competitors in your testing which prevented selection of certain guest operating systems, but there are many guest operating systems which are supported by all three and which could have formed the basis of an honest test.
As for Xen and Windows 2008 Enterprise Server, might not the compatibility problems be rooted in the same issues that prevented Lotus 1-2-3 for Windows from functioning properly under Microsoft Windows? Might not a proprietary and undocumented para-virtualization interface be the issue?
VMWare is quite careful about certain guest OS vendor sensitivities in a way that Xen is not. VMWare therefore might have help from Microsoft that the Xen folks do not. A rather dramatic case-in-point concerning this behavior is VMWare's agreement with Apple not to allow virtualization of Leopard Server on anything other than Apple hardware -- and to not allow virtualization of Leopard client at all, even on Apple hardware. I'm sure that VMWare arrived at this agreement to avoid having Apple go out of its way to make sure that Leopard won't virtualize at all under VMWare's solutions. Various Apple-fanatics have torn into the VMWare code and have found special checks associated with installing a client OS which appear to be there only to enforce Apple's EULA [since disabling the checks allows successful installation and running of Leopard client; disabling other checks may allow running Leopard as a guest on non-Apple hardware, aka Psytar].
"have found special checks
"have found special checks associated with installing a client OS "
man, do you have any proof ? I think you souldn't talk about things you don't know.
Man, do I have proof!
google the following terms: vmware fusion leopard client
or, for a single search result, see http://www.macosxhints.com/article.php?story=20080805081909302
Note reference to Apple's EULA in the above search result.
Accessing various other search results, you will see that VMWARE is indeed preventing the installation of Leopard client, but that if any of several methods are used to hack the VMWARE install to make Leopard client appear to be Leopard server, the install will complete and the resulting virtual machine is viable.
If you're going to attack the verity of another poster, please do the necessary finger-work to determine out whether or not the guy you disagree with has actually stated a falsehood.
Please beware of making false accusations.
You have no idea of what you're talking about in this regard.
XenServer from Citrix has had no support (even now) for Windows 2008 Server. They claim to have it soon. We're likely to cover other Xen variants, including Virtual Iron, Novell's Xen components and others very shortly.
As far as I'm personally concerned, Apple's Leopard is a nice operating system but I care not one whit if it can't be virtualized-- it's not an enterprise server OS. The currently popular phrase 'lipstick on a pig' comes to mind as Apple has specifically tied support for Leopard to its own hardware, and nothing else. That hardware really isn't a good virtualization platform in the classic sense. So the point about virtualizing Leopard is moot, IMHO.
Your artfully constructed comment just doesn't hold water. Sorry.
Holding Water -- Gotta Go.
When someone artfully constructs a review so that one of the three main competitors can't even reach the starting line, it's not much of a review, is it? Just pointing out the obvious -- that the guys who made it past the starting line have a rather incestuous relationship relationship with Microsoft -- wait a minute! One of them IS Microsoft! Whodathunkit?
I'm not much of a Leopard fan either, but if VMWARE is cozy with Apple, one has to wonder how cozy they are with Microsoft. There may well be a bit of cooperation going on, even though the two are ostensibly competitors.
I think my comments hold far more water than your review.
Who wrote this headline?
I don't know who's responsible for the headline, but I find it disingenuous. "VMware edges out Microsoft..." makes it sound close, when the fact is that in some cases, VMware absolutely beat the pants off Hyper-V. The fact, if true, that Microsoft is only supporting two operating systems should have been commented on more. The authors mentioned it but hardly criticized Microsoft. Perhaps this is because this was strictly a performance test, but it seems as if the article was giving Microsoft more than one benefit of the doubt. It's as if Microsoft was taking an SAT exam, but asked to be excused from the math section because it forgot to study and the proctor agreed.
One other question: why didn't the authors test with Windows Server 2003?
millions of I/O per second?
Clearly, VMware ESX I/O is way better than Hyper-V. But I suppose the word "millions" was accidentaly inserted.
Post new comment