So, which of the hypervisors is fastest, in terms of raw performance? That's what we set out to determine in the latest issue
of the magazine.
We could have looked at many different hypervisors, but decided to limit our testing to three -- ESX from VMware, XenServer from Citrix, and Microsoft's Hyper-V. One of the reasons was space -- in a print magazine, you simply don't have the real estate to make a story as long as you want.
The other reason was that most customers trying these things out will most likely be looking at one of these three. ESX is obvious -- it's numero uno when most admins start kicking hypervisor tires. Microsoft has put a lot of its marketing power behind Hyper-V, and it's quickly gaining a lot of recognition in the industry. XenServer, now that it's being given away with a lot of free enterprise-class features, is sure to go on most admins' short list.
The results, from writer/online columnist Rick Vanover, were startling, to say the least. The Porsche of hypervisors? XenServer. Raise your hand if you saw that coming. It outperformed Hyper-V and ESX in most categories. The pokiest? ESX. Again, not at all what I expected. In fact, even in the few tests ESX came out on top, it barely edged out the competition. Microsoft did well across the board, and is definitely a fine product.
One note about these results: First, as some of you know, VMware is very careful (some might say paranoid) about allowing benchmarking and performance results to be publicly released. Rick Vanover and I discussed our methodology with VMware, and the company agreed that our testing was fair. Note that we made no changes -- at all -- to our results, or the reporting of those results, due to our conversation with VMware. It's an important point, so you know that what we're reporting hasn't been altered in any way by any vendor.
So, what does this test tell you? One thing, and one thing only: how these specific hypervisors perform under these specific test conditions. That's it. Don't extrapolate that to mean we're saying XenServer is the "best" hypervisor, or ESX is the "worst"; that would be a mistake. As Rick points out, we didn't judge critical factors like management software, availability of third-party applications and so on. Finding the right hypervisor is a crucial decision for your shop, and a multitude of factors go into that decision. Speed is just one aspect; how important that is will depend on your virtualization goals.
What this does tell you is that from a pure performance standpoint, XenServer gets you there from here the fastest.
You'll notice that we also did a separate test with ESX, using memory overcommit to determine the effect on performance. This was for informational purposes only. Memory overcommit, available only with ESX, is quite important in many virtualization implementations, so we felt it would be valuable to show you the performance hit on your datacenter while it's in use.
Keep in mind that Rick, the author, is a confirmed VMware man, and uses ESX extensively on his production network. That makes it harder for those who want to scream "anti-VMware bias!" at us. As always, we call 'em as we see 'em here at Virtualization Review.
Do these results give you pause? Do they make you want to try XenServer or Hyper-V, when you wouldn't have previously? E-mail me or respond in the comments section below.
Posted by Keith Ward on 03/09/2009 at 12:48 PM10 comments
Brian Madden has an interesting post
on what he sees as the real value of client hypervisors. It's a good read.
His basic theory is that while running multiple OSes on a desktop or laptop will be nice, the most important functionality will be the ability to run your Windows OS on multiple devices, including your smartphone or cellphone. He's right, I think.
If that's the case, it will be interesting to see how Microsoft develops Windows. Since I believe Windows 7 will be a huge success, and continue to secure Redmond's dominance in client OSes, it will be incumbent on the company to develop smartphone/cellphone versions of Windows 7, unless it's going to cede that space completely to Windows Mobile (and Mobile 6.5 looks like a ginormous step in the right direction) and keep Windows 7 for large form-factor devices.
On the other hand, users might prefer the Windows desktop OS paradigm on their handheld devices -- having used Mobile 5 and 6 extensively, I can say I hated them with a passion, but many users may opt for the devil they know instead. Although Windows Mobile 6.5 and onward is optimized for the smaller screens, the desire for a familiar UI may trump those benefits.
Since I don't cover all things Microsoft the way I used to, I may be totally off-base here. If anyone can shed light on Microsoft's development plans for Mobile vs. Windows 7, I'd appreciate hearing from you and getting your take.
Posted by Keith Ward on 03/09/2009 at 12:48 PM6 comments
The Feb./March issue of our magazine is out now, and on the Website, so I wanted to give everyone an overview.
First, I've made a change to our news/reviews section of the magazine, which is called Upfront. I've moved much more into product reviews, since virtualization still has that new car smell, and many (or most) of you haven't used these product before. It's not like trying out another IDS system for Windows, after all.
We have three of them this month: Reflex security, VMware's ThinApp application virtualization offering, and Microsoft's System Center Virtual Machine Manager 2008. And while you're at it, let me know what you think of this new focus on products vs. news coverage and analysis. My rationale is that the nature of print magazine doesn't lend itself to that kind of coverage anymore, since it's all Web based.
There are a couple of features, written by yours truly. One deals with virtualization on the chip: I take a look at what Intel and AMD are doing to further the use of virtualization technologies by building more and more virtual functionality on the processor, and what we can expect to see coming in the future.
The other feature is a case study of a virtual desktop infrastructure implementation at the University of Maryland. The admin there needed a better way than sneakernet, and found it with Pano Logic zero-client devices. It was an eye-opening story for me to do, and showed the potential of VDI, and why I can see this technology making serious inroads in the near future.
Our sensational columnists weigh in with their latest as well. Chris Wolf, your Virtual Advisor, counts the many ways that virtualization complicates troubleshooting, and gives some suggestions for vendors on making it easier.
Virtual Architects Danielle Ruest and Nelson Ruest take a different look at Live Migration. Is it always necessary? They say that even Microsoft says it's not necessary in every case.
Finally, my Take 5 column is one I really enjoyed writing. It's about surviving in these worst of times. What should you and your company be doing?
As always, let me know what you think of the issue. Which stories did you like, which ones did you think were a waste of space? What would you like to see in upcoming issues? Tell me all...
Posted by Keith Ward on 03/06/2009 at 12:48 PM1 comments
So, Gartner says
that virtualization sales will continue to increase.
Um, this is news how?
"Virtualization helps organizations to cut costs, better utilize assets, and reduce implementation and management time and complexity, all of which are crucial in this economic environment," says a Gartner research director. Wow. What bold, fresh analysis. I'm glad Gartner's around to enlighten us.
Here's more stunning insight: Gartner expects server virtualization to continue to be the point of entry, and hosted desktop (VDI) to make bigger inroads. Man, knock me over with a feather.
Does Gartner really have nothing better to do than release "duh" papers? Well, at least some folks at Gartner are doing interesting research.
Posted by Keith Ward on 03/05/2009 at 12:48 PM5 comments
There's so much happening in virtualization these days that it's easy for stuff to slip through the cracks. New hypervisors, for instance, are still coming out. But, to my shock, one long-promised hypervisor is still waiting for its grand entrance: Sun xVM Server
I didn't even realize this until yesterday, when I was doing research for an unrelated story. I guess I simply assumed that such a fundamental part -- the fundamental part -- of any vendor's virtualization strategy would be one of the first products released. But my assumption was wrong.
Sun has xVM Ops Center 2.0 out, and they're at VDI 3.0, having fully integrated VirtualBox. But no base hypervisor. Am I the only one who thinks that's weird?
I know xVM is on the drawing board, and should be out soon; the problem is that Sun has been promising that for a long time now. When does soon become "it's here!"? Check out Sun's "xVM Central" blog, for instance: lots of news about Ops Center and VDI -- but nothing but the sound of grass growing when it comes to xVM.
It's hard to imagine Sun has given up on xVM, and decided that there are enough hypervisor choices out there that one more would simply be white noise. If that's the case, they should tell us. If that's not the case, they should be giving some kind of real update on xVM availability. Maybe the recession has forced development priorities elsewhere.
Whatever the reason, every day that Sun drags its feet on releasing xVM is another day that it falls further and further behind aggressive virtualization companies like VMware, Microsoft, Citrix and, now, Red Hat. The pity here is that I've always liked Sun's virtualization approach and vision in this space; like IBM and only a few other companies, it can offer end-to-end, hardware-to software virtualization. I guess we'll know soon enough how serious Sun is about slugging it out in this virtual ring.
Posted by Keith Ward on 03/03/2009 at 12:48 PM4 comments
If you follow virtualization at all, you probably know who Chris Wolf is. Chris is a virtualization analyst for Burton Group, and for my money, the most knowledgeable analyst in the industry. That's why I wanted him as a columnist for Virtualization Review
magazine, and was delighted that we were able to snag him.
Earlier this week at VMworld Europe in Cannes, France, Chris gave a fascinating session on hypervisors. The basic idea was to compare hypervisor readiness for enterprise computing. A hypervisor had to have certain capabilities to be considered worthy in areas like high availability, networking, management and storage. He compared four hypervisors: VMware's ESX, Microsoft's Hyper-V, Citrix' XenServer 5, and Virtual Iron's Xen-based hypervisor.
Only one hypervisor met all of the criteria in the "Required" category of functionality: ESX. All the others came up short in one or more areas. Hyper-V, for example, was lacking in some high-availability areas like the ability to prioritize VM restarts in the event of an outage -- in other words, if the server goes down, which VMs get restarted first -- and a fault-tolerant management server. Chris did, however, point out some areas for improvement, like dynamic I/O buffering and virtual hard disk compatibility.
Some of XenServers shortcomings included the lack of 802.1Q VLAN trunking, and directory services integration for better security.
Virtual Iron's was the lowest-rated hypervisor of the four, lacking an enterprise support policy among other faults; it's worth noting, however, that Virtual Iron, by and large, prefers to stick to the small and medium-sized business shops, which have fewer needs.
There's a lot more to Chris' findings than that; he and several Burton Group colleagues spent three months doing comparison tests and going over every inch of these hypervisors; their results are quite thorough, objective and eye-opening. It's worth checking out as part of your virtualization decision-making process.
I will note, also, that we add some more perspective to these findings with our separate coverage of the raw performance of three of the four hypervisors Burton Group looked at (minus Virtual Iron), and come up with some pretty interesting numbers of our own.
Posted by Keith Ward on 02/27/2009 at 12:48 PM7 comments
CANNES, FRANCE -- Oh, how I'd love to be writing that dateline for real! No, I'm still here in the States, in my humble hovel in northern Maryland, rather than gazing out at the Riviera. But I can dream, can't I?
Oh well, on to business. Our fabulous "Everyday Virtualization" blogger, Rick Vanover, has given a nice overview of announcements at VMworld Europe, and I wanted to follow up with some basic analysis of my own. (BTW, if you're not a regular reader of Rick's blog and articles here, you'd be wise to add it to your bookmarks. It has a real-world flavor you don't often find in this industry.)
If I had to give an overall opinion of the news coming out of Cannes, I could do it in one word: Yawn. It's clear that, although this is an important show, VMware and the other vendors are mostly gunning for the U.S. conference later this year (I would add, however, that attendance is larger than I expected -- about 4,700 folks. Not bad at all, and a good sign for the industry).
First up, the former Virtual Datacenter Operating System (VDC-OS) is now vSphere. Oh boy! I like that they've shortened the ungainly acronym, but vSphere is less than overwhelming. Too much like IBM's WebSphere for my money. Not much new stuff in terms of functionality.
But Alessandro Pirilli at virtualization.info is reporting that CEO Paul Maritz claimed that once vSphere is shipping, there's no reason that 100 percent of your datacenter can't be virtualized. In my opinion, that's simply marketing hyperbole at its best. Until virtualization can make the claim of absolutely native performance, there will still be situations where you must have one OS and one application per box. Remember also that we're right at the beginning of things like network virtualization and I/O virtualization; until that crucial plumbing has been proven as robust and reliable as their physical counterparts, companies should rightly be wary of going 100 percent virtual.
Of more interest to me was the announcement of the VMware Client Virtualization Platform (CVP) bare-metal client hypervisor. This has been known for some time, and it's good to see it formalized. CVP will be part of the VMware View family of desktop products, and available, it appears initially, only with Intel vPro procs. That makes me wonder if a similar deal will be forthcoming with AMD, or if this will be an exclusive contract.
Bare-metal hypervisors are one of the coolest current areas of virtualization research, and I'm interested to see how far they go -- on laptops and desktops for now, but soon coming to a smartphone near you. It also gives more impetus to the desktop virtualization movement, which continues to gain steam. Citrix, of course, beat VMware to the punch with Independence, and you have to think that Microsoft isnt' far behind.
Hmmm, this blog is getting long. I'll break it into another entry or two, coming later. In the meantime, let me know what you think of VMworld Europe's announcements, either below or via e-mail. Am I too blase about it?
Posted by Keith Ward on 02/26/2009 at 12:48 PM5 comments
I need to turn to you, knowledgeable readers, for help in answering some questions. In a followup to yesterday's announcement
by Red Hat about its virtualization roadmap, I asked the company some questions about the new Enterprise Hypervisor.
Specifically, I wanted some details about how KVM would work as a standalone hypervisor, since my understanding is that it's hosted inside the Linux kernel (i.e., a Type II hypervisor). The response I got from Navin Thadani, senior director, virtualization business at Red Hat, threw me for a bit of a loop. He says KVM is a bare-metal hypervisor (also known as Type I), and even tries to make the case that Xen is a hosted hypervisor. Here's his comment in full:
It is a myth that KVM is not a Type-1 hypervisor. KVM converts Linux into a Type-1 hypervisor. There is only one kernel that is used (and that is the Linux kernel, which has KVM included). On the flip side, I can make an argument that Xen is not a Type-1 hypervisor, because the CPU and memory is scheduled by the hypervisor, but IO is scheduled by Dom0, which is a guest (so it's not bare metal). In the KVM architecture, the CPU, memory, and IO are scheduled by the Linux kernel with KVM.
On the other hand, other folks fall into the "KVM is a hosted hypervisor" camp, exemplified by this snippet from Brian Madden:
Xen folks attack KVM, saying it's like VMware Server (the free one that was called "GSX") or Microsoft Virtual Server because it's really a Type 2 hypervisor that runs on top of another OS, rather than a "real" Type 1 hypervisor. KVM responds "So what? Why should we rewrite an OS from scratch when something like Linux is available? And if you want to use a KVM machine as a dedicated VM host, then fine, just don't install anything else on that box."
So, is KVM hosted or not? Is Xen hosted or not? Is Red Hat full of hot air, or are they onto something? I'll be honest and say that I just don't know. Thadani upset my hypervisor apple cart with this comment. But, with your help, I'll get to the bottom of it. Weigh in in the comments below, or e-mail me directly. As a follow-up, tell me what difference you think it makes in the grand scheme of things ("it" being whether KVM is hosted or not, and the same thing about Xen). I'll post comments as they come in. If you do not want your name used with your comments, let me know.
Posted by Keith Ward on 02/24/2009 at 12:48 PM17 comments
You may have heard that Citrix, in what can only be described as an exceptionally bold move -- and risky, calculated gamble, all at the same time -- is now giving away
a new version of its XenServer hypervisor that has capabilities way beyond the typical free hypervisor.
Among the goodies in this version are its version of live motion, called XenMotion, which also incorporates multi-node resource sharing; management for an unlimited number of servers and VMs; and integrated storage management, among other stuff. Wow; that's a lot of virtualization power, and is now free for the taking.
It's hard not to applaud Citrix for making XenServer free. You can't get that much capability for free with any other solution, including VMware. This is no small thing. I mean, if you're an admin with virtualization responsibilities in this recession, don't you at least now owe it to yourself and your company to see what XenServer can do?
Of course, this is exactly what Citrix wants you to think. And when you've got XenServer set up (and if it works well), of course you're going to at least consider Citrix Essentials on top of that, right? And then, since you have so much Citrix in your datacenter, why not Citrix XenDesktop, goes the thinking to which every Citrix sales rep in the world wants you to subscribe.
So goes the theory, anyway. Make no mistake: XenServer is a fine hypervisor, which will be borne out in our next print issue, which should be hitting your mailboxes any time now, if it hasn't already. We prove that unequivocally. It still hasn't really taken hold in the buying public's imagination, however -- at least not yet. Let's face it: When folks think hypervisor, product No. 1 on the radar is VMware's ESX. Product No. 2 is Microsoft's Hyper-V, which is pretty remarkable considering how new it is. XenServer (or Xen in general) is usually next on the list. Will this souped-up version, which may well be enterprise-worthy right now, change that pecking order? Citrix sure hopes so.
The question that must be answered is this: How much free stuff can Citrix give away and continue to be viable? It just released an awful lot of intellectual property, for no discernable return. Sure, there's hypothetically something at the end of all this (an enterprise presence and upsell of other products), but absolutely no guarantee that any of that will occur.
If XenServer now does what you want, why spend $1,500 per server to upgrade to Essentials? Yeah, you'll get StorageLink for better storage management, and lab management software if you want the Cadillac edition, at double that price. But XenServer already has storage management; granted, it's not anywhere near as powerful or feature-rich as the Essentials technology promises to be, but does it pass the "good enough" test? Can you get by for a year or two or more without it? And if you can, why upgrade at all?
That's the scenario all those Citrix reps do not want you to envision. That's why this is so risky for Citrix, and so fascinating for us virtualization watchers. It's clear that hypervisors are commodities; are things like live motion, distributed resource scheduling and P2V going to be next? If so, for what capabilities, exactly, will a vendor be able to charge? I don't know if this is a slippery slope or not, but it sure is exciting.
What say you? Comment below or e-mail me and tell me if a) this will convince you to give XenServer a whirl, and b) if Citrix is making a good or bad move for the long haul.
Posted by Keith Ward on 02/24/2009 at 12:48 PM6 comments
Lots and lots of virtualization news
this week. I'll break it all down in separate blog entries, to keep things manageable.
First, let's take a look at Citrix
. The way to think about this is that it's another version of the
Terminal Services/XenApp (formerly Presentation Server) relationship. Citrix and
Microsoft have carved out a partnership unparalleled in this industry. They make
a product that basically does the same thing: provide remote access to
and delivery of an application (TS and XenApp, in this case). TS is more
for smaller shops; XenApp is the enterprise version, adding functionality that
makes it a better fit in the datacenter.
In the meantime, Microsoft
continues to improve TS and add features, but not to the point that it can be
seen as a XenApp replacement. It purposely cedes the big datacenter to Citrix,
apparently content to make less money for Citrix' benefit. Citrix, meanwhile,
thrives. Presentation Server, for many years, was
its business -- and
still is, in every meaningful way.
But since the purchase of XenSource,
Citrix is starting to see itself as a major virtualization player. It's a
good fit, because XenApp has always been a virtualization product in a way. But
it's thinking bigger now -- much bigger. Citrix knows enterprise at one level,
and thinks it can ride Microsoft's coattails right into the next level of
And it's counting on the fact that Microsoft
will constrain itself to providing the base hypervisor -- Hyper-V -- and the
management component (i.e. System Center Virtual Machine Manager (VMM)), and not
step over the line into the key piece of the Essentials pie, which is its
StorageLink storage technology. Citrix has well-founded hope in this regard, as
Redmond has never been one to dabble too much into storage. It's not hard to
imagine Microsoft saying "Hey, you guys can have that. It gives us more
resources to pour into making VMM the dominant virtualization management app."
Just like TS/XenApp, Microsoft and Citrix hope to divide the spoils without
encroaching on each other's territory.
However, in this situation there's
a difference, and it's a huge one: Citrix and Microsoft pretty much own the
remote app delivery space. That is not the case with virtualization. To get into
those enterprises, Microsoft and Citrix have to offer a solution that's as
good, or nearly as good, as what VMware offers, for significantly less money.
Will Essentials, in combination with Hyper-V and VMM, be enough, technology- and
price-wise, to overcome VMware's lead in reputation and installed base?
As far as Citrix is concerned, a hidden benefit of
Essentials is that it takes some ammunition away from those who charge that
Citrix is all about XenDesktop and nothing else nowadays. This is a big
datacenter virtualization play, and shows that Citrix is continuing development
on that front. Even that, however, has a desktop virtualization (i.e.
XenDesktop) benefit: VDI requires many, many more VMs than server
virtualization, which is still the virtualization king. Where will those desktop
VMs live in an enterprise setting? You guessed it -- the SAN, which StorageLink
is all about. Very smart move by Citrix.
Posted by Keith Ward on 02/23/2009 at 12:48 PM0 comments
The recent speculation
that Cisco is eyeing VMware for a buyout led
Parallels CEO Serguei Beloussov to contact me with a response. Beloussov is one of
the real smart guys in this industry, with a proven track record going back many
years. It's worth listening to what he has to say. Here's his take:
If there’s any truth to the rumor that Cisco is pursuing VMware, I
believe it's unlikely that either VMware or EMC would be sold at the speculated
prices. VMware is performing well and growing quickly, so there's no rush to
sell when the markets are down. It is also unlikely that EMC would sell VMware
separately, as VMware is the main growth engine and largest technology asset for
EMC. On the other hand, Cisco is unlikely to want to own EMC's high-end storage
With that in mind, if a deal
does go through, Cisco will become a dangerous competitor to IBM, HP, Dell and
others, which would create complications that need to be considered. However,
Cisco has done an amazing job of integrating various assets -- I found its
recent integration of telephony companies very impressive. If Cisco did come
to own VMware and/or EMC, the IT industry would change a lot, but it would
have a positive impact on Parallels, as we're an ideal partner for IBM, HP,
and Dell, among others.
Putting aside the P.R. aspects of Beloussov's last
sentence, I agree that Cisco will become a chief competitor for some of the
large companies mentioned. But more than that, they'll become serious rivals
with Microsoft in the virtualization space. I doubt Microsoft would welcome such
a move. It's hard enough trying to keep up with VMware as it is; if the huge
resources of Cisco are thrown into the mix, that could make keeping up with the
Joneses that much harder.
What's your take? As Frasier Crane
says, "I'm listening
Posted by Keith Ward on 02/18/2009 at 12:48 PM1 comments
It's well known that going green saves money as
well as dwindling resources. But how much green, of each kind, is saved?
you're not the math type, never fear -- several vendors offer "green calculators"
that will do the hard work for you. While doing some research for a sister
publication, I came across a number of different calculators for finding out how
much green you can save through virtualization, and other common-sense steps.
Novell's PlateSpin, which makes a
number of green technology virtualization products, has a detailed 'go-green'
consolidation-based calculator. Simply input factors like
number of physical servers, average power consumption per server, cost per
kWh, processor utilization before and predicted utilization after
consolidation, etc. The calculator spits out at the other end savings in kWh
Note that these are rough estimates only. Your mileage can, and almost
certainly will, vary, perhaps significantly. But it serves as a good benchmark.
Other large virtualization
vendors have similar green calculators, including VMware, which is more basic than PlateSpin's.
Avaya has an interesting take on
going green. It offers a calculator to tote up the environmental savings by
turning commuters into telecommuters. It claims that working from home three
days per week saves about 200 gallons of gas per year.
consultancy 1E has an “Energy Savings
Calculator” that computes the greening of
your business from the simple act of turning off computers not in use. It even
translates the savings into reduced carbon emissions and trees.
The Uptime Institute bills itself
as a vendor-neutral organization concerned with increasing efficiency in
enterprise settings. It publishes a "True TCO Calculator," which acts as a guide for
building a high-density, high-performance, green datacenter. This would be of
more value to companies starting from scratch, rather than those making
changes to an existing datacenter.
Many computers already have
power-saving measures built in, but not activated, in the form of Energy
Star-approved computers. The Website GreenerComputing has an Excel
spreadsheet available to calculate how much greener each computer could be
if the power-saving features were activated.
On a personal level, Google has started
the "U.K. Carbon Footprint Project." Going through its calculator gives you an idea of how much pollution you
and your household contribute to Great Britian. Then, you can input your
location and statistics, and compare your usage vs. others who have done the
same. Different-colored balloons separate the good from the bad -- from the
ugly. Very cool.
Posted by Keith Ward on 02/18/2009 at 12:48 PM0 comments