Best Argument: VMWare
The disruption of virtualization hegemony
Jason Perlow: For the first time in four years, Windows Server and Hyper-V are not only at parity in terms of basic enterprise virtualization functionality with VMware vSphere, but in a number of respects actually exceeds it in terms of features offered and encompasses the functionality of a number of other VMware products that would be considered expensive add-ons, as well as 3rd-party enhancements for VMware that you’d have to go to other vendors and spend big bucks for
The obvious manageability, scalability, ease of deployment and significantly reduced total cost of ownership advantage of Windows 8 Server and Hyper-V will finally force VMware into a corner as CIOs examine the licensing bottom line and the value to that Windows Server 8 and Hyper-V brings to their environments as a complete end-to-end virtual infrastructure and private cloud solution in a box, for a fraction of a cost of the industry leading enterprise virtualization stack.
For those of you who still feel that VMWare's position is safe, I ask that you remember in the mid-1990's when Novell Netware was network operating system king and the newcomer, Windows NT seemed unfit for the enterprise. It's going to happen again, folks. Get your affairs in order, VMware. 2013 is going to be remembered as the year that the virtualization hegemony was disrupted.
Microsoft a minority player
Ken Hess: VMware has the clear advantage in the x86 virtualization market, since it began the market in 1999. Prior to this breakthrough, virtualization was something that IBM did on mainframe computers. VMware brought virtualization to the desktop and to the data center for everyone.
Microsoft is a late arrival in the virtualization market. Although, tardiness doesn't imply failure, I believe that Microsoft hopes that Microsoft shops and SMBs will embrace Hyper-V as their virtualization solution. Microsoft has a chance to claim a small percentage of the market's growth over the next few years but will never encroach on VMware's pole position.
Even if you take VMware out of the picture, Microsoft is still a minority player in the x86 virtualization space. Red Hat's KVM and Citrix's XenServer are two compelling and capable alternatives. Hyper-V has little chance of market penetration in light of its well-established competition.
For my money, there's VMware and then there's everyone else.
Great Debate Moderator
How do we tell who's winning in 2012?
In 2012, what are going to be the most important factors to watch in the Hyper-V versus Vmware race to tell who's winning?
Hyper-V will gather significant market share
Again, this isn't about winning the entire market, the focus of this debate is "will the MS hypervisor make important inroads" and will it have new wins. I believe when IT environments start looking at how they need to consolidate imporant, performance based workloads such as SQL server, Sharepoint, and Exchange, and start thinking about creating large shared infrastructure that can be easily provisioned, then we are going to see Hyper-V gather significant market share. It doesn't make sense to put these types of workloads on a much more expensive virtual infrastructure solution like VMware.
The consumer wins.
2012 isn't the deciding year. 2013 won't be either. Your probably looking at 2015 before you'll see any sort of uptake or adoption of Hyper-V, if any. Why? Because of maturity. VMware is a mature product currently in its fifth major iteration. This isn't VMware's first rodeo, as we say in Texas, and by the time any companies might begin to embrace Hyper-V, VMware will have produced another major version or two, while Microsoft will still be patching Hyper-V version 3 with weekly "Patch Tuesday" updates or releasing its R2 version. It's really a question for the consumers who, on this scale, are your cloud service providers, your major IT outsourcing companies and your large independent companies whose CXOs make the real decisions about technology direction. There, of course, will be a few mavericks who adopt Hyper-V right away and I wish them well. And, I hope they have a large budget allocated for additional hardware and stand-by staff. They'll also need Microsoft Hyper-V support on speed dial.
Great Debate Moderator
Thank you for joining us
Perlow and Hess will post their closing arguments tomorrow and I will declare a winner on Thursday. Between now and then, don't forget to cast your vote and jump into the discussion below to post your thoughts on this topic.
Great Debate Moderator
What about the Netware comparison?
Perlow brought up the NT versus Netware race as comparable to Hyper-V versus Vmware. How and why does this comparison work, or why doesn't it?
Because VMware only has virtualization
It works because everyone thought that Novell had an unlimited sustained business model for their network operating system, but all they had was their NOS, nothing else. The same can be said for VMware because they only have a virtualization stack, they do not have a NOS, they do not have end-to-end managment, and they do not have the application infrastructure. When compared to the overall MS ecosystem, VMware only has virtualization.
VMware is no Novell.
The comparison doesn't work because when NT 3.x workstation hit corporate desktops and NT 3.x server hit corporate data centers, it made sense to integrate and standardize your servers and your workstations on a single platform because of their similarities. It was right for the time. Netware was really on its way out by the mid-90s in favor of server operating systems that did more than just serve files and provide print services. They were great for their time but their time had come and gone. Hyper-V, on the other hand, is really an unnecessary effort on Microsoft's part. These days data centers and desktops need no such consolidation or standardization. Management is (or should be) web-based so that workstations can be Linux, Mac, Windows, Chrome or mobile devices.
Great Debate Moderator
How about "private cloud"?
I hate the term "private cloud" but a lot of what is being called private cloud is just a highly virtualized private network. Will the ability to intelligently connect the private cloud and public cloud turn into an important factor and an advantage for Microsoft, which is in the process of building a massive public cloud infrastructure?
Microsoft is the only company that can
The cloud, both private and public, is about managing applications and efficiently leveraging resources to do so. Microsoft is the only company that has in-production private and public cloud offerings today. Organizations are going to need the flexibility of running some of their their infrastructure in-house and some of it outsourced to the public cloud, in a fairly dynamic manner and managed in a completely transparent fashion, particularly if they have variable capacity needs. With Hyper-V, that's actually an option today. With VMware, not so much.
If you don't mind MS-branded handcuffs.
For some companies, their private cloud will be a stepping stone to the public cloud. For others, it will be one-half of their hybrid cloud infrastructure. Having the ability to integrate the two is a feature that is absolutely essential. If you want to lock-in to a vendor for public/private cloud services, then Microsoft might be for you. Sure, they'll tie-in your private cloud to their public cloud but what if you want to go with another vendor for your public cloud tie-in? Will that be a seamless connection? What about those who've invested years of development effort in Amazon's public cloud? Unless Microsoft can give up on its proprietary, locked-in world view, then that's another gotcha in the Hyper-V hype machine.
Great Debate Moderator
Is the pain justified?
Does the pain of migrating justify the cost-savings? Explain.
With the VMware pricing model, the continued growth of a VMware environment will inevitably result in higher and higher costs. Essentially, a VMware infrastructure leads to a variable cost for virtualization in an age where very few IT environments have variable budgets to play with. A Microsoft-based infrastructure brings the predictability and scalability to costs that customers have now. Even with a high migration cost of $600/VM, migrating from VMware can produce a savings of $4 million over 5 years, for a 1000 VM environment. In a typical customer environment, the migration cost is in the range of $100-$200 per VM.
Absolutely not. It would be foolish for a company to migrate to Hyper-V from VMware on any kind of scale. First, because of the sheer costs of doing so from a labor perspective. Second, from a hardware point-of-view, you'll probably have to upgrade your hardware to run Hyper-V. Yes, even Server Core. Think of the Celsius to Fahrenheit temperature conversion formula when you think of converting from VMware to Hyper-V; double your current hardware requirements and add 32. This is the step known as the "Microsoft Tax." When have you ever known of any Microsoft upgrade that didn't require a corresponding hardware upgrade? Also add in that you probably won't get the same virtual machine density, so add in more hardware for that gotcha too. Third, are you going to throw away everything you've built with VMware to migrate to Hyper-V? Your licensing, your staff's training, your networking, your storage and all that you've built will have to be trashed. Finally, what's your service stability worth to you? If you can afford to weather the problems of migrating to an unproven technology, please be my guest.
Great Debate Moderator
The migration issue
Let's talk about the installed base issue. Vmware is the runaway marketshare leader. How easy is it to migrate from a Vmware infrastructure to a Hyper-V infrastructure?
I don't see people throwing their existing investment in VMware infrastructure out in the garbage -- after all, this debate is about whether Hyper-V can make significant inroads, not completely displace the competition. However, that being said, its very easy to migrate form a VMware-based virtual infrastructure to a Microsoft-based one. First, System Center can manage both a VMware and a Microsoft environment, allowing co-existence and easy migration. Additionally, Hyper-V comes with Windows Server, so most customers already have it. VMware administrators can jump right into Hyper-V by leveraging their existing virtualization skills and combining them with the Windows skills they already have, making the transition very easy. It is probably also worth mentioning that Microsoft has a distinct advantage with Hyper-V over VMware because it owns the source code to Windows, and thus can achieve much higher levels of performance and integration by using their own hypervisor. Additionally, from a skills acquisition and training perspective, Hyper-V will be built into the Windows 8 client, allowing anyone with a standard, low-cost commodity PC to learn how to use Microsoft's Type-1 hypervisor on their desktop, eliminating the need for products like VMware Workstation. Server 8 and Hyper-V can also be installed on commodity 64-bit PC hardware, unlike ESXi, which primarily requires VMware certified enterprise-class servers to run.
Two steps forward. Three steps back.
I've never done it nor do I know anyone who has. My guess is that currently, you'd have to reconstruct all of your virtual machines in Hyper-V. VMware, on the other hand, will use a Microsoft Hyper-V disk image for a reverse conversion. VMware realized early on that people would want to do that. There are third-party applications that will perform live migrations from VMware to Hyper-V. What you really need to know is how difficult will it be to change back to VMware from Hyper-V, once you realize your mistake.
Great Debate Moderator
Does Hyper-V really save money? How much?
When we compare Hyper-V to Vmware, how much of a cost savings are we typically talking about for the Microsoft solution?
VMware costs can run 10 times Microsoft's costs
Ken, "I haven't put pencil to paper" should be counted as a gaffe on your part, seriously. It makes it sound like you haven't done your homework. When you include the management solutions from both vendors, VMware costs can be at least 10 times Microsoft's costs, depending on how many VMs are on each server. For example, in a very large environment with 5000 VMs with a very high consolidation ratio, such as 12-18 VMs per host, VMware is more than 12 times as expensive. You can perform the calculations here for your own environment at http://cloudeconomics.cloudapp.net Also another note: Ken is in error with this statement that Hyper-V is not a free solution. Microsoft has been shipping a free version of Hyper-V Server since October of 2008, and I have received confirmation from Microsoft that they will be shipping a free Server 8 version of Hyper-V Server as well. http://www.microsoft.com/en-us/server-cloud/hyper-v-server/default.aspx
You get what you pay for with VMware.
I haven't put pencil to paper comparing the two but I don't really think you can measure it in such black and white terms. If you are talking about starting from absolutely no infrastructure, the Microsoft solution will seem cheaper on the surface. But how about four years into your plan? I think you'll see that both technologies will cost about the same over the long-term. VMware seems expensive but it's really the difference in investing in a tried-and-true technology or betting on a longshot. I know that Microsoft touts their new Hyper-V as a "free" solution but it isn't free. You still have to buy the license for the base operating system from Microsoft. And, Frankly, I've never seen anything free from Microsoft. Besides, if we're talking free, why didn't decision makers flock to Red Hat's KVM for free server virtualization? That technology actually is free and it has higher performance than Hyper-V and very high virtual machine density. There are many factors to consider in calculating costs such as hardware, support and maintenance as well as licensing. TCO isn't two-dimensional as many believe. My lack of concrete calculations isn't a "gaffe," because you can't calculate a production-level solution like VMware 5 when MS's product hasn't been released yet. MS doesn't even know what it's going to charge for Windows Server 8 yet. It's pure speculation at this point. If we're going to compare a production system to an unavailable version, then let's do it fairly: Windows Server 8 Hyper-V vs. VMware vSphere 6. You can't just count license fees. You must also factor in VM density into your calculations, which you also don't have any numbers on for comparison. Any calculations made now are pointless. Not in error. Hyper-V isn't free. You have to purchase the base OS, which is far from free. There might be a "free" version but you don't get any Windows licenses with it. That's useless. There are other free hypervisors available if you're looking for free ones. To say that Hyper-V is free is really not being honest. If it's totally free, then can I run it on Linux? No, I have to buy a MS licensed OS first. That really isn't free.
Great Debate Moderator
The cost issue
How much is cost really an issue when we talk about virtualization? After all, when it comes to enterprise data centers, isn't virtualization software a small fraction of the overall cost?
The advantage of Microsoft's processor-based cost model
It is hard for any company to simply ignore the costs of virtualization. Enterprises are starting to realize that with VMware???s new consumption-based pricing model, the more they scale their environment and the more they increase server density, the more they pay. This is radically different from Microsoft???s processor-based cost model, which provides customers with unlimited replication capabilities. Additionally, Hyper-V delivers both OPEX (Operating) and CAPEX (Capital) savings. One of the biggest datacenter costs is OPEX and, with System Center 2012 improvements like automation, automatic remediation and strong cross-component integration, customers can significantly reduce their OPEX costs. Because of its focus on Infrastructure, VMware can only deliver CAPEX savings. While these costs have significant impact on the Total Cost of Ownership for enterprises, it should be noted that Hyper-V starts to really prove its case in a cloud-provider or strategically outsourced virtual infrastructure. When a Cloud provider or datacenter/hosting outsourcing business like an HP, an IBM, a Verizon or a Rackspace needs to provision thousands of virtual machines for customers, they are essentially selling service level agreements (SLAS) and not platforms. In a completely strategically outsourced infrastructure, the customer primarily cares about the SLA and that they can be provisioned server resources as quickly as possible. These customers aren't doing any of the infrastructure management. When your VMware licensing costs start to run in the millions of dollars, Hyper-V begins to make a heck of a lot of sense.
What's the cost of downtime?
You're right, Jason (Hiner), cost isn't such a big issue, when you talk about virtualization. It isn't really that much of an issue compared to other costs. C-level execs look for the right tool for the job. The proven tool. When it really boils down to dollars and cents, what's the cost of downtime for your server infrastructure? It's for this reason that customers bank on Oracle, for example. No, it isn't cheap but it's the best RDBMS there is. VMware has that same level of competency. It isn't the cheapest technology but you get what you pay for in this particular instance. A lot of people are saying that VMware has become the Oracle of Virtualization. They say it like it's a bad thing. It isn't. Everything works on paper but the proof is, as they say, in the pudding or the data center, in this case. If you're merely quoting price as a strategy, then Red Hat's truly free KVM or Xen are better answers than either Hyper-V or VMware. Amazon uses Xen. Rackspace uses Xen. Actually, if you want ultra high density and inexpensive virtual servers, then Parallels beats everyone. But, for rock solid stability and true enterprise-level features, VMware has the upper hand here. Tell someone who uses Oracle that they can use MySQL (free) or PostgreSQL (free) or MS SQL Server (not so free but not so expensive as Oracle) and they won't change. Why? Quality. Reliability. Stability. But, never Price. However, I'm still not convinced that there is a lot of price/feature disparity between the two.
Great Debate Moderator
Ken brought up other virtualization products from Citrix and Red Hat. Are these going to be major players in the data center over the next 2-4 years or is this destined to turn into a two-horse race?
Other products will be the exception, not the norm
I believe that Red Hat and KVM-based offerings from the Open Virtual Alliance will offer a compelling solution for cloud providers who for whatever reason, will want a 100 percent open source virtualization solution and will want to roll their own infrastructure or rely on systems integrators to tie in most of the pieces. But I think this will be the exception rather than the norm -- the majority of enterprises are unlikely to go this route, they will want fully supported vendor solutions if they end up doing things in-house. Hyper-V is the clear commercial alternative to VMware. Still, I see customers implementing heterogeneous virtualization environments with different silos of hypervisors used to solve specific virtual infrastructure problems.
One leading VMware horse and a distant pack.
Since I brought it up, I'll tell you that it's already a two-horse race: VMware and the rest of the pack (The other horse). Companies have standardized on the best technology, which is VMware's. Microsoft might gain but it will be a small percentage of companies who are virtualization dabblers and Microsoft bigots. Though Hyper-V has some compelling features, VMware has them plus it has everything else too: Management, performance, stability, security. First, VMware is now what we used to call ESXi, which is a very light bare-metal hypervisor. Some people complain about VMware's pricing but those are not the decision makers, they are the techies. People who have the financial responsibility for SLAs and customers aren't going to bank on an unproven technology. When the techies are home playing video games or geeking out over a new gadget, the C-level executives are planning and constructing next year's budget and their long-term plans for expansion and they want stability, scalability and VMware's experience behind that.
Great Debate Moderator
Are both of my debaters online?
Ready to rumble
Good luck, Ken
Ready to Rope and Tie
This Texas boy is ready to ride.
Great Debate Moderator
The state of virtualization
Sum up the current state of virtualization. Is there still a lot important stuff left to virtualize in the data center, or are we heading into maintenance and optimization time?
There's much work left to be done
Currently available statistics show that only around 50% of workloads (e.g. Exchange, SQL Server, SharePoint, etc.) are currently virtualized. In my experience as a infrastructure architect and data center migration engineer, this data is consistent with the applications and workloads that I have run into. While a good number of servers in the datacenter have been virtualized, there is still a significant amount of work left to be done.
Much more to do for VMs.
I believe that virtualization is at an inflection point in that heavy operating systems for hypervisors and for VMs are out. There's still much to do in the virtual desktop infrastructure (VDI) space and in server virtualization. Operating systems need to move out of the way in favor of deployable workloads. Working on hypervisors is really the wrong direction to go. It's the VMs that need the work. As far as "important stuff to virtualize" is concerned, yes there's much work yet to do. VMware proved that, with the proper underlying infrastructure, you can virtualize just about everything. Their concept of virtual networking is the most elegant bit of virtualized hardware and its related services. VMware is really on its way to creating a truly plug and play virtual infrastructure. I don't think there will ever be a maintenance and optimization time because there's too much to do for the creation of the ultimate virtual data center. That time is still at least a couple of years away.
Great Debate Moderator
What are the functionality differences between Hyper-V and Vmware? Is there true feature parity?
Hyper-V features far exceed what VMware offers in the box
Windows Server "8" Hyper-V brings to market a number of features that far exceed what VMware offers in the box. These include: Hyper-V Network Virtualization: To deliver a hybrid cloud solution for customers and hosters need a secure, multi-tenant solution. Share Nothing Live Migration: This provides the ability to Live Migrate Virtual Machines with nothing but an IP address at each end of the network connection. Hyper-V Extensible Switch: VMware provides a closed, replaceable switch. Because of this closed architecture, there???s only one option, the Cisco Nexus 1000. With Windows Server "8" Hyper-V you now have access to an open, extensible switch, providing customers with flexibility and choice. Encrypted Cluster Volumes: With Windows Server ???8??? Hyper-V, you can create the largest and most secure clusters using Bit Locker encryption to protect your data at rest. Again, VMware offers no such option. GPU-enhanced hardware-accelerated VDI: Hyper-V in Windows Server ???8??? is the only virtualization solution with in-guest hardware GPU, providing for an extremely high-performance Virtual Desktop (VDI) experience when using Microsoft's Remote Desktop Connecton RemoteFX feature that closely resembles localized desktop performance. Virtual Infrastructure Scale: Windows Server ???8??? Hyper-V offers the largest clusters (64 Nodes per cluster for Hyper-V vs. 32 for vSphere), the most VMs in a cluster (4000 VMs per cluster for Hyper-V vs. 3,000 for vSphere), and the largest Virtual Disks in the market (64 TB for Hyper-V vs. 2TB for vSphere). Hyper-V also permits 1TB of RAM per VM, 160 virtual processors per virtual host and up to 32 vCPUs per VM, and that applies to Linux VMs as well, not just Windows ones.
DISparity not Parity for Microsoft
It's very difficult to compare a list of features between two products like VMware's vSphere 5 and Microsoft's Hyper-V 3. Why? Because Hyper-V 3 isn't available until next year and vSphere 5 will go through at least one major feature update between now and then. Hyper-V really comes out of the gate in catch-up mode and it will always be in catch-up mode compared to VMware. Plus, I've seen some awesome features proposed and included in Microsoft's beta systems over the years only to be disappointed at release time. And, if you're a Microsoft fan, you have to admit that you'll probably wait until the Windows Server 8 R2 version and a couple of service packs before you take the plunge for anything production-oriented. If you don't, then you haven't leared any lessons in the past two decades. Hyper-V 3 is Microsoft's Great Virtual Hope because they realize that heavy, non-virtual operating systems are about to die a painful and malingering death by attrition. Companies that produce operating systems need to focus their efforts on creating them as virtual or at least virtual-aware.
Great Debate Moderator
So why hasn't Hyper-V already crushed Vmware?
Since as early as 2008, we've been hearing that Microsoft's Hyper-V was going to crush Vmware. However, in the big virtualization run-up of the last few years, it's been Vmware and not Microsoft that has been the big winner. Why?
Has now achieved feature parity -- and beyond
Despite excellent performance and overall value compared to its much more expensive competitor, the product was missing a number of key virtual infrastructure management and high availability features that were necessary to seal the deal for large enterprises in order to consider it to be in the running for x86 server virtualization platform of choice. This included things like Live Migration, which was introduced in version 2.0. Version 3.0 that comes with Server 8 is now at feature parity with vSphere 5, and also has a number of important enterprise features that VMware does not offer. Additionally, early versions of Hyper-V were not well-suited to hosting Linux-based solutions when compared with VMware, so this stifled its adoption in heterogenous environments which had significant Linux-based infrastructure. All of this has now changed since the synthetic device drivers in Microsoft's Linux Integration Tools became open source and have now been integrated into the mainline 3.4.x Linux kernel. This means you simply just install your Linux VM on Hyper-V 3 on Windows Server 8 as if it was native and running on the metal, without having to install or compile any additional drivers. Even VMware cannot claim to have this level of openness with their tools integration.
VMware is the Victor not the Vanquished.
VMware has the experience and the customers. It's a proven technology. VMware innovates at a somewhat conservative pace but you can bet on its stability because of that pace. VMware's position won't be easily surpassed, even by Microsoft. Microsoft's Hyper-V version 3 is a good product but it has yet to prove itself. If I were in product marketing at Microsoft, however, I'd tell you that it's a VMware killer. That's just good business. Also, if you think that VMware isn't going to continue to innovate and produce in the field that it created, you're sadly mistaken. Microsoft's Hyper-V is basically an attempt to enter a committed market. Even Red Hat and Citrix realize that it's a VMware-centric world but I don't think they're gunning for VMware like people assume that Microsoft is.
Hyper-V: Best in class and more
In grand, blustery Texas form, my opponent has relied on the classic tactics of FUD to defend the cash cow that is VMware. He offers no solid technical arguments as to vSphere's superiority as an enterprise virtualization stack as it compares to Microsoft's Hyper-V; all he can do is cast aspersions as to the product's quality and the need to maintain an IT quid pro quo.
Neither does Ken offer a truly compelling argument for vSphere 5's superior Total Cost of Ownership or the VMware platform's ability to deliver actual significant dollar reduction in operational expenses by providing an end-to-end virtual infrastructure management, multi-tenancy cloud and network operating system solution. He can't, because only Microsoft can.
Why the FUD? Because he has no defense against Hyper-V's superior value except to say that you should continue to throw millions of dollars at licensing fees at a product which is going to have a virtualization feature set that is not only going to be commoditized by significant efforts at Microsoft, but also by Open Source solutions such as KVM.
VMware has a single area of specialization -- Virtualization, for which it charges a heavy premium. And that single area of specialization is an exposure when your competitor has a complete solution across the entire stack and your entire reason for being is to provide virtual access to your competitor's operating systems.
With Hyper-V in Windows Server 8, Microsoft will offer you a best in class hypervisor along with great deal more built-in features for a heck of a lot less money, which has become scarce in today's shrinking IT budgets. These financial constraints have been pressuring CIOs to do a lot more with a lot less, and it's a trend that is not going to change anytime soon.
The bottom line is that Hyper-V has been a stable, proven, high-performance virtual infrastructure solution for at least the last two years, and it has already been gaining some traction in enterprises for its ability to consolidate high-performance Windows workloads. Environments should not throw out their existing VMware infrastructure, but if your organization is looking to grow its virtualized Windows footprint, you'd be foolhardy not to give Hyper-V a very close look.
With the feature set I have described in my arguments above, which includes superior scalability to its competitor, not to mention the bottom line -- a tremendous cost savings when applied at both a large enterprise and small/medium business scale -- the answer to the question of the debate at large "Will Hyper-V make inroads against VMware?" is without any doubt a resounding yes.
VMware: More soldiers than Microsoft
A scene from the Movie, 300.
Daxos: I see I was wrong to expect Sparta's commitment to at least match our own.
King Leonidas: Doesn't it? [points to Arcadian soldier behind Daxos]
King Leonidas: You there, what is your profession?
Free Greek-Potter: I am a Potter, sir.
King Leonidas: [points to another soldier] And you, Arcadian, what is your profession?
Free Greek-Sculptor: Sculptor, sir.
King Leonidas: Sculptor. [turns to a third soldier]
King Leonidas: You?
Free Greek-Blacksmith: Blacksmith.
King Leonidas: [turns back shouting] Spartans! What is your profession?
Spartans: WAR! WAR! WAR!
King Leonidas: [turning to Daxos] You see, old friend? I brought more soldiers than you did.
VMware has more soldiers than Microsoft does. Virtualization is VMware's only profession. Microsoft is a highly diversified software company that produces hundreds of different software programs including games, desktop applications, server applications, operating systems, servers and more. VMware invented x86 virtualization. Hyper-V is Microsoft's attempt to remain relevant in this cloud-oriented, virtualization-focused world.
Microsoft, like the Greeks, may share the battlefield alongside a professional virtualization company but, in the end, VMware's phalanx will better protect your investment against the Xerxesian virtual machine horde that it faces.
Tough to bet against VMware
This debate was quite a doozy. Both Perlow and Hess made a strong case, and this is a topic that has very important implications for the future of the data center.
Hess is right that VMware's installed base is a huge competitive advantage that won't be easy for Microsoft to overcome, but Perlow is on target that Microsoft has the potential to win big with the integrated software stack. Microsoft could make inroads with service providers based on features and cost, but it's likely going to have a difficult time dislodging VMware in big companies that have invested heavily VMware products in recent years.
For that reason, it's tough to bet against VMware and so I've got to give Hess the razor-thin victory on this one.