There are many ways to compare alternatives, but for simplicity, things often come down to cost. Unfortunately, it’s not always easy to separate the true comparisons from creative marketing — let’s take a look at some key factors that can help sort the apples from the oranges.
Case Study: The Very Expensive Laptop
Imagine you are looking to buy a new laptop, equipped with various accessories and software. Part of the decision is to select an appropriate office suite to include in the configuration:
There are a number of office suite options: the less capable choice adds $80 to the price, and the one you really need to do your work adds $230 — almost three times the cost. Outrageous!
Laptop with Professional Suite Costs Three Times More!
That’s the headline anyway. Obviously, it’s far from true because it only focuses on one component and misses out on the big picture.
From laptops to cars to private clouds — cost comparisons that focus on just one portion of a solution don’t make much sense.
Compare Entire Solutions
The truth is, comparing just one component of a solution is arbitrary and meaningless. To truly compare two alternatives, one must look at the complete solution.
In the case of our laptop, it turns out that the professional configuration costs about 17 percent more than the one that doesn’t meet all requirements. A far cry from the three times hyperbole. Not only that — it has capabilities the cheaper alternative cannot match, a difficult aspect to quantify when performing a pure cost comparison.
Practical Application
Now we’ve seen how cost comparisons can easily be manipulated by focusing on just one aspect of a complete solution. Are there any real-world examples of an out-of-context cost comparison without regard for the big picture? Yes…
Allow me to introduce the Microsoft Virtualization Cost Comparison Calculator:
By using the exact same trick to artificially compare the cost of VMware vSphere to a handful of System Center products, the Microsoft Virtualization team has settled on a marketing message: VMware costs N times more than Hyper-V.
Three Times More? Five Times More?
Would a reasonable person choose to spend over three times more for something if a suitable alternative is available? No! Just as a vSphere solution is not multiple times the cost of Hyper-V, neither is Hyper-V a suitable replacement for the industry leading virtualization platform.
For a true cost comparison, with line-by-line transparency, check out the VMware Cost Per Application Calculator and see for yourself how the complete solutions compare.
We know VMWare invented the commodity PC virtualization; however, in 2011 and in the SMB space, all these comparisons of technical superiority are non-factors. Why? MS has it out of the box to turn on, and the technological superiority of VMWare has shrunk and continues to shrink with every release of Hyper-V. Why did VMWare begin to release ESX platform previously, because of Hyper-V being free out of the box, and since then MS has increased features and capability. I can honestly say Hyper-V in SMB is perfect in terms of skills required to maintain it and the cost involved in Backups and Recovery – all that matters in SMB is cost and easy accessibility to the product. VMWare should focus on ideas to combat that growth – how to get VMWare products to the mass?
Meant to say that due to initial release of Hyper-V being free, VMWare began to offer ESXi (smaller and less featured than ESX) as hosts of other companies – Citrix’s Xen, and Red Hat’s KVM offering their virtualization products freely. The primary difference between all is the management applications, and third party costs to support the particular Infrastructure. But just as VMWare commoditized x86 virtualization in early 2000, Microsoft is commoditizing the virtualization landscape in shorter period of time because of the cost to maintain / support it, ease of use out of the box, and availability out of the box. Cost wise and Support wise, Hyper-V is already ahead in the SMB space.
Both tools are biased :p
I just ran the comparison against Citrix. The calculator said I needed 5 Xenserver hosts to do the work of 3 VMware hosts… That doesn’t sound right.
Settings: 50 apps, Enterprise Plus, everything else medium.
…see my post – the entire tool is predicated on memory overcommit. Since XEN doesn’t have that feature they are likely saying that the ESX servers are more dense and you’ll need two more hosts to have the same number of appliations…er, VM’s.
The tool is so fundamentally flawed – and as I mention, even more so now with Dynamic Memory with Hyper-V.
Even in the cost example they used at VMworld – they make it all up by saying that with any other platform you need more hosts which of course increases the hardware and software costs so that they can catch up to VMware’s costs…all the while saying that overcommit is the hero here and solves all the cost problems. In many of their examples they claim to get outrages consolidation rates as well – much higher than anything you typically seen in production.
For heavy Windows shops like us…VMware cost more plain and simple. VMware it technically better…if you use the parts that are better. The gap has closed so much that it got hard for us to justify VMware anymore. SCVMM 2012 closes it down even more.
Cost per App calculator is total crap. First of all, how pretentious of VMware to presume that a virtual machine is an ‘application’ to begin with.
Most importantly, the cost per app calculator doesn’t factor in the most important thing – WHAT IS RUNNING IN THE VM?? Their supposition is, if you run a bunch of VM’s that are sitting doing nothing, well, you can get more on ESX than anyone else (the whole calc is based on memory overcommit)…hmmm…who does that? What if you turn on the wonderful “Fault Tolerance”? What if you are doing guest clustering where VMware recommends you disable memory overcommit? What about Dynamic Memory in Hyper-V which completely makes their calculator even more worthless?
I attended a compete session at VMworld where the presenters showed a slide and SAID THAT IF YOU DISABLE OVERCOMMIT HYPER-V/MS SOLUTION IS LESS EXPENSIVE. This gets even more lopsided once you factor in dynamic memory since they refused to talk about it at VMworld because it was ‘beta’.
Jesus this blog is crap. I attended an event last friday and had conversations with 4 other CIO’s of large companies, all who have Hyper-V either in production now or in their lab doing testing – several for no other reason than, “VMware is killing us with the cost…”.
Proof is in the pudding Eric – you can spin it every way you want and try to correlate it with an Office example – but you aren’t kidding anyone…
Careful mentioning SC-VMM 2012 .. and unproven.. Yet to be released (except in beta) version of a cloud solution. I attended TechEd this year and saw the feature set expected in the offering and I am impressed that they now meet nearly if not all Gartner standards of an enterprise-ready cloud solution. But if you want to compare apples to apples then you may want to put that solution up against the new vCloud offering coming out from VMware later this year (version 5 of their Hypervisor). If you compare Microsoft with VMware TODAY there are many differences in not only cost but also VM density per host (often due to hypervisor footprint, memory management, efficiency of SMP), Host hardware capacity, VM size capacity, layer2/3 Network features, security… and the list goes on. All of which you may want to compare with the like-generation solution. To rightly judge each solution you may want to start with a common requirement criteria. As mentioned above “in the SMB space” and set those criteria. Do the same for small, mid and large enterprise deployments. Compare direct features: live migration to vMotion, (until 2012 there is no VM load balancing solution from MS) to DRS, Hypervisor aware network and system security (native and 3rd party vendor), MSCS vs. HA&FT (individually then combined), Network I/O comparison on like hardware, Storage I/O on like protocols (minus NAS as it is unsupported in current MS solution), SRM vs. MS DR solution (until VMware ver.5 both are dependent on storage vendor sync technology). Everyone seems to want to market their products, that’s fine… but “we” the technical community need to do an honest side by side bake-off. You need similar requirements (or sets of requirements for small, med and large deployments), identical hardware, and industry standard testing tools that the most relevant. The technical personnel are often drawn in to “false advertising” marketing wars. We need to take the Kool-Aid IV’s out of our arms and do the actual work. Some features fall down or are often not as advantageous as were hoped (FT comes to mind as an unfortunate disappointment, but so is current lack of VM load balancing). We can do better than the hype.. we can do the work.
all good points..
the real question is (and since this post focuses on price), do you want to pay even MORE with VMware – their cloud model with vDirector, etc… goes to a PER VM price. Some of the rough ballpark numbers I’ve seen on this at this point are outright scary…if you think VMware was sucking your budget dry before…wait until you go ‘cloud’…
So typical…people complain about VMware costing too much in comparison to Microsoft!?! Microsoft? Really? W a k e U p ! It’s 2011 and they are only trying now because they have no choice. There is no passion (for making a difference) in their innovation, technology, leadership. So will I spend a little more for VMware, no matter the calculator? Yes. Why? Because I TRUST them and they do QUALITY work.
Quality matters in everything. When did people stop caring about quality! I have been burned by Microsoft too many times over the last 15+ years. Quality, leadership, and innovation…they’re now playing catch-up with everyone. How can they innovate competing against the innovators Google, Apple, VMware…arrogance. The SMB market is being played by a big marketing machine in Redmond. Think about it.