The Hoard Facts

Blog archive

VMware Calculator Aids, Assists Archrival Microsoft

Microsoft is in full gloat mode after running across VMware's cost per application calculator and using it to prove that vSphere 5.1 is more expensive than Windows Server 2012 and System Center 2012 combined. In its Nov. 15 Server & Cloud Blog, Redmond was quick to point out that this is not just about the cost of license acquisition, but also the CAPEX costs for items such as power, space and server hardware.

Anxious for customers to view the situation first-hand, Microsoft urged them to plug the following values into the calculator:

  • Number of VMs: 100
  • Virtualization host type: Server
  • Network storage type: iSCSI SAN
  • Compare to vendor: Microsoft
  • VMware vSphere 5.1 edition: Enterprise Plus
  • Management deployed on physical or virtual: virtual
  • Electricity: low
  • Real estate: low

At this point, Microsoft says it is important to understand that the aforementioned values are not random, but represent a common datacenter virtualization scenario.

Turning to results, Microsoft says that in this scenario, the cost-per-application to virtualize 100 apps using VMware vSphere 5.1 Enterprise Plus edition is 19 percent higher than with Microsoft Hyper-V and System Center 2012.

According to Redmond: "However, we firmly believe you will save far more with Microsoft. According to VMware's calculator, Microsoft's total software cost ($974) is much lower than VMware's ($1,491), but the infrastructure cost ($1,198) is higher than VMware's ($1,083) infrastructure cost.

"Why is this the case? This happens because the calculator assumes that a VMware ESXi host can run 20 percent more applications than a Microsoft Windows Server 2012 (Hyper-V) host -- an assumption with little credibility or real-life customer evidence."

Microsoft goes on to say that the calculator bases the "run 20 percent more applications advantage" on an August, 2011 report VMware commissioned to a third party that compares vSphere 5 to Microsoft Hyper-V 2008 R2 SP1, and that Dynamic Memory, which was initially unveiled in Hyper-V 2008 R2 SP1 has been improved in Hyper-V 2012. Microsoft calls that "a fact that VMware simply ignores." Continuing, Redmond says you cannot apply the results of a test performed with a previous version of the product (Hyper-V 2008 R2 SP1) to the current version (Hyper-V 2012), and assume that "everything remains constant."

"Why would VMware choose to base results on older technology?" Microsoft asks, referring to a white paper about the advantages of Hyper-V over vSphere 5.1 which Microsoft says "provides the likely answers."

Microsoft goes on to point out that the August 2011 report fails to build on a realistic customer scenario, declaring, "VMware was able to show an 18.9 percent performance improvement (and higher consolidation ratio) only when using many VMs running the exact same workload with the exact same data and overcommitting the host -- under specific VM configurations and settings. Ask yourself: do you ever run the exact same workloads with exact same data on a host and overcommit in a production environment? Your most likely answer is no."

Microsoft further notes that when users purchase Microsoft's Windows Server 2012 and System Center 2012, they get a complete private cloud solution, and any realistic cost comparison with VMware should include VMware's private cloud solution, vCloud Suite 5.1, which was recently introduced.

Having thus dismantled VMware, Redmond goes for the kill shot, saying if you re-run the cost comparisons for 100 VMs using vCloud Suite 5.1, you will find that a VMware solution does not cost 19 percent more, but approximately 440 percent more than a Microsoft solution with Windows Server 2012 and System Center 2012.

This is a feel-good development for Microsoft, which just loves sticking it to VMware. From VMware's point of view, it's a thumb in the eye they could do without, but the train keeps on rollin'.

Posted by Bruce Hoard on 11/16/2012 at 12:48 PM


Subscribe on YouTube