Tuesday, June 09, 2009

Response to Steve Kaplan blog on Virtualization

On June 8, Steve Kaplan posted a blog analyzing Microsoft's Virtualization Calculator, which we at Alinean developed, vs. VMware's calculator (which we also developed). Both tools are powerful instruments, allowing organizations to explore various scenarios for virtualization and analyze savings, return on investment and payback.

The ROI / TCO tools can be found at:
http://www.microsoft.com/virtualization/why/roi/default.mspx
http://www.vmware.com/solutions/cost-savings/

The blog can be found at: http://roidude.typepad.com/my_weblog/2009/06/dont-believe-any-numbers-you-dont-make-up-yourself.html

This blog post is a response to Kaplan's post:

1) Personalization required - First, we could not concur more with Kaplan's assertion that the metrics you collect and enter into an ROI model are the most important. Whenever we develop an ROI model, certain industry metrics and assumptions are made. These are average metrics, and do not apply to every organization, however, most organization have difficulty collecting thier own metrics, and to provide initial estimates and a starting point, all our tools provide the ability to start with these metrics and common assumptions for a quick analysis. However, we ALWAYS provide the ability to scrutinize and edit all metrics and assumptions. Every metric highlighted in Kaplan's blog can be reviewed and edited - and moreover should be to create a completely personalized ROI analysis.

2) Head to head comparison methodology - Admittedly we never like doing head to head product comparisons. Why? Because prices change constantly, configurations are very sensitive to the environment and advances in technology, and it is always difficult to do a head to head comparison when solutions are different. So we concur with Kaplan that the comparison may not be apples to apples. We clearly state this in the comparison section, highlighting that

a) We assume that the benefits for VMware are the same by default, and that all pricing,
b) ALL configuration assumptions are from Microsoft directly and not Alinean.

As Kaplan points out, there is conflicting information from each vendor on which solution provides higher consolidation ratios, management benefits, integration, etc..... As part of our comparison we are clear to highlight that we do provide a price comparison, but we do not provide a benefit head to head comparison because of the conflicting information and no definitive research one way or the other.

HOWEVER, to allow users to dig further and customize the comparison, we provide for the capability to review and override the assumption that another virtualization solution and Microsoft provide identical benefits. Our suggestion - run the VMware TCO Calculator or Citrix XenServer Calculator and put in the benefit calculation if it is superior to Microsoft's consolidation / savings.

3) Training Classes - Kaplan assets that Alinean made an error regarding Training costs being included for VMware or other providers and not for Microsoft. The assertion Microsoft made in this section was that since Hyper-V is integrated in the operating system, and admins are already familiar with it, this integrated approach alleviates the need for training. I believe this is a sound assumption, although many organizations already do have a VMware skillset as Kaplan highlights. Don't like the assumption? If the user does not agree, it can be changed easily.

4) Are CALs required or not? Kaplan asserts that CALs are included for VMware and not for Microsoft. According to Microsoft, and as programmed into the model, CAL upgrades are not required for Microsoft users who are hosting existing Windows Server 2003 servers onto Windows Server 2008. This information can be found at: Volume Licensing Brief. As indicated prior, Microsoft was responsible for defining the comparison and configuration criteria (as clearly stated in the model we did not independently create this section), and specified removal of CALS, however, we agree with Kaplan and believe that the issue of CALs should be removed from both sides of the analysis equation and will be posting an update to proactively address.

5) Consolidation ratios can change the configuration - We agree with Kaplan that the consolidation ratio drives higher or lower differences in costs between Microsoft and the other provider's offerings. Using the default, can provide a good starting point and initial estimate, however, the consolidation ratio is easily changed, and needs to be personalized from our iniital estimates (based on utilization, as is server configurations and to be servers selected). Licenses are automatically matched based on needs, however, these too can be reviewed and easily customized to match actual pricing and configurations.

To help users with this consolidation sizing, the latest version of the Microsoft Integrated Virtualization ROI / TCO tool includes integration with Microsoft's MAP tool, which can help organizations more precisely analyze current servers and utilization, and assure that consolidation ratios / target server configurations are appropriate based on real world analysis.

6) Backup Software - counter to Kaplan's assertions, notes were provided in the tool regarding backup software and its seperate pricing for other virtualization solutions compared to the Microsoft configuration using System Center DPM (also at a cost).

7) Return on Investment calculation error - not sure what is being referred to, but there is no double counting of savings, or error in our own review.

Kaplan makes a number of good points in his article regarding personalization of any ROI report, and not just taking the vendors initial word for the results. We encourage and enable this more than any other tool provider, and welcome the continued debate in order to help users make the best, educated decision regarding virtualization.

No comments: