VDI and ROI - Mutually exclusive?

I got an interesting post from Jimmy the Mad Salesman the other day:

'For the last 2 days, I have been looking in trade journals, google, and tech blogs for a strong ROI to move to desktop virtualization. Whenever I read an example, the ROI is always vague. So far the only example I have seen is a deal HP and VMware did with Collier County, FL but again only soft ROI are revealed. so far I see it makes sense from a management and security (perspective).'

ROI is not that hard. I think of it like this for VDI:

1. What do I save in licensing costs by virtualizing?
2. What do I save by not having to license certain applications (OS, A/V, etc)?
3. How many support calls were logged in the past year?
4. What were the support calls related to - apps, os, SaaS, new users?

On the flip side if a company virtualizes Windows, that will help w/support and management of a machine, however little if anything is saved.

If a company goes down the VDI path with Ubuntu (for example) all of the agents required to protect Windows from the several hundred thousand threats that exist go away, Citrix goes away, RAM to support Vista evaporates, and the list goes on.

In my opinion this means that if you virtualize Windows, you gain a measurable ROI, but centainly far from breathtaking.

Burton Group Dazzles

I am 2/3 of my way through the day at Catalyst and have been truly impressed by the quality of material I have had exposure too. I spent the day in the Data Center/Virtualization track and it did not disappoint. Notable take aways:

There is a right way and a not-so-right way to roll out true HA in a virtual way. The right way is simpler

'When we talk about security around a virtualized environment we get dumb again

I asked the question to one of the co-founders of Xen (who was speaking about VDI) how do you avoid the (insert software vendor here) tax? I mentioned microsoft btw. I did not get a straight answer which means one of two things - that's not the point (Citrix and Microsoft are two peas in a pod on this), or we haven't figured it out yet.

I was happy to get a premise validated in a left handed way which was that we are on the right track here at virtualizationstuff in rolling out an Unbuntu based desktop running an xTP protocol that will cost an organization $50-100/year per device running the solution. Including email using whatever client you want. The future looks bright

Catalyst is officially underway

I was at the opening reception last night decked out in my Scottish regalia to celebrate the kickoff of Burton Group's Catalyst conference. Some notable observations thus far:

My identitystuff crew is here
Ian Glazer is the newest member of the Burton Group
The folks from Boeing were dazzled by my kilt
National City Bank's architect flagged me down and promised to wear his next year

This remains one of the best conferences in the IT industry in my opinion - definitely worth checking out.

More on ROI at the desktop

I was speaking with a buddy at a Fortune 10 and they are going through the usual hand wringing about virtualizing desktops and asked me to help on coming up with defensible models for their finance team. Here were some of the things we looked at and I wanted to share them...

Agents - there are 12 agents on every desktop which means 12 additional pieces of technology and associated infrastructure components to purchase, deploy, and manage. Average savings was $580 per desktop on agents alone year one/per new build.

Patching - with our process the patching occurs on boot in as close to real time as possible. An average of 15 minutes per desktop/laptop * (enter device count here)/4 = hours per device per month.

Backup - Increasing storage is spend, but it is more than offset by forensic investigations ($5,000 per request/investigation, plus retainer) and calls to support. Google or Microsoft search is free so when the device is reset on boot up the search is available and the cost is then borne by the user, reducing calls to support 30%

Hardware - Life expectancy for a thin client linux on a non-beefy machine doubles its useful life and allows for further depreciation and flexibility when accounting for costs of assets. Ace up the sleeve for finance.

All in average savings was $1100 per desktop with better support, easier and more efficient management, and more control.

What's the Point?

I have been mulling this quetion over in my head after finally being able to catch up on some reading of the industry rags out there -

What is the point of virtualizing the desktop if you stick with the same OS and Approach to management?

Is it Microsoft's virtual license fees? No. You still need to license the OS.
Is it security? The Microsoft Tuesday updates are more secure than last month?

I don't get it. In the articles I have read, they all talk about saving $X after they have spent $Y on infrastructure, LICENSING FEES, etc. and it makes no sense if what you are trying to do is to cut costs.

The smart CIO's will go another path and deploy a desktop that is more secure, more controlable, consistent from dev to production, includes the back up, key apps, etc. required for users to do their job, and do it either on their concrete or through a managed service.

Why would I want to pay licensing fees for OS's that will require new gear and skillsets to manage in a way that is no better than I have today.

What's the point?