Does startup Liquidware Labs hold the secret to VDI success?

What happens when one of the founders of Foedus (the professional services firm VMware bought in 2008), the founder of Vizioncore (the software vendor Quest bought in 2007), and a software company called vmSight get together? You get a new company called "Liquidware Labs," a software company hoping to take the guesswork out of implementing VDI solutions.

What happens when one of the founders of Foedus (the professional services firm VMware bought in 2008), the founder of Vizioncore (the software vendor Quest bought in 2007), and a software company called vmSight get together? You get a new company called “Liquidware Labs,” a software company hoping to take the guesswork out of implementing VDI solutions. Their first two products were released yesterday, and they look pretty badass.

The ten-second overview of Liquidware Labs is this: their software analyzes your pre-VDI physical desktop environment to figure out what users are actually doing, then it gives you metrics and scores to help you decide which users you can convert to VDI, what apps and peripherals you'll have to support, what remote display protocols you'll need, how you should size your servers, and how many different disk images you'll need.

After all, if you don’t know what your users REALLY do, it’s impossible to guess how many servers you’ll need to host your user VMs. Do your users really use Flash websites? Do they watch video? How much time do they spend in each application? How much is active versus passive?

Liquidware differs from traditional monitoring tools like Lakeside Systrack or Citrix EdgeSight because Liquidware takes the data and turns it into specific scores and reports that show the VDI and user experience "fitness" level of each physical client. (UPDATE: So it turns out that Systrack does have a VDI assessment product. I'm obviously not familiar with it, but I've scheduled a demo so I hope to share more in a few weeks. Stay tuned.)

How Liquidware Labs came to be

As I wrote in the intro, Liquidware Labs is based on people coming together from Foedus, Vizioncore, and vmSight. Foedus was a consulting company that did a lot of early VDI work. After selling to VMware and putting in his requisite time, Foedus partner J. Tyler Rohrer was interested in trying to build software to model the best practices experience they developed doing VDI implementations. He got together with David Bieneman, founder of Vizioncore software who also found himself with some time on his hands after he sold Vizioncore to Quest. The two of them talked about what they wanted to do and decided to buy vmSight, a new (yet already in existence) software company that focused on VDI compliance and audit software. That brought Steve Gant and Jonathan Alexander to the team.

And so it was. This core group built-out the rest of their team, their plans, and got to work re-tooling the vmSight stuff to build software to put their VDI best practices in a box.

Product plans

Ultimately Liquidware Labs plans to have five products, although only the first two are available today. From a functionality standpoint, their products will include:

  • Assessment module (available now) Gathers configuration details of physical desktops and measures actual workloads, establishing the baseline for the VDI environment. Creates VDI "fitness" reports and identifies clusters of desktops, resource requirements, etc. (more on this later)
  • Diagnostics module (available now) Builds upon assessment data to collect detailed usage information about apps, networks, storage, etc.
  • Capacity planning module (not yet available) This is the “what if” engine... What if I decided to go win7, or switched hardware, or let people use hulu? You could even stand up and pilot and then model it to scale.
  • Migration module (not yet available) This is a component that can automate some of the actual migration work, like maybe tying into app compatibility lists and packagers, making sure that the right people have the right apps available (based on the assessment data)
  • Support Center module (not yet available) This is a tool for Level 1 and Level 2 support personnel, with potential for user self-service. It will hook into the VDI environment and pull data from the other modules.

How it works

Liquidware’s goal is to remote the question marks people have about going to VDI. They want to answer the basic questions like “what do you have?” and “what do you use?"

The first thing they do is interrogate the machine via the Assessment Module. “Tell me about you... your CPU, your drives, your network drives (since they need to know about I/O patterns), your printing habits, network connectivity, devices, peripherals, USB use, phones, iPods, screens and resolutions... everything” Then they watch the machine to find out what it’s actually doing via the Diagnostics module. They look for logons and logoffs, application use, how intensive your graphics operations are, etc. They also have some pretty cool patents (via vmSight) for injecting hidden signatures into network packets going in and out of a desktop, so they can measure round trip latencies and things like that.

Liquidware came up with their own set of metrics for this which is sort of a “fitness” level for VDI. Ultimately they want to classify the various desktops and plot them onto a scattergram with red/yellow/green colored dots to show how well they could fit with VDI. A green dot would be a good fit. Red might have some financial or technical implications that would cause you to want to think twice before virtualizing, and yellow might be something that could go either way, like they saw AutoCad on the workstation but the user never used it.

Ultimately this interrogation and profiling helps Liquidware find the real-world use cases that you would otherwise just not know about. I mean even a “simple” task worker might be using 500MB spreadsheets.
This “VDI fitness” is an actual metric they calculate into a numeric value based on all the stuff previously mentioned. They also calculate a numeric value for what they call the “VDI User Experience” that’s based on things like logon duration, CPU queue, page faults, network latency, graphics intensity, non-responding app times, etc.

And of course, customers can change the fine-tune the algorithms used to calculate both of these indexes, so, for example, you can increase the importance of CPU load if you just wanted to focus on user density and didn’t care about user experience. ;)

While they’re calculating these two indexes, Liquidware also lets you take a look at commonalities. Regardless of the VDI technology you’re going to use—linked clones or flex clones or Provisioning Server or app virtualization or whatever—at some point you need to start figuring out how you’re going to slice and dice your images and applications. So Liquidware lets you do that with the data you’ve collected, allowing you, for example, to only show data from Windows XP workstations with SP3 and at least 512MB of memory. And then you can keep on drilling down into the bubbles that can apply to the most narrow scope of multiple machines.

Again, all of this is so you can start your VDI project with real data and not just guessing. They have literally thousands of different metrics you can slice and dice, so now you can know (again, for example) which users can use RDP or ICA or who needs PC-over-IP. (And when you decide to use RDP instead of PC-over-IP, you can know exactly who will be mad and why.) They can also look at the SAN for things like IOPS per LUN and stuff.

Show us the money

So it all sounds good. I haven't personally been able to test it since I don't have an environment full of a bunch of desktop computers. (Actually, that's not true. I now work out of TechTarget's San Francisco office which has about 50 users. I wonder if they would mind if I popped the Assessment Tool onto our network?)

The Liquidware Labs version of the Assessment and Diagnostics modules just came out yesterday, so it's too early to get broad opinions (although partners have had access for awhile, and vmSight's tools were out for awhile too). That said, does anyone have any hands-on experience they can share? Personally I think this is one of the cooler products I've seen in awhile, and certainly one of the Top 5 companies I'm most excited about right now.

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

I was fortunate to see the details of this technology at Synergy in a 2 hour sit down with Tyler. It is truly incredible. No more "best guess" approaches or the use of old, obsolete or inherently inaccurate data. These tools provide the ability to segment your users with real live information . Makes the all to common guessing game approach obsolete. The ability to build a ROI analysis and VDI justification / business case just got very real.


From what I have seen of some of the detail - I would also agree - in fact it also looks like a great tool to do an even better version of an Infrastructure Assesment prior to implmenting a  XenApp Farm?


So I guess the questions that come to mind are: How much will this cost, how complex is it to implement and how long does it take to gather the metrics?  As a consultant does the VAR pay for this and provide it to the customer at no charge to generate the business or is this something that would have a use beyond the initial scaling/sizing window?  Just the type of things I know will be on customer minds.  Not doubting the value of it up front.


I have been using Stratusphere for about 2 weeks in a 3000 seat customer environment here in the North East [full disclosure, I am a consultant), and I have to be honest, it is almost surreal how much they SEE and index in their database. VDI is a complex enough solution, with a ton of question marks, so being able to look at application activity, user activity, os activity, network, Badass for sure. This product has some serious potential.


@glenda Canfield - All awesome quesitons:

1. OVF is about 400mb=10 minutes

2. Install and Config=      30min-hour (based on end user infrastructure)

3. ConnectorID deployment= based on eSoftware distribution +- 1-30 min (assumes 10-1000 end points)

4. Data Collection = instant, all metrics begin to write to the DB

5. Relevance of data can begin immediately, day one, day 7, or day 30, or ONGOING depending on what you seek

6. Two pricing models: One VAR's can use for services as trusted advisor, yearly fee, no limits on use....Two: End users may purchase from vars (in increments as small as 100 users) for use in their internal environments for ongoing system intelligence and diagnostics!

Hope that helps



@kevin goodman - agreed !  I keep missing you at all the industry events -something magic about BriForum - hoping to connect soon!


Congrats Kevin !!