VMWorld's third day came complete with discussions of the pros and cons of VDI (Virtual Desktop Infrastructure), Virtual Appliances, and the biggest party I've ever seen at the Las Vegas Speedway.

VDI is this year's hot topic at VMWorld but I suspect that it's embers will die before VMWorld 2009. No matter how you slice it, VDI is still a heavy technology using bloated, full-blown operating systems, network bandwidth hungry protocols, and performance that is lackluster. VDI is just a way to sell more server hardware and to lock in companies to Windows operating systems. If you opt for VDI, in the traditional sense, a thin client (hardware), a Virtual Host machine, and a Virtual Guest OS (The Desktop Install); you may find that you really aren't saving much money, if any, at all and you're binding yourself for 5 or more years to Microsoft.
To find out the long-term solution for this problem, go to my Linux Magazine Virtualization Blog/Column.

Virtual Appliances are another hot topic. Everyone is using them or building them or both. Using SuSE Studio and VMware Studio, you create your own Virtual Appliances and upload them to their respective community sites for all to share.
Several companies offer a physical hardware appliance and a virtual appliance with the same functionality as the physical device. There are appliances for firewalls, intrusion detection, intrusion prevention, virtual machine management, web-based applications, and so on. To see some of the possibilities, go to and download all you want. Be sure to download the so that you can test these little gems.

Finally, the VMWorld Party at Las Vegas Speedway. Imagine, if you can, 10,000 IT people, racetrack staff, and performers all running loose at this full-sized, world class racetrack with only one thing in mind: Have fun.
The entire place was decorated specifically for this party. It was all about VMWorld 2008. Free food--good stuff too; shrimp, imported olives, pate', BBQ, Mexican food, Pizza, and even Mango Gazpacho. Oh, and did I mention the free beer, pop, and liquor? Oh yeah, it's all good. Your old Linux guy stuck with Diet Coke--I had work to do--ah the life of an inefficient and hungry journalist.

But wait, there's more.

They were giving race car rides on the track to anyone who wanted to ride with speeds topping 120MPH. You could have a neck-snapping, tire burning one-on-one ride in a Mustang GT on the inside track or in a more family type sedan around the big track. They had professional drivers do the driving but still--how awesome is that?
I have never seen anything so over the top all out crazy in my life. Pictures and film don't do it justice but if you search in a day or two for VMWorld 2008 Party, you'll probably see what I mean.

VMWorld 2009: San Francisco -- makin' them reservations now.

I have just one word for my experience at VMWorld 2008: Incredible.

Dani AI

Generated

A concise technical follow-up to ’s Day Three recap: the post captures the 2008 VMworld tone—VDI skepticism, appliance momentum, and a memorable racetrack spectacle. The technology points raised are still useful as decision criteria rather than absolutes; this note focuses on practical evaluation and a short pilot workflow that answers the common performance, cost, and manageability questions in the thread.

VDI is a tool with clear tradeoffs. It can simplify endpoint management and centralize policy, but it shifts costs into compute, storage I/O, network, and licensing. Match delivery model to user class: non‑persistent pools work well for task/kiosk users; persistent or GPU‑accelerated hosts are needed for designers and multimedia users. Lock‑in concerns are valid—verify migration paths, image exportability, and licensing terms before committing to a single vendor.

A compact pilot checklist that addresses the main risks:

  • Segment users by workload (task, knowledge, power, kiosk).
  • Baseline existing desktops for CPU, memory, app launch time, and disk I/O.
  • Run a representative pilot (dozens of users across segments).
  • Test both persistent and non‑persistent images and measure login/boot storms.
  • Capture storage latency and I/O patterns under load; test caching/SSD options.
  • Validate peripherals, printing, multimedia, and remote (WAN) latency scenarios.
  • Model 3–5 year TCO including servers, storage, licensing, helpdesk, and ops.

Virtual appliances deserve separate caution: they speed deployment but carry patching, support, and configuration visibility risks. Treat appliance images as versioned artifacts, scan for vulnerabilities, and test upgrades in an isolated lab before production.

The racetrack party is a good reminder that conferences amplify vendor messages. Technical choices should be driven by measured pilot data and clear exit strategies, not by the size of the party. ’s enthusiasm about the report fits that balance between fun and practical takeaway.

Great report, Ken. Wish I was there with you!

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.