Now that XenDesktop 4 is a few days old and people are starting to digest the many new capabilities of the product, I thought I’d spend a little time what exactly “FlexCast™ delivery technology” means.

I find that it helps me to think of FlexCast more as a strategy for delivering desktops, than as a specific technology. It’s about thinking of all your virtual desktop and application delivery methods as a toolbox that enable you to directly address the different performance, security, personalization and mobility requirements of all your users. This approach is dramatically different from the way most other vendors are looking at desktop virtualization. They try to force fit a single desktop virtualization approach to all use cases, even when it doesn’t make sense. FlexCast is revolutionary in comparison, but the idea is quite commonplace if you think of some analogies.

For example, instead of talking about how to run an IT service, let’s say I’m running a different kind of service – a restaurant. A hundred people come to my restaurant one night, and they walk in with the expectation that they can order what they like to suit their tastes and hunger. But instead of taking their orders, I don’t even offer a menu, and I serve everyone a plateful of chicken, green beans with almonds and potatoes. For the lighter eaters, this is way too much food, and for the folks that skipped lunch, they are woefully underfed. Then there are those that have strict dietary requirements, like a nut allergy, and they can’t eat any of it because almonds touched the plate! You might please a few patrons, but you’ve wasted your inventory by giving too much food to some, the wrong food to others, and not enough to the remainder. And at the end of the night, none of them want to come back ever again.
Of course, this would never happen in the real world, but for some reason this one-size-fits-all approach is how other technology vendors have treated desktop virtualization. XenDesktop 4 with FlexCast is your extensive menu of options for giving users what they want and what they need.

That said, I’ll give a quick drill down on each of the FlexCast delivery models and describe them each as plainly as I can. The names we chose to assign each model might be a little different than the current lingo – but this was necessary to be more precise, because no one has ever really talked about them all in the same context before:

  • Hosted shared desktops. If you are a XenApp customers today, you think of these as “published desktops,” and we estimate that you are using this delivery model for over 10 million users today. The problem with that name is that the term “published” can describe nearly every delivery model! Not very helpful. Hosted shared desktops are built on the Microsoft Terminal Services (or now Remote Desktop Services) platform, where users share effectively one configuration of a Windows Server desktops via independent sessions. In this model, there’s a lot of IT control over the configuration, and personalization is minimized or disallowed. This model is attractive not only for the standardization it enables, but also the maturity of the technology, its massive scalability and low TCO – up to 500 users could share a single server. These traits are what make it ideal for factory workers, retail clerks, bank tellers, nurses’ stations and the like.
  • Hosted VM-based desktops. You know these as VDI or hosted virtual desktops. Each user’s desktop runs in its own virtual machine, enabling multiple users to share a single physical server while running their environments in isolation from each other. This affords each user more potential personalization, a familiar Windows desktop environment, and compatibility with applications designed to run on a desktop OS. Scalability is good, with about 50 production desktops per server, and getting better all the time – but still not of the scale of hosted shared desktops.
  • Hosted Blade PC desktops. In reality, these could be blade or rack workstations, or simply PCs relocated in the datacenter. In this model, you have one user per hosted blade PC, so clearly massive scalability isn’t your goal. Going back to the restaurant analogy, this is what you pull out of the kitchen when the local football team’s offensive line comes in to eat . Workers with heavier computational requirements, like engineers, scientists, researchers, etc. would get these.
  • Local Streamed desktops. This model is truly one of the hidden gems of desktop virtualization, often overlooked but very useful and cost-effective. You may have heard this type of model referred to as “network boot ” or diskless PC. Let’s say you have an environment with lots of standardized PCs that you’ve just purchased in the past couple years. Perhaps they are attached to some specialized peripherals that are particularly “chatty” with the desktop OS. Desktop streaming enables you to leverage the CPU and RAM of that PC and give a truly local experience, but also centralize the management of those desktops. A “provisioning server” in the datacenter streams the OS bits needed to run the desktop to local memory. A single server has enough horsepower to serve a few hundred users, so the scalability is somewhere between hosted shared and hosted VM-based desktops. The re-use of existing PCs also contributes to the cost-effectiveness of this model. You need fairly standard PCs and a LAN connection, but this works great for all those users that work from the office primarily anyway.
  • Virtual apps to installed desktops. For current XenApp customers, this is what most of you are doing today. You are either hosting or streaming apps to rich clients with locally installed OSes. You get the benefits of reducing overall desktop management costs by simplifying application management – the more apps you virtualize, the lower your costs. This model is often the simplest, most “traditional” way to start with desktop virtualization. Virtual apps can be used both online and offline for mobile workers. Only issue is that you still have to deal with the OS at the endpoint, and this is something many IT shops are looking to desktop virtualization to solve.
  • Local VM-based desktops. This model is enabled by a client hypervisor, which, to be clear, isn’t quite here yet, although Citrix has publicly discussed and demonstrated XenClient. This model means that virtual machines live on the endpoint, and virtual desktops are delivered into those VMs. A robust solution incorporates encryption, security policies, and synchronization for OS, apps and user data with the data center. This client hypervisor would enable a centrally managed, virtual desktop to be taken offline, so it’s ideal for the fast-growing laptop using population. This isn’t part of XenDesktop 4 today, but it is definitely a significant part of our overall FlexCast delivery technology strategy.

So there you have it – a quick overview of FlexCast delivery technology in all its flavors. It’s a powerful concept, and a technologically involved one. The real takeaway is that, for Citrix, despite involving many technologies, it is indeed ONE comprehensive strategy, designed to meet many requirements.

Learn more about Citrix XenDesktop 4

Follow XenDesktop on http://twitter.com/xendesktop