HDX 3D - What Happened to Projects Pictor, Apollo and Prism? » ocb

Post to Twitter | Comments (6) | Views (2938) |

04 Sep 2009 10:06 AM EDT

HDX 3D - What Happened to Projects Pictor, Apollo and Prism?

[ Tags: xendesktop ]

I don't care what anyone says, it's really not that hard to virtualize a desktop. The hard part is delivering virtual apps and desktops to users in a way that's so fast... so smooth... so seamless, they don't even realize it's not running locally. Believe me, THAT is no easy task. Introduce a slight mouse delay or drop a few frames, and users go nuts.

Citrix has spent more than 20 years refining and improving user experience for virtual desktops and apps. During that time, we've faced every imaginable combination of users, devices, networks, drivers, peripherals and applications. And for each new challenge, we've gone back to the drawing board to develop, optimize and refine technology from the datacenter to the device to ensure a "high-definition experience," using the least bandwidth possible. When you do something for more than 20 years, you tend to become fanatically focused on the details. You also tend to get really good at it. That's exactly what the HDX Technology in our Citrix XenDesktop and XenApp product lines is all about.

But while I'll stand our HDX Technology up against anything in the industry, there is one class of applications that have admittedly been almost impossible to do well in a virtual desktop environment in the past. Apps with high-end 3D professional graphics such as Dassault Catia, Autodesk Inventor, ESRI ArcGIS or Siemens NX are so demanding, they have simply been impractical to do in most cases without running the app and all its data on a local PC or workstation.

This is unfortunate since the manufacturing companies who are the heaviest users of these apps are often those who have the most to gain from the flexibility, security and cost saving benefits of virtualizing these apps and running them centrally. Because many of their users are in offshore locations, intellectual property concerns are high. And since offshoring partners tend to be a long distance away and can only work with data that stays securely in the customer datacenter, WAN performance is absolutely critical. When it comes to engineers and designers working with complex 3D models, you simply can't afford to settle for anything less than a true high-definition experience.

Earlier this week, we officially unveiled our new HDX 3D technology, breaking through this barrier for the first time ever. Many of those familiar with Citrix, however, know that we've been working on the technology needed to deliver these high-end graphical apps (and the enormous data models they use) for a few years now. During that time, we've had several high profile projects underway to address various aspects of this problem, including Projects Pictor, Apollo and Prism... all of which you've probably heard about in blogs, events and user conferences. So which of these efforts lead to the final breakthrough that is now HDX 3D? Here's the inside scoop on how it played out.

From early on, there were a number of technologies on the market designed to "remote" 3D graphical apps. At the same time, it was pretty clear from the beginning that none of them worked in a true WAN environment. While this was obvious from our own testing, it was even more apparent by the number of customers who asked us for a solution even after purchasing other 3D graphical application remoting solutions (some quite expensive, I might add). Issues like lag, overshoot and mouse delay were driving users crazy as they constantly chased their models around the screen and waited for skittish refreshes. As I mentioned above, it's not hard to virtualize a desktop or an application - it's just hard to make it feel like it's local.
While figuring out how to perform seamlessly over a WAN was obviously critical, we also found out that unlimited bandwidth LAN environments could be just as much of a challenge as tightly constrained WAN scenarios. While virtualized 3D apps tend to perform better on the LAN, customers quickly discovered that they could easily consume up to 30 to 40 Mbps of bandwidth on average.

Clearly this too was a non-starter. In many ways it reminded me of when I bought my last car. After all the papers were signed, the salesperson walked me through all the buttons and dials, and casually slid into the conversation the fact that the car took only the most expensive grade of fuel. Not the kind of surprise you want to find out about after-the-fact.

Probably the best news of all from a customer perspective is that HDX 3D is a built-in feature of Citrix XenDesktop Advanced, Enterprise and Platinum editions at no additional charge. It runs on a dedicated blade or server with a dedicated GPU, but is managed alongside regular XenDesktop users - providing a single infrastructure that serves everyone.

It's pretty easy to sum it all up: HDX 3D for does the same thing for 3D graphical applications as the rest of our HDX technologies do for regular applications. At the end of the day, it all boils down to user experience and cost. You can talk all day about the flexibility and security benefits of centrally managed desktops, but if users have a bad experience... or if you blow out your network costs in the process, you haven't solved the problem. Luckily with HDX you have both. And with our new HDX 3D technology, we've finally broken this barrier for the most demanding professional graphics apps in the industry.

But now for the REAL question on the mind of all you Citrix fans who have been following us on this journey... "Which of those graphics acceleration projects we heard so much about became HDX 3D?" Was it Project Pictor? Or Project Apollo? Or maybe the more recent Project Prism?" Or were these all really just the same thing under different names to keep people guessing?

The answer is actually pretty simple. Project Pictor, our first major investment in this area, was based entirely on a XenApp architecture. Project Apollo, a parallel effort that started a bit later, set out to accomplish the same goals on a XenDesktop based architecture. As we worked through these two projects, it became clear that both had strengths and weaknesses. At the end of the day, we decided to combine the strengths of both projects under a new initiative dubbed Project Prism. It is this final combined effort that will debut later this month as the new HDX 3D feature of XenDesktop.

I have personally tested all of the other solutions in the market today and can assure you that XenDesktop with HDX 3D provides the best user experience over the WAN hands down... AND uses dramatically less bandwidth over the LAN. Getting there was no easy task, but I can promise you that taking the time to get it right was well worth the journey.

Labels

Comments (6)

Anonymous says:

Barry Phillips says:

Stephen Greenberg says:

Sridhar Mullapudi says:

Anonymous says:

Barry Phillips says: