Citrix XenDesktop 5.5 has made some significant enhancements that will help many more customers adopt desktop virtualization as they strategically rethink if their client architecture capabilities will carry their business forward. As Citrix continues to make these enhancements, a greater diversity of client architectures will emerge as customers pick the right desktop for different use cases within their enterprise resulting in a greater need for user virtualization that abstracts and enables centralized management, and on-demand delivery of the user experience across multiple client architectures.
Let’s start with the continuing saga of the protocol arms race and dispel a myth right off the bat. Delivering user experience is not just about the protocol. For me it’s logic 101 to understand that delivering the best user experience in a world where there are multiple delivery scenarios will require numerous techniques. This is why Citrix uses the HDX brand to describe the delivery of user experience. While it is true that in many cases the delivery of user experience will require enhancements to the core delivery protocol, that is a very one dimensional view of the user experience.
Ok ok, the cynics out there are already starting to raise their eyebrows and refusing to let go of their never ending love with the “my protocol is better than yours” arms race :-) So let’s just get that out of the way. When I was Desktop CTO at Citrix, we started project Nitro which was designed to take many of the core protocol components to the next level. I’m really pleased that my former colleagues have delivered on many of those promises by providing solutions for the delivery of a rich experience for Flash, video, voice and graphics. I am sure the Citrix blogs will give a lot more detail about these, but it is significant to note IMO, that Citrix is and continues to invest in the core delivery experience. Combine this with the Citrix Receiver that is now available on more platforms than ever, and the value of centrally hosted desktops can be extended to many more consumers devices while enabling IT to maintain security centrally. For many, multimedia use cases have held back the adoption of desktop virtualization and XenDesktop 5.5 represents a significant step forward to removing those barriers.
HDX driving adoption with the sponge cake.
I hope you enjoyed your protocol fix. Now we can move on to discuss the other significant HDX update. The addition of personalization via the acquisition of RingCube as an early access release. I think this is a great move forward and much like the HDX multimedia improvements, HDX personalization will result in greater adoption of VDI for certain desktop use cases as customers continue to diversify their client architecture strategy. In particular I am really pleased that Citrix is making this available for all releases. Why do I think this is really important? It’s a lot more than arguing I don’t want to be forced to upgrade from one edition of XenDesktop to another. It’s more about adoption across our industry.
Desktop virtualization has been around for several years now, and Windows 7 migrations are going to have to be fast tracked with Windows XP end of life being April 2014. This is a long awaited opportunity for many to finally get some investment in the desktop, however we are at a critical stage in the maturity and lifecycle of desktop virtualization. Clearly adoption is critical as we move forward and without broad successful adoption by the mainstream (beyond the early pioneers), our industry will suffer as customers lose confidence in their ability to execute on a strategic client re-architecture. Therefore, steps any desktop virtualization vendors takes to drive adoption are critical to the success of our industry which I embrace with enthusiasm.
As with any new technology architecture introduction, there is always the question as to what the impact will be on existing technologies. It’s certainly a question I entertained while at Citrix when we were considering acquiring RingCube in early 2011 and again when I decided to join AppSense as CTO. It helps to first understand what RingCube is at it’s most fundamental level. It’s a difference disk that lets you separate your stuff from the base build. It may be tempting to confuse this with VMware linked clones technology but RingCube IMO is simpler by design and does not rely on a model of creating dependent replica VMs linked to a master VM, with per user data/persistent disks and apps delivered via application virtualization. For me this model makes sense if you are looking to replace roaming profiles, redirect My Documents data to a separate user data disk, are willing to bet on application virtualization across the board and are happy with View Composer scalability.
The RingCube technology does not require you to change your existing management tools and hence I think it is simpler to adopt for mainstream customers.I am sure many could debate many virtues of either approach but that is not the point of this blog. The salient point I want to make is that they are different approaches and could be confused if you are not careful which does not help drive adoption as confused customers only evaluate indefinitely…… I used to describe the RingCube approach as the 2 layer sponge cake – a pretty simple mental model to understand. Think of the lower layer of the sponge cake as your C: drive, the upper layer the D: drive. The C: drive is managed centrally and D: can be used to personalize your experience. This provides lots of benefits.
Firstly if the C: drive is managed centrally then it becomes possible to share the C: drive amongst many users and reduce the cost of VDI storage moving towards the vision of pooled desktops. Patching becomes easier as it’s a single point of central update as opposed to trying to manage multiple distributed C: drives that overwhelm so many desktop organizations today. In fact I believe that many 1-1 implementations that leverage local storage can also take advantage of this approach. Since your existing management tools and processes can still be used, the adoption is less disruptive making it easier to get started for certain use cases. So this technology has the potential to drive a lot more IT efficiency which is goodness. Once the C: drive is patched, the personalization changes stored in the D: drive can be reapplied bringing back a consistent user experience. Since the D: drive RingCube difference disk does not require any infrastructure, scalability is not a concern, it’s as scalable as the core XenDesktop infrastructure. I know first hand that this was always a big architectural concern that could cripple the viability of a solution for many customers, hence the simpler approach is prudent IMO, especially when first starting out.
Not all agree and for an alternate opinion, Ron Oglesby from Unidesk provides for an entertaining read. I could go on to argue the merits and likelihood of success of a simple difference disk solution vs. a complete rethink of desktop management using multiple layers but despite what Ron thinks my position is, I am somewhat neutral since driving adoption is what I am most focused on.
Will the apps work?
What’s more important with respect to disk layering technologies is customer success with application compatibility. There’s always been the question in our industry of will apps work in a layer? Will third party vendors support their application running in a layer? That has always been a barrier to adoption for any of the small innovative companies to truly get traction. So with an industry giant now making a clear statement that layers are ok, this will go a long way to legitimizing layers as an approach for certain use cases and this will further help drive adoption and spawn further innovation. This is great news for AppSense, as this reaffirms our belief that a complete user virtualization platform should include this type of component technology to enable some use cases. This is something that we have had in development for several years and at AppSense we continue to evaluate technology approaches to solve for additional use cases that enable organizations to empower their users in new ways. Watch this space!
So is personalization a silver bullet?
To answer that question it’s first important to understand things at a fundamental level. We’ve already established that RingCube will add personalization with the addition of a D: drive or vDisk as Citrix calls it. However we need to think about that a little and understand in reality what that means. If you keep adding stuff to your D: drive guess what happens? You end up with a really big D: drive that erodes your benefits. It’s like having a really big fat profile. So it’s obvious that during real life implementation you need to think about how to leverage this technology to drive your solution(s). Depending on your environment you can of course say personalization to you means collapsing all your non core image apps, data, preferences into a single layer and have a simple architecture that scales. I have no issue with that, but I believe over time as things change you will need more management and greater abstraction and granular control. After all we are trying to move away from a world where the desktop was managed as a single monolithic entity.
You can achieve some of that additional abstraction via a profile management solution like Citrix UPM which is part of XenDesktop. Citrix understands that and hence why RingCube vDisk technology is not being positioned as a replacement. However that will only take you so far because there is so much more to consider when managing across multiple client architectures, more on that below. In reality what I think most people will do with the technology with it’s current set of capabilities is use it for Departmental Installed Apps. Brian Madden discussed this quite well here.
There are other use cases as well but simple personalization and departmental installed apps IMHO will be most pragmatic. In summary I believe the D: drive will be used to group together sets of applications for departments to help manage against bloat of the D: drive eroding TCO benefits. This also helps mitigate against application compatibility concerns because one would assume the layered apps are tested. This would still be a simple architecture that scales using your existing management infrastructure and over time more sophisticated layers management could be added and even additional nested layers. Additional abstraction to provide further flexibility and use cases such as working in non VDI use cases like XenApp will have to be provided by other means.
So when do I need user virtualization?
A lot of people confuse user virtualization and think it’s just a profile feature, or it’s just personalization and perhaps if you have read this far you also now know that it will include additional component technologies like layers. User Virtualization focuses on the “complete abstraction and management of the user” away from the underlying desktop and applications. It’s about delivering the user experience everywhere. By the nature of our virtualization, abstraction and management, the right tool can be used for the right component. Many XenDesktop customers use this level of component abstraction to enable many new use cases to simplify the adoption of multiple client architecture such as PC, VDI, XenApp, App virtualization and laptops.
Without a sophisticated set of capabilities that enable IT to manage across these heterogeneous client architectures, one ends up with several methods to replicate functionality in multiple environments that leads to complexity. Disk differencing technology does not work across all these client architectures and also does not provide the same capabilities as user virtualization. Some of these additional user virtualization capabilities can be summarized as follows:
- Robust granular set of and application of contextual policy controls and GPO.
- Per user, per application management of the users personalization.
- Application control to enable granular execution context.
- Application licensing and metering.
- Application and user rights management to mitigate data center risk due to user actions.
- Performance management.
- Cross platform support that enables roaming between locally installed, virtualized, and published applications.
- Cross application delivery support.
- Superior controlled model for roaming between VDI and native PCs.
- Proven scalability of solution in the most demanding enterprises.
- Vendor independent approach
For over a decade customers have understood the value of user virtualization and the need will continue to grow as Windows client architecture diversity continues. Reassuringly the technical community is beginning to see clarity about the distinct differences and need for user virtualization. Even Microsoft with Windows 8 is upping their focus on enabling a user centric world that enables people with the creation of a User-Centered Experience team.
It won’t stop there. The post PC Era is upon us. Today we are faced with increasing complexity due to the need for greater diversity of client architectures within Windows. User virtualization will play a major part to enable that world. However IT will need to transcend that world and learn how to embrace an emerging world that through the forces of consumerization and IT as a service, will enable many more client use cases. User virtualization will need to rise to the challenge and enable us to re-imagine a whole new generation of innovation to empower users and IT that goes far beyond disk differencing which is only applicable to Windows. Competitive, pretenders that deny user virtualization and limit their vision to only the PC Era coupled with arcane patent leverage to force marketing agreements under duress will continue to lag years behind in innovation and continue to operate from a position of desperation and weakness.
AppSense achieved record growth of 54% and revenue or $71m in FY 11. Yes we published the numbers because they are real and not rah rah statements like some….
…This record growth was not driven by the inclusion or exclusion of layers. Customers will continue to need technologies that allows them to abstract complexity, manage centrally and deliver on demand user experiences for the right use case. The addition of XenDesktop 5.5 with HDX improvements for both multimedia and the addition of a disk layer technology will drive greater adoption and is a welcome evolution in our industry. AppSense will continue to remain a strong Citrix partner driving adoption of XenDesktop with user virtualization and is committed to evolving the user virtualization platform to enable its customers to embrace the post PC Era.
Harry Labana, VP and CTO