When layers become a problem

I have seen a lot of discussion on the subject of layering of software images to deliver virtualized client computing. There are some interesting opportunities here but also a number of serious potential issues. This is too big an area for a single blog post so I am going to look at this over a number of post in the coming weeks.

There are number of different ways the word ‘layers’ is used in client computing. Firstly, when we think of our software stack we unsurprisingly think of the layers that make up that stack. Those layers are typically operating system, applications and user environment. We then think about how we manage those layers, either as the complete unwieldy stack or individually. As  such we are thinking about layers as a handy shortcut for what we want to manage individually.

Another way that the word ‘layers’ is being used is as a way of splitting the stack into a large number of separate layers. A number of small vendors are launching early stage products to try and deliver solutions based on a layer per application with those layers being delivered selectively depending on the user. The problem here is that each of the layers then needs to be managed separately including the impact of every combination of layers. Many of the organizations I talk to have thousands of applications in use across the business, that means thousands of layers and a huge combinatorial problem.

The change between the two uses of the word ‘layers’ is dramatic: In one case we are referring to three layers that we want to manage individually in the other case attempting to manage thousands of layers in a way that may well prove to be impossible. As I think of managing all those layers I am reminded of the mess that usually results from my attempts at eating a Napoleon, also known to the French as a ‘Mille Feuille’ – thousand layers.

Integration between internet or cloud delivered applications

Brian Madden has started a thread about one of the areas that I see as being absolutely critical to the delivery of client computing from the cloud:

 http://www.brianmadden.com/blogs/brianmadden/archive/2009/10/14/we-need-a-new-user-settings-framework-before-rich-internet-apps-will-replace-windows-apps.aspx?CommentPosted=true#commentmessage

As I see it, in all of the discussion about cloud computing there is one aspect of cloud delivered client computing that has got little public attention so far and that is the subject of integration between Internet or cloud delivered applications. To illustrate what I mean; imagine a user who has finished editing an image and wants to send it to colleagues. In a traditional desktop the image editing program knows where to look to find an email program that works on behalf of the user. If the applications are now provided by different suppliers then there is currently no way for one application to signal that it wants the help of another.

Being able to do this is taken for granted on traditional desktops because everything runs in one place and a mixture of the registry, inter process communication and configuration files take care of communications. In moving to the cloud one of the great benefits we get is the ability to buy services from different suppliers and to move freely between suppliers. However, this breaks the integration model. You could say that these applications would have to be provided by a single supplier, but that argument quickly leads to all applications being supplied by a single service provider. This is not cloud computing it is outsourcing, a valid model but not the same as cloud computing.

What is needed is needed is a mechanism that allows unified policy and personalization to be delivered between multiple cloud providers. Sounds a lot like User Environment Management, doesn’t it? One of my goals is to establish such a mechanism so that customers have a single place to control all of the interactions between their cloud delivered services. This is the only way organizations will have an efficient way to add additional services and the freedom to move between providers.

Windows 7 roll out will drive desktop virtualization adoption

Following a major gathering like VMworld it is little surprise that a number of people come away with a similar take on a key area. The adoption of desktop virtualization to support Windows 7 roll out is just such an area. I recorded a video interview on this last week and this week I have seen articles by Bernd Harzog and Jon Wallace both commenting on the same underlying point.

 http://www.virtualizationpractice.com/blog/?p=1316

http://www.insidetheregistry.com/content/viewarticle.aspx?articleid=1768

The roll out of Windows 7 is going to be a key to the adoption of desktop virtualization for a number of reasons:

• The fact that most organizations did not roll out Vista has given IT departments time to reflect on how they want to manage client computing. Organizations that previously went from one OS migration to the next want to get off the treadmill and get more control over their own destiny.  Desktop virtualization allows them to do this by componentizing the software components of a desktop and bringing them together for a user whenever needed. This removes the bulk of the issues of application compatibility that have made past migrations so difficult and expensive. The move to Windows 7 should be the last time organizations have to do a brute force migration.
• The economic situation over the last year or so (plus the skipped Vista migration) has left many organizations with a very elderly desktop estate that needs to be replaced. With so much of the estate overdue for replacement there is a great opportunity to introduce a fundamental change such as thin clients and host virtual desktops or client virtualization.
• The technology for desktop virtualization is ready now for some user groups and will fit for a far larger proportion of the user base at the point that Windows 7 will roll out. Technologies for delivering operating systems available from Citrix, VMware and others. Applications can be delivered in a host of ways to support different use cases. User environment management products such as AppSense’s deliver all user aspects of the desktop into these standardized environments and aid migration from XP/Vista to Windows 7. They will become an essential part of how you manage desktop virtualization both day to day for future migrations.

In short, several key events have happened at just about the same time and, on top of the roll out of Windows 7, will cause a rapid move to desktop virtualization.

Progress in application virtualization

It seems like an eternity ago now but back in December Virtual Strategy magazine asked me to make some predictions about what I expected to happen during 2009 in desktop virtualization. The article is here:

http://www.virtual-strategy.com/December-2008-Executive-Viewpoint/Executive-Viewpoint-Martin-Ingram-AppSense.html

I was reminded of this article while at Briforum a couple of weeks ago. My first prediction was that the application virtualization vendors would improve application compatibility so that a far greater proportion of applications could be successfully packaged. This is important because, for application virtualization to be useful in delivering componentized virtual desktops, organizations need to be able to deliver the majority of their applications in this way. In talking with the Symantec people at Briforum they now believe that almost all applications can be delivered with the Symantec Workspace Streaming (the new name for Altiris SVS). This is really encouraging and, if similar progress can be made by the other vendors in this space or through some of the emerging technologies in this space, will get us to our objective of being able to deliver a desktop composed of standardized components plus the user environment. This will give IT a more manageable platform while users continue to get a familiar and productive working environment.

Incidentally, looking at my third prediction I see organizations are making significant progress there too. But that will be the subject of another post.

AppSense at Virtualization East

A few months ago I presented at SysCon’s Virtualization Conference in New York. To anyone who has not been to one of these events I would thoroughly recommend them -good sessions and plenty of interesting conversations to be had in the corridors. My interview with Jeremy Geelan where we talk about desktop virtualization now and in the future has just been published. The video is available here:

http://tv.sys-con.com/node/1000899

If that interests you, I will presenting at the Santa Clara virtualization conference in November. Hope to see you there.

Gartner has been busy

In just the last two months Gartner have published three documents that talk about the importance of Persistent Personalization, their term for a subset of User Environment Management (UEM). They have approached it from three different directions: One document looks at Persistent Personalization itself, one for hosted virtual desktops and one for client virtual desktops. As you would expect with three documents in such a short period of time there are some very solid pieces of information and some long shots. In this article I want to pull together the principal themes and clarify some of the points they make.

Out with the old in with the new
Organizations have been implementing VDI since around 2005 but the early implementations were all reliant on keeping a complete copy of the user’s image for each user. This meant that the use cases where it made sense to deliver desktops in this way were economically limited. Today organizations are implementing VDI through componentizing the image and then automatically delivering those components on demand allowing IT to standardize the components and hence reduce costs and improve service delivery for the broader user base. One of the critical components is the user environment which represents all user related information in the image and which is delivered into the standard components to give the user a productive and familiar experience.
Up to now there has been some debate on what the critical dates for the adoption of this new componentized model and hence VDI in general would be. Gartner’s take is that VDI is ready for task workers now and will be ready for deployment to a broad range of users in 2010, driven by the actions of vendors such as AppSense introducing functionality such as user installable applications to allow greater user freedom. They also see that the componentized model will drive the older single-image-per-user deployments obsolete by 2012.

Persistent Personalization (UEM)
Unsurprisingly Gartner see Persistent Personalization as critical for the success of desktop virtualization with some stages of its development being contingent on developments in UEM. In particular Gartner cite the capability for a user to be able to install applications and I would also point to improvements in the manageability of the platform.
In debates on technology it is critical to keep in mind the overall objective and so not get diverted into unproductive side alleys. The objective of desktop virtualization is to improve the manageability of the desktop platform and hence deliver a better service to corporate users at lower cost. The way that we will achieve this is by standardizing the components of the image and managing these components across the business and so achieving economies of scale. Consequently, key to reaching our goals is that we can manage the components effectively. This means being able to manage the delivery of the components but also manage within the components so that we have appropriate visibility and control into the components. We must not slip back into the unique and impenetrable image problem we have in current PCs, generally referred to as ‘the blob problem’. This means that manageability of the user environment is key so delivery can be done efficiently and so that any user problems can be isolated quickly and effectively.

And a new category: Workspace Virtualization
Within Persistent Personalization, itself a subset of User Environment Management, Gartner have introduced a new category called Workspace Virtualization to cover vendors such as RingCube, MokaFive and UniDesk. Some of these vendors are completely new, some have had offerings in different markets before but this is their first recognition as part of broader corporate desktop virtualization. Consequently it is worth thinking about how their technology could contribute to the goals of desktop virtualization.
Amongst the vendors with products the common theme is that they split the image stack in a different way to typical desktop virtualization. Rather than creating a division between operating system, applications and user environment they split purely on the basis of operating system and everything else.
This has the benefit of simplicity but does it address the problem we need to address and hence achieve our goals? The solutions in this category are very new to market and will undoubtedly mature over time but I have two concerns. Firstly, can these solutions manage the components of desktop virtualization so that we achieve economies of scale or are we going to find ourselves with an unmanageable blob per user, much as we have now? Secondly, can these solutions become more than a point solution for just a small proportion of users leading to another management tool which must also be managed? It is too early to tell with the solutions we see now and it will be interesting to see if this sub category can carve a niche for itself.

 

In conclusion it is great to see Gartner getting behind the importance of Persistent Personalization and its importance to desktop virtualization. Three papers in two months is recognition that Persistent Personalization and the larger category of user environment management is critical to the development of desktop virtualization. This will help the broader market understand the role of UEM in improving the overall management of corporate desktops and hence deliver better service for users at lower cost.