Disclaimer:  I am not a vendor certification kind of guy.  Until recently, I have never taken a vendor certification exam, never wanted to, never thought I would.  I’m still not a cert kind of guy, but at least I now have a little more respect for those who are.

The Vendor Creates an Exam 

Enter the 3rd Party Exam Development Company 

Enter the Exam Question Writer

Enter the Subject Matter Experts

Enter the Exam Training Class

There was also a training class official curriculm developed by the vendor (likely also outsourced) I was not involved in the development of the official curriculum training class for this exam.  I believe that was a separate process completed independently and in parallel using the same detailed version of the sylabus.  In our case I think the class had been written prior to our test scrubbing, however none of the SMEs had seen the class materials.

Enter the Exam Taker 

How I Would Prepare in the Future

Read More on Tim Mangan

The IPD represents the most critical design elements in a App-V design:

• Step 1: Determine the Project Scope
• Step 2: Determine Which Model(s) Will Be Needed
• Step 3: Determine How Many Instances Will Be Needed for Each Model
• Step 4: Client and Sequencer Considerations
• Step 5: Design the Streaming Infrastructure
• Step 6: Design the Full Infrastructure

This guide addresses considerations that are related to planning and designing the necessary components for a successful App-V infrastructure:

• Planning for streaming applications or deploying via software management systems
• Planning for centralized, decentralized, or departmental use
• Planning for isolated use, such as a lab or classroom
• Planning for Internet-based clients
• Rapidly deploying a new version of an application

The IPD guide can be downloaded here

Read More on Ruben Spruijt

For an organization that uses System Center Operations Manager already, this free management pack
from Microsoft can be a welcome addition that allows some limited monitoring capabilities. It is most
appropriate in larger organizations where the division of responsibility leads to a list of servers and
services that need to get checked off of a “to do” list by personnel who do not really even need to have
a cursory knowledge what the server or service does. In that regard, additional configuration of the
pack, through disabling of monitors or setting overrides, will need to be performed by someone more
familiar with App-V to eliminate unwanted alerts and state display warnings.

The complete whitepaper can be downloaded here

Read More on Ruben Spruijt

My friend Kevin Kaminski and I have been getting our updated "Masters" Level training class ready; putting the new App-v release candidate through its paces, checking out the new SCCM R2 integration, Ops Manager integration, stand-alone client, and even the "lightweight" server.  A lot of new stuff, for sure.

Anyway, if you missed your chance to see the new release at the Briforum session I did in June, I have a new video posted to my site today, at www.tmurgent.com

In Episode 8, we use the App-V 4.5 RC1 sequencer to sequence OpenOffice 3.0 Beta 2.  As part of the sequence, we create the MSI, and then use it to deploy to a "stand-alone" 4.5 client.  We hit a snag, and debug the sequence to solve.    36 minutes in length.

PS:  If in the Boston area, catch me speaking at the Boston Area Windows Server Users Group "Virtualization Deep-Dive Day" on August 15th.  http://www.windowsboston.com/ 

Read More on Tim Mangan

The “Application and Desktop Delivery solutions” diagram has been developed in order to be able to provide a complete overview of the various applications and desktop delivery solutions. This article was written by Ruben Spruijt in order to introduce the highlights of the delivery solutions in 30 minutes.  There are so many delivery solutions that the functionalities can be confused through incomplete knowledge. The point of this article is not to describe all of the application scenarios or the technical advantages or disadvantages, but purely as a high level, vendor dependent overview of the start of technology in the applications and desktop delivery segment. Hopefully this overview will be helpful!

Workplace scenarios

Trusted and Untrusted workplace scenarios . Trusted workplaces are devices that have a network connection to existing IT backed infrastructure via the LAN  or WAN.

Untrusted workplaces are devices that have no secure LAN or WAN connection with the existing IT backed infrastructure. Examples are devices that are active at home, at a stage work station or in connection with security in a separate network segment.

Each organization has various work station and application delivery scenarios.  It is important for the IT department to have insight into the different workstation and delivery scenarios.

This reflects how the users are working with or would want to work with the applications.

Secure Access

Secure Access solutions assure secure access of untrusted devices to corporate IT. The symbol consists of two parts, the shield stands for secure and the stoplight stands for access. The access can also be close linked depending on the chosen secure access solution. Solutions that realize secure access scenarios are, for example, Cisco ASA, Citrix Access Gateway and Juniper SSL VPN.

Web Application Acceleration

Web Application Acceleration solutions assure acceleration and security of web based applications.  Today we all make use of these solutions.  The largest number of the internet applications that we all use, such as Google, MSN, eBay or marketplace, make use of these applications.  Web application acceleration solutions are not only for the large internet organizations, but also for your web applications. Solutions that make web application acceleration and security possible are, for example, Citrix Netscaler and F5.

Desktop broker

The desktop of connection broker determines which remote desktop will be made available to the client. With this it is possible to make available a dedicated or a pool of remote desktops. The automatic turn on, deletion or pausing of remote desktops is a functionality that can be provided by a desktop broker. There are various suppliers of connection brokers. Citrix with XenDesktop, Provision Networks VAS and VMware with VDM are the most well known solutions.  Depending upon the supplier, the connection broker can have additional functions. Functionality such as a web interface that assures secure (SSL) and easy access to the remote desktops, Active Directory integration, USB port redirection and integration with Terminal Services in order to provide access to a Terminal Server or a personal Remote Desktop through rules set by IT.

Application Streaming and Virtualization

With the aid of application streaming and virtualization, windows applications can be used without any changes to the local operating system, let alone that application software is installed on a workstation.  In other words: the application is implemented, saves data and prints as if it is locally present, without anything being changed on the local client.  Sources such as CPU, memory, hard disks and network cards are used for the execution  of this application. Application Streaming and Virtualization assure the availability of applications on desktops, laptops, VDI and Server Based Computing platforms whereby the application is executed on the “client” platform.  No changes are made to the platform.

A number of advantages for Application Virtualization are:  installation, upgrade, roll back and the ease of application support.  Installations of applications is now in the past; conflicts are not longer possible.  It creates a dynamic application delivery infrastructure.

Solutions for Application Streaming and Virtualization are: Microsoft Application Virtualization (App-V), Altiris SVS, VMware ThinApp, InstallFree and Citrix XenApp client side virtualization.

OS Streaming

OS streaming makes it possible that VDI, SBC and desktops start up and work from an image file saved on the network.  A single image can be used by multiple workstations simultaneously.  The advantage is that complete operating systems, including applications and clients can be made available quickly and securely.  The availability of a single image on multiple VDI, SBC and desktops is possible without conflict. Through this, an upgrade or roll back of an OS is possible quickly, easily and without great risks. When virtual desktops make use of OS streaming in a VDI environment, this solution also saves valuable storage and the administration of the virtual desktops is simplified.  Virtual or physical machines that make us of OS streaming thus become “stateless devices”. Citrix Provisioning Server is a solution that makes OS streaming possible.

Virtual Desktop Infrastructure

VDI, Virtual Desktop infrastructure = “Dedicated Virtual Remote Desktop”

Virtual Desktop Infrastructure (VDI) is a solution for remote access to Windows XP/Vista or Linux desktops that are implemented on a virtual machine in the data center.  VDI can be a server hosted solution (online computing) or a client side solution (offline computing). This overview describes VDI from the server hosted solution. With this, access to the desktop is not bound to one location or end user device. Each user possesses a unique personal desktop environment. Program execution, data processing and data storage take place centrally on a personal desktop.  The information appears on the client screen via RDP/ICA/VNC/RGS or SPICE. The protocol for the reproduction of the correct information dependant on operating system, bandwidth, application properties and technical or company requirements.  Just as other solutions for desktop delivery, VDI consists of various infrastructure components that assure administration, load balancing, session control and secure access to virtual work stations. Suppliers of complete VDI solutions are VMware and Citrix. Suppliers of Virtual Infrastructure solutions are VMware, Citrix, Parallels and Microsoft.

Bladed workstations

Bladed workstation = “Dedicated  Physical Remote Desktop”

BladePC, a physical professional workstation in a blade enclosure, offers many of the advantages offered by VDI. A BladePC solution consists of hardware as well as software. The hardware consists of a physical professional workstation; the software assures that access to the physical workstation is possible.  The software consists of a minimum of two components: a transmitter and a receiver. The transmitter is installed on the BladePC and the receiver on the client desktop, laptop or ThinClient. The BladePC solution offers, in addition to the VDI advantages, even extra advantages such as:

Access to graphic intensive applications: In combination with HP Remote Graphics Software, this solution provides graphic intensive applications just as fast as the end user would expect.  2D, 3D and multimedia applications are executed centrally on the physical BladePC and reproduced locally on the client workstation via the RGS protocol.

Use of resource intensive applications: Resource intensive applications make full use of the resources present on the physical machine. It is a workstation-class centralized workstation. Performance is maximized through this.

Server Based Computing

Server Based Computing (SBC) = “Shared Remote Desktop”

SBC is a solution for access to desktops or loose applications on terminal servers in a data center. Access to the desktop or application is not bound to a location or end user device and program execution and data processing occur centrally on the terminal servers. The data are saved on a fileserver.  The information appears on the client screen via RDP or ICA. SBC consists of various infrastructure components for administration, load balancing, session control and support.  Some advantages of SBC are the rapid and secure availability of applications, low TCO, location and workstation independent application access.  Suppliers of SBC solutions are, for example, Microsoft, Citrix and Provision Networks.

Client Management

Each professional IT organization will make use of a Client Management solution. Client Management provides, for example, OS deployment, patch management, application and client deployment, asset management, integration with service desk and remote control.  Client management solutions are, for example, Microsoft System Center Configuration Manager (SCCM), RES Wisdom, Altiris Deployment Solution, LANdesk Client Management and Novel ZENworks.

In conclusion

The solutions that are reproduced in the solutions diagram assure that applications and desktops can be offered in various manners and in an effective and dynamic ways.

What is THE best solution? There are various business needs and technical requirements that together determine which solution is the best one for you.

In order to make the correct choice, it is important to know the pros and cons of the various solutions. Workshops are often held for this purpose. In addition to this form of information provision, there are various technical in depth articles written by Ruben Spruijt.

Read More on Ruben Spruijt

Read More on Ruben Spruijt

The complete article in English can be downloaded here, a Dutch version is also available here

Read More on Ruben Spruijt

Presentation #1, Hyper-V unleased can be downloaded here: http://www.virtuall.nl/seminars/BriForum2008/BriForum08HyperVunLeashed.pdf

Presentation overview:

• Microsoft Virtualization 360
• Virtualization within WS08
• Hyper-V benefits and disadvantages
• Managing Hyper-V and ESX with System Center Virtual Machine Manager
• 'War of the Virtual worlds'

Presentation #2, Streaming Smackdown 2008 Edition can be downloaded here: http://www.virtuall.nl/seminars/BriForum2008/BriForum08StreamingSmackdown2008.pdf

Presentation overview:

• Application and Desktop delivery solutions overview
• Application Streaming (Endeavors Application jukebox)
• Application Virtualization (Microsoft App-V, Altiris SVS, Citrix XenApp client side virtualization, Thinstall, InstallFree and Xenocode)
• Application Analysis (AppDNA)

At BriForum alot presenters and experts submitted there baby photos and 'fun facts' so Brian Madden could create a nice presentation for the attendees. After some searching in old photo books I found a nice picture of myself which fits great in the streaming smackdown session http://www.virtuall.nl/seminars/BriForum2008/BriForum08StreamingSmackdown2008.jpg

Presentation #3. Top10 challenges with SoftGrid (and other application virtualization solutions) can be downloaded here: http://www.virtuall.nl/seminars/BriForum2008/BriForum08Top10challengeswithSGandOthers.pdf.  Shawn Bass was also heavily involved in creating and presenting this session.

Presentation overview:

• OS/Shell integration
• Core 'vs' Virtuall
• IE, WMP and Plugins
• Services and drivers
• Sequencer gotchas
• Troubleshooting
• High Availability
• Branch Office
• Offline use /DR
• Performance

These three BriForum days were awesome and both from technical and personal point of view worth visiting. CU next time!? 

Read More on Ruben Spruijt

Below is the complete description of App-V package generation and all the corresponding terminology.

Package Generation: 30,000’ View

The life of a package file begins with the Sequencer.  The process of producing a package consists of three stages: you install the application, you run the application, and you save a package.  The files and folders that are created and modified during installation and execution become part of the package.  Prepare to dive to about 5,000 feet.

Package Generation: Sequencer Application Installation Stage

The installation stage is driven by the Installation Wizard, shown here at the top.  Below this, is a data flow diagram (or DFD) that roughly aligns with the stages of the wizard.  Circles represent processes or activities, rectangles represent external processes or activities, and the open boxes are data stores (both volatile and persistent).  Unlike similar looking state diagrams, the direction of an arrow shows movement of data.  This is typically from a provider process to a data store, from a data store to a consumer process, or directly from a provider to a consumer.  The label on an arrow clarifies the type of data that is flowing.

ACLs

One of the changes in 4.5 Sequencer is that the Installation Wizard prompts you for the installation directory when you start monitoring (and before you have installed the application) rather than when you stop monitoring.  SoftGrid 4.5 supports ACL capture and propagation from the Sequencer through to the Client.  This change was necessary to ensure the ACLs we capture are accurate and match what the application intends.  After you have specified the target installation directory—the last component of which becomes the package root—the security descriptors are initialized to known values.

Monitoring

The sequencer then begins the process of monitoring all file operations for new processes spawned by the shell or select system services.  This is when you run your installation program.

SxS

SoftGrid 4.5 includes better support for applications that install and use SxS assemblies.  One of the improvements is to dynamically privatize public assemblies that are installed by the installation program so they are immediately available to the application.  In the past, we have encountered application installers (such as Microsoft Money 2007) that needed these assemblies for later installation stages and would not work with the post-processing privatization that we provided in 4.2.  This is especially true if an application installer offers to check and install upgrades at the end of its installation.

Configuring Files

When you stop monitoring, the list of all files and directories that were created or modified by the application installation populates three lists: the list of user files, the list of application files, and the list of VFS files.

VFS Files

Any files that are created outside the specified installation directory become part of the VFS.  Copies of these files are moved to the VFS directory under the package root and into subdirectories based on the name of the CSIDL.  CSIDLs are symbolic machine-independent names for common file locations.  For example, CSIDL_PROGRAM_FILES refers to the program files directory, which could be named and located differently on different systems.

Each file and directory outside the installation directory is parsed against all standard CSIDLs and a VFS mapping is created.  This process is a vital part of the way SoftGrid makes applications independent of the system on which they are sequenced and run.  For example, if an application stores a file in C:\Program Files, but the application is run on a system where C:\Program Files is named something different, perhaps a localized version of the name, the process of parsing the VFS mappings is reversed so the file will appear in the correct location even though the target system is configured differently.

Package Generation: Application and Save Stages

The next stage is the application installation stage, which is managed by the Application Wizard in the Sequencer.  You can see from the DFD below that the overall process is very similar to that of the installation stage.  There are, however, a few of differences.  First, the VFS exists in this stage so file operations that occur to files and directories with VFS mappings are captured and handled by the VFS.  Second, the rules for determining what is application data and what is user data change.  Third, the monitoring includes file I/O (reads and writes), not just operations like creates, renames and deletes.

Save

For our purposes, the most interesting is the I/O monitoring since this is what drives the final stage, saving the package.  Because the Sequencer knows which byte offsets of which files are accessed during the application stage, it is able to place those blocks in feature block 1 and thereby create the minimum set of data necessary to satisfy the application execution scenario.  When you launch an uncached application, you only wait for FB1 to be loaded; all other data is brought down on demand or in the background.

All other file data, ACLs, and SystemGuard configuration also are written to the package along with the metadata that identifies the SoftGrid attributes of each file.

Application versus User Files

The SoftGrid package format and file system identifies package files one of four different ways: application data, application configuration, user data, or user configuration.  This is historical since the SoftGrid 4.x file system no longer distinguishes between data and configuration files.  In other words, user data and user configuration are treated the same on the client, and application data and application configuration are likewise treated the same.

Regardless of whether a file was created during the install or application stage, if it is an executable or SxS manifest, then it is application data.  From there, the logic differs depending on stage.  For files created during the installation phase, anything in the user profile is user data, and anything outside the profile with a .dat, .ini, .dot, or .mdw file type is user data.  All else is application data.

In contrast, everything created during the application stage is considered user data.  This is one reason why it is important to do perform the right operations in each sequencing stage.  If you do something like installing the bulk of the application during the Application Wizard, then you risk miscategorizing files.

There is a design change under consideration to simplify this logic so that, roughly, files in the user profile are user files and files outside are application files.  The Sequencer would still provide you a means for overriding this default heuristic.

Global Volumes

The SoftGrid file system manages package data using a collection of six container volumes.  You can think of these volumes as self-contained file systems within a single file.  Each volume stores different information based on the type of data, where it is located, and who modifies it.  The next two slides describe each of these volumes and illustrate when the volume is first initialized, what it contains, and where it lives.

The first two volumes are global across all users and packages.

By default, these volumes are both stored in the shared All Users or Public documents folder.  Users have only read access to these volumes and because they are opened by the file system at system start up, access to these volumes is usually through the SoftGrid file system itself.

Package Volumes

The next four volumes are unique to each package.  The first three, shown in red, are stored in the shared profile, and the last one, shown in blue, is stored in the user profile.

All these volumes are created when a Load operation is performed, or upon shutdown of the first time a package is launched.  The user package volume, because it is stored in the user profile, is only created for the user that is performing the load or the launch.

Data Organization: File and Directory Locations

This slide shows the locations of all these volumes as well as the temporary copies for a package.  The SoftGrid file system ensures the integrity and consistency of package data by always making modifications to a temporary copy of the file system volume, which then replaces the package volume when the application is no longer in use by the user (in the case of the per-user volumes) or all users (in the case of per-system volumes).  If the system is disrupted before the package is shutdown, then it automatically rollsback to the original volumes when the package was launched.

Data Management: Launch and Shutdown

This slide describes the activities that occur when a package is first launch by a user, first launched on a system, or subsequently launched by a user or on a system.  Four combinations can occur.  The two obvious ones are the first launch on a system by a user, a subsequent launch on a system by a user.  Two less obvious cases are a new user on a system on which the package has already been launched (such as would happen on a terminal server), and a user that has previously used a package on a new system, such as might happen in a terminal server farm.

First Launch

The first time you launch a package, the package data is streamed down to the client and stored in the file system data cache.  Temporary versions of the four package-specific volumes are initialized but don’t yet contain actual data.  As the application modifies files, the modified versions get written to the appropriate volumes.  When the application exits and the virtual environment closes, the temporary volumes are copied to their permanent locations and names.

Subsequent Launch

The second time the user launches the package, the permanent volumes already exist so they are copied to the temporary versions.  All modifications are made to the temporary versions and when the package is shutdown, they replace the previous permanent versions.

New User

Consider the new user on a Terminal Server.  The per-system volumes already exist, so it behaves as a subsequent launch.  However, the per-user volumes do not exist, so it behaves as an initial launch.

New System

Conversely, on a new system, the per-system volumes are initialized by a user that previously used the package behaves as a subsequent user launch.

Package In Use

Lastly, if a package is in use, then only the first user incurs the cost of copying and mounting the per-system volumes.  Other users only incur the per-user volume setup.  Conversely, the per-system volumes are only save when the last user of the package exits the application.  This has some subtle side effects in terms of performance and when changes are committed.  The first user to launch a package will have a slightly longer launch time than subsequent users.  And since changes to per-system volumes are only committed when the last user of a package exits the application, a system disruption will cause changes to the per-system volumes to be rolled back to when the first user launched the package, while changes to the per-user volumes will be rolled back just to the start of a user’s launch of the package.

Data Management: Administrative Operations

The three primary client management console commands that affect the file system volumes.  When you load an application, in addition to populating the file system cache with the package data, it pre-initializes the two global package volumes.  Unload deletes the two global package volumes.  The repair and clear options will delete the user package volume (and also delete the settings.cp file).

Load

There are several management activities that directly affect file system volumes.  When you perform a package load, it initializes the per-system volumes and initializes the per-user volumes for the user performing the load.  This operation behaves very much like a launch: at the start of the load, temporary versions of each volume are created and when the load completes, these are moved to their final package locations.

Repair/Clear

Repair and clear are the same from a file system perspective: they delete the user package volume for the user performing the repair.  Because the user’s SystemGuard personal settings live in this volume, a repair also has the effect of eliminating any modifications to the VFS or Virtual Registry made by the user.  Clear differs from repair because it also removes knowledge of the application, its FTAs and shortcuts from the user. 

Unload/Delete

Unload and Delete are likewise similar: they have the same effect on file system volumes.  The per-system volumes and the application data isolation volume are removed, and the package data is evicted from the file system cache.

Delete, like Clear, removes information about the application for all users.

4.0 Data Robustness Improvements

Data consistency - Metadata is maintained with file data in file system container volumes and updated as a single unit.

Data Integrity - Window of vulnerability for cache flushed reduced to just the periods when the file system cache is being modified. Modifications are made to working copies of file system volumes and synched on package shutdown.

Efficient - SystemGuard settings.cp stored in the user package volume and always reflects state of file system files.  (Shared profiles and profile errors may orphan files due to inconsistent file system and VFS state.)  Only data for package files modified by the user are stored in the per-package user volumes.  (Commingled package storage and copy-on-access leads to large profiles with many files in AppFS.)

Override and Permanent

With the change in 16476 to always delete the application data volumes on an upgrade, the purpose of the override and permanent bits is lost.  It is a loss we accept, but one we should lose consistently across the product.

Historically in 4.0, an application-specific file (‘application data’ or ‘application configuration’ SoftGrid file attributes) is replaced with a newer version in an upgrade package unless the Permanent flag is set and the Override flag is clear.  This is actually different than the 3.x behavior, which ignored the Permanent flag.  The table below shows this difference.

In both 3.x and 4.0, application-specific files that were modified by the user but were unchanged in the upgraded package were left alone during a package upgrade.  With the 16476 change, these are lost.

The security changes to isolate application-specific changes to a user via the application data isolation volume and the intended behavior on package upgrade to delete all application-specific data (now covered in 16476), means 4.2 and 4.5 no longer care about the Override and Permanent flags.

Since their behavior has never been well understood outside the Sequencing and File System teams, I suggest these options be completely removed from the Sequencer.  The Sequencer already marks all application-specific files with the Override flag by default and should continue to do so for backwards compatibility (whether or not that compatibility is officially sanctioned, since it is needed for phase-in and test).

Read More on Ruben Spruijt

• HTTP streaming
• New Sequencer UI
• Dynamic Suite Composition (DSC) for MSI packages
• Improved integration with SCCM 2007 R2
• Reporting. Application usage information is now recorded locally on each client and then sent to a App-V Management Server during Publishing Refresh. This means that offline usage of applications or usage of applications when streaming from different sources is now all accounted for properly in the App-V database. We are also providing new guidance so that customers can run their own customized reports.
• New Manageability tools. We’ve added a OpsMgr 2007 Management Pack to improve the monitoring capabilities of your App-V servers. We’ve added a ADM template to make it simpler to manage common client configurations via group policy. Finally, we’ve created a VSS Writer so that backups of the App-V server can leverage the latest technology
• Client cache improvements. 
• New MSI package capability. 
• Documentation, I reviewed the documentation. It's pretty good!

The RC version can be downloaded from http://connect.microsoft.com

Read More on Ruben Spruijt

[For reference, when Microsoft purchased Softricity, I spoke with Brian Madden in one of his podcasts  I also wrote an article   last year at the one year mark.]

When Microsoft announced the purchase, I stated that the real value of Softricity to Microsoft might not be in the "product as a whole", but in the pieces.  With the changes to the product being released this year, we are seeing some of that come true.

Today, when I look at what application virtualization involves, I break it into three parts:

• Application publication
• Application delivery
• Encapsulation/Isolation

What we had

The Softricity SoftGrid product did all three parts. 

The client would pull down the application publishing information at logon.  This  involves the shortcuts, icons, and file type associations needed, plus the OSD object to make the magic happen.  Some customers, especially when using Citrix and/or RES PowerFuse, were already turning off this part of the functionality and doing the application publishing with those products.

The application delivery with SoftGrid was performed by a client pull mechanism using the RTSP(S) streaming protocol.  Softricity could never get over just how cool this was and were effective in their Marketing message of a kind of "just in time" delivery.  In practice, we found we had to invent ways to get the client cache pre-filled with the assets to that the user never needed to wait for the streaming and could also work (disconnected mode) if the client lost connectivity with the server for some reason.  Technically, this streaming uses a special ("remote sparse matrix") file system that was specifically developed for the product.  At the time of the purchase I felt that Microsoft would probably find other uses for technology.

The actual virtualization engine doing the encapsulation and isolation (SystemGuard) was of course the real heart of the product.

What has happened so far?

 First of all, Microsoft has managed to not screw everything up.  It is hard for a large company to successfully bring in a start-up company and not destroy the product.  So let's give them credit (Citrix also seems to be doing this well, currently) for that.  First, Microsoft decided to give away all the server and sequencer pieces and charge a client license only.  This was good.  Yes, there are many customers unhappy with the client licensing model.    The SoftGrid for Windows Desktop client, was bundled into the Microsoft Desktop Optimization Pack (MDOP) and offered for a ridiculously low price.  This is a problem, however, because to purchase MDOP, the client OS has to be on subscription (Software Assurance), and SA is considered expensive given the long time that lapsed between XP and Vista (assuming the customer even wants Vista - but let's not go there).  Ultimately, I think Microsoft intends to make desktop SA work.  The fact that you can also get VECD (which you will want for VDI) only if you are on SA makes the cost of SA look a whole lot better.  At any rate, although non SA customers are unhappy, Microsoft has been shipping an awful lot of MDOP (reported at 6 million copies in the first 6 months).  In addition, we have seen good improvements to the virtualization engine.  Support for virtual services has been added, as well as for applications that are .net based, and a bunch of other things which increase the range of applications that can be supported. 

Finally, they also quietly included a new option for application publishing and virtualization without using streaming for the delivery.  This major new feature was included in a hot-fix rollup pack in February of this year.  And it was described as for use in one-off situations like for a contractor in India  and you don't to try to stream stuff over the internet.  But this week, I suggested using this feature (MSI based publishing) across the board to a customer with 50,000 SoftGrid client seats.  Basically, you take the publishing info and create an MSI that "installs" the publishing info using whatever push mechanism you use for non-virtualized apps (such as SMS).  The MSI populates the shortcuts and icons and file type associations, and them imports the asset file (sft)  to pre-fill the cache.  Configure the client to never look for a server and boom - you have a client with virtual apps with no back-end infrastructure to worry about.  Fewer moving parts is reason enough to give up things like app usage reporting in my book.  And if that reporting is important enough for you just use a tool that will monitor the clients and report both the virtual and non-virtual apps in one report.

What is coming soon?

The new 4.5 release coming out this year (Microsoft only says Q2/Q3 but the rumor mill puts money on a release candidate soon and RTM in probably September).  Also important to some customers will be SCCM R2 which will have a slightly later timeframe.  So what is in 4.5?

First, we have renaming happening.  The Softricity name clearly will be gone.  Microsoft is also trying hard to eliminate the SoftGrid name, but that is turning out to be more difficult than they realized.  "System Center Microsoft Application Virtualization" sounds awful to say even the first time.  Oh well, I'm just going to say MAV and move on to features.   The beta only included a small set of these, but the others were shown or discussed at the Microsoft Management Summit ( which I blogged about here).

Starting with the virtualization engine,  "Dynamic Suite Composition" (DSC) is the first attempt at solving a problem created because application virtualization virtualizes too well sometimes.  The problem is that when we containerize apps and run them in independent virtual bubbles, they also can't talk to each other.  We usually work around this by packaging multiple apps together into a single package.  If the application has too many dependencies this starts to be too much and we sometimes decide to install the major application (such as Microsoft Office) natively on the clients and virtualize the rest.  DSC allows us to make individual app packages and have the client perform a form of "mash-up" to bring them together into a single virtual bubble on the client.  This paragraph makes it sound better than it is, as there are special steps involved in the package creations due to the dependencies.  This will limit the use of DSC to a few things like Office and browser plug-ins.

There is a version of the MAV Server, which they call the "Streaming Server" (I call it the "lightweight server  because that name is less confusing).  The intent is to place this in a branch office.  The only purpose of the streaming server is to deliver the application (sft) via the normal pull streaming mechanism to the client.  This server does not deliver shortcuts, nor does it talk to the database.  Client overrides (registry settings at the client) are used so that the client can pull publication info from the home office at logon, and then stream everything locally.  You loose app usage reporting from these clients since nothing talks to a database if that is important.

4.5 will also automate the MSI option that came out in the hot-fix.  Rather than a separate tool to generate the MSI, it is embedded in the sequencer as an option.  By the way, the sequencer is getting a UI overhaul in the final release.  The three wizards are combined into one, and some options that were formerly in the wizards are now just new tabs in the sequence editor.

It is the delivery part which will have a lot of new architectural options.  In addition to MSI, we also have HTTP, SMB, and Bits.   While you can still use RTSP(S) with the VAS  (which is now called MAV Management Server), the new SCCM R2 release will allow you to ditch the VAS, SoftGrid Management Server, and SoftGrid Database and use SCCM instead for all of your virtual and non-virtual application distribution needs.  You will be able to specify HTTP(S) streaming from an IIS distribution point if you go this way.  There is also an SMB option ( file : // ) which I don't yet fully understand.   We also have background streaming using Bits (which may be related to the SMB option), but again you need to be on SCCM R2 so that you have the BITS distribution point.

Summary

The segmentation of the technology into those three parts, and providing new options for implementing them is a smart long-term move.  Maybe I'm not in love with SCCM, but the concepts work for any ESD to incorporate also.  Many of these options make us (myself included) look a little silly at poking fun at Citrix for not having "real streaming".  Streaming is a cool option, but quite frankly I would rather obtain delivery via other means.

Application Virtualization continues to make an awful lot of sense, not only for desktop and terminal server use, but for VDI in the future as well.  The concepts of encapsulation, segmentation, and layering are important to all forms of virtualization.

On the competition front, application virtualization continues to draw interest.  Two years ago we were not sure that Citrix would continue with AIE.  They did (now renamed) and have continued to improve their capabilities.  While the virtualizaiton engine of Citrix does not handle as high percentage of application as MAV, they have continued to make good progress.  Thinstall (now Vmware ThinApp) has arrived on the seen, as have others.  The Altiris/SVS/AppStream combo is also out there (I don't run into them, but if I don't mention them I'm sure "guest" will).

I think that Microsoft has done OK so far.  I might not make all the same choices as they made (like failure to deliver a 64-bit client for Terminal Services), but I would not have thought to make some of the right choices either.

Read More on Tim Mangan

The show continues to grow, with over 4000 attendees this year, and the registration had to be closed 4 weeks early due to space issues.  The exhibit floor was well visited by the crowd.  The isles were (artificially?) small the first day with the reception and with small tables in the isles forcing the crowd to within touching range by vendors in the booths.  This seemed effective in getting attendees to stop in the booths, but it sure made it hard to get around.

In addition to Microsoft, Intel, Dell, HP,  AMD, and  Citrix were present.  Additionally there were a number of vendors for products that add into this space.  There were also a few consultants with booths.  More on some of these booths later.

There were three keynote addresses at the show, one for each of the main days.  The day before and after also had many technical sessions - and I am kicking myself for not staying the extra day at the end because they added a couple of things I would have liked to see. 

Tuesday Keynote

On Tuesday, Bob Muglia, Senior Vice President of Microsoft's Server and Tools Business  was the primary speaker.   Titled, "Dynamic IT: Transforming Management & the Datacenter", Bob once again updated us on the "Dynamic Systems Initiative" (DSI) that he has been talking about for a few years.  This initiative is how Microsoft describes their efforts to help IT manage their systems and major applications.  (See this video from last year for more info on DSI) Microsoft has been building management capabilities for several years now, and will be continuing to do so.  Bob brought out several speakers to help address specific products and run demos.

Michael Kelly, Lead Program Manager for SCCM, demonstrated SCCM 2007 SP1[CHECK].  I noticed two new (at least to me) things in this demo.  First, they are working with partner vendors to help in using SCCM to configure vendor specific settings.  In the demo, he picked a Dell extension called the (Bare Metal) Server Deployment Pack  that you would download from Dell (available "before fall")  that added options for a PowerEdge Server.  This allowed him to create a new server configuration for that hardware that set up the bios for virtualization and enabled a raid-5 array, in addition to the usual OS configuration. (I talked to HP later on to see if they had such a pack.  They do not, but they have a white paper (see www.hp.com/servers/integration/microsoft ) that explains how to make your own based on  their "SmartScript".   Second he demonstrated the Microsoft Deployment Manager, which formerly was known as BDD.  This rename reflects that we are dealing with servers and desktops.  The second new item was that the Deployment Manager now support the optional use of Multicast.  This would allow multiple machine deployments of a common configuration using less bandwidth (but of course you have to be willing to enable multicast in your routers, which a shop large enough to be interested in this feature probably will not enable).

Next up, Rakesh {?}  talked about the new version of System Center Virtual Machine Manager, and he and Bob announced the availability of the Beta today.  The existing VMM supports the older Virtual Server product, but not Hyper-v.  This means that Hyper-v based machines (Hyper-v has a release candidate out but is not yet released) must be individually managed until the new version is released.

In the demo, Microsoft highlighted the multi-vendor support built into SCVMM.  They demonstrated using SCVMM  managing both a Hyper-V  and a VMware ESX cluster.  This included visibility of state and performance parameters.  They also demonstrated  moving a hyper-v based virtual machine from one hyper-v box to another.  Note that the virtual machine was not live, and Bob explained that live migration is coming but will not be in the release.  But they did demonstrate using SCVMM to do a v-motion on the VMware cluster.  With all this talk about being the first vendor with "heterogeneous " support for virtual machine management, it was not lost on me that somehow the word Citrix never came up in this demo.  So Microsoft can manage hyper-v and VMware (but not Xen, today) making the case that you need heterogenious support.  Citrix can manage Xen and Microsoft (I think). Hmm....

Also shown in this demo was the use of powershell to do all of the work.  Rakesh claimed that all of the commands run on the hyper-v servers are viewable and editable as powershell scripts which should allow people to easily develop custom scripts to do things that Microsoft does not supply out of the box.

Also shown, but not well enough explained was a checkbox when defining a server labeled "High Availability Server".  The claim was made that checking that box would automatically do everything needed to add the server to a HA cluster.

Day 2 Keynote

Brad Anderson, General Manager, Management and Services Division, ran the day 2 keynote.  Brad dove down into DSI in a keynote titled "Managing the Dynamic Desktop".  He began with a message here was not unlike that of Citrix - young people entering the workforce have different expectations of how they do things and called it "Extreme Mobility".  But the central core of the message was that we need to think user centric, not pc centric in our thinking.  This is a message that I really resonate with.  To the user, it is all about the user experience - having their "stuff", applications and data.  And this means whatever device or location they happen to be using.

DSI was described with four  core areas: 

• User Focused: " Delivering the right resources to users the right way".
• Unified -virtualized: "Managing the full combinations of all types of virtualization and physical"
• Process-Led/Model Driven: "Capturing Best Practices & Processes though Collective Knowledge"
• Service Enabled: "Guidance on how applications should be architected".

Edwin Yuen, Sr. Product Manager System Center.  Demoed a desktop with OS delivered via SCCM and with virtual applications on Vista. One underappreciated change in Vista is "offline file and folders".  This was used to keep files associated with application use centrally , not unlike "roaming profiles".   Also shown was the same user using Terminal Server to access the same virtual app and files, and then a VDI equivalent (using Xen Desktop), all with a "consistent working experience" . 

SCCM 2007 SP1 will be in May, and the R2 Release Candidate in July.  This will include updated capabilities of the  Asset Matrix  purchase to "Assess, Deploy, and Update from the Desktop to the Datacenter & Beyond".

MDOP 2008 was "announced" for Q3 release (not new news).  This includes Microsoft Application Virtualization (SoftGrid) 4.5, updates to Desktop Recovery, and other component who's name I did not catch.  They stated that MDOP 2007 was the "fastest selling version 1 product in the history of the volume licensing program".  They also stated the intent to make MDOP updates every 6 months in the future (in another session I heard "no less than once a year").

Dave Randall, Program Manager System Center performed demos around Unified-Virtualized.  Primarily highlighted integration of System Center and Intell VPro.  It is kind of like doing RDP into the NIC.  He Talked about having "15 scenarios" out of the box for out of band management.  So in these scenarios, you can manage changes to a powered off PC remotely.  They showed an example of powering off machines not being used.  They showed powering into bios only and making bios changes remotely - a remote control session without booting into the OS. They also talked to  an IDE redirection scenario where you have a bad OS and can boot the PC from another remote disk disk and access PC disk to repair.   Not necessarily new stuff, but integrated right into System Center.

For the Process-Led/Model Driven part of the keynote, the discussion was around driving operational processes based on models.  System Center uses four model types: Configuration, Health, Business Process, IT & Security Policy, Regulator Compliance, Capacity.  Individual products apply to subsets of these four model types.  These models are also extensible, such as Configuration Packs for SCCM, Management Packs and Reporting Packs for SCOM, and Solution Packs for System Center Service Manager (whenever that ships).  "Microsoft Operations Framework V4" was announced as being released to the web today.  This describes the framework in which to produce these packs.

Bill Anderson, Lead Program Manager,  System Center.  Performed NAP demos.   Protection/Quarantine, and remediation were shown.  In this demo, a PC was brought online and NAP checked it out.  The user was immediately notified of non-compliance and put into a quarantined state, then repaired and the user was notified and could access resources.  They also talked to supporting "what if" enforcement policies (i.e. apply a policy that doesn't quarantine to determine how many devices would be effected if  enforcement was turned on.

For the  Service-Enabled section,   they started with a describing of two types of services, "Finished Services" and "Attached Services".  The distinction was clearly lost on the crowd.  "Finished Services", in this sense are services offered  in the web.  Windows Update was stated as the largest existing example (600 million PCs were updated last month).  Also Asset Inventory Services.  "Attached Services" are services that the enterprise houses for lan-attached use. WSUS is the counterpart example to Windows Update;  Asset Intelligence is the other.  Maybe someone understood why this distinction was important?

Neal Myerson, Lead Program Manager System Center, gave a demo of something new.  This was a "live preview", not a product nor announcement yet, called "Attached Knowledge Services".  It is intended to be used to allow a company to ask the question, "how do my operations compare to other companies".  Basically, a company would send SCCM/SCOM data to Microsoft with identifying information stripped out.  A "Finished Service" there would process and present a scorecard to compare how you are doing in relation to other companies.  So , example, you could compare your client up-time stats to others.  You can filter who to compare against by things like industry type or company size.  You can also drill down into another company (without seeing company name) and see what hardware and what software and configuration they are using to achieve their results.    Eventually, Microsoft hopes to have a notification engine added so that Microsoft can advice/alert customers of potential problems based on their configuration when Microsoft sees and/or solves an issue for another customer.  I am sure that some will have issue with sending the data, even scrubbed, but there is potential here for a new kind of collaboration.

The keynote wound down with a display of the System Center roadmap for 2008/2009 releases.  We will have to wait for the slides to be made available, but there was probably nothing new on it that wasn't already known.

And the final teaser in the keynote was a demo was a thumb drive plugged in and giving access to virtualized apps and data.  I think it was Kidaro and SoftGrid, but they really didn't say.

Speaking of Kidaro,

I saw it for the first time here.  While Microsoft doesn't close on the purchase until sometime in May, they are moving forward quickly with plans.  Kidaro sits on top of Virtual PC to create a more seamless user experience.  It provides application shortcuts on the main PC that launches a virtual PC session in the background, runs the app there, and gives the user a seamless window like experience.  Copy/paste between sessions is supported also.  There is a bunch more as well (but also glaring gaps like not having file type associations for those apps taken care of).  Anyway, it will become part of MDOP as well.  There are some interesting licensing challenges on usage of Kidaro for some of the use cases they tout, like contractor PCs and Employee personal notebooks that Microsoft has not figured out yet.

Speaking of MDOP components…

We had an opportunity to see features of the upcoming MAV 4.5 release that have not been publicly seen yet.    In the public Beta we have seen the "Dynamic Suite Composition", which allows two virtualized apps (typically a base app and a plug-in) to be sequenced separately but run in a single virtual environment.  Here, the MSI output choice is native to the sequencer, Sequencer Gui changes try to simplify the process by moving lesser used functions out of the wizards and into editor tabs.  Also, there are some new deployment options.  In addition to RTSP and RTSPS, and MSI, we will be able to use HTTP or HTTPS streaming.  Microsoft stated in one session that initial tests were showing better performance with HTTP than RTSP (which sounds like performance bugs in RTSP to me).  Also, there is a sequencer output option to a file in addition to a sft (you get both the sft and an msi).  This would be distributed through whatever method (file share, thumb drive, etc) and would fill the cache without streaming, as is the case with the current external msi utility.  RC0 will be out "this summer" and release by the end of Q3.

Day 3 Keynote

Debra Chrapaty, Corporate Vice President of Global Foundation Services spoke today.  Her group is responsible for the strategy and delivery of the major infrastructure services within Microsoft, including things like Microsoft Live and Online Services, MSN, Hotmail, and Microsoft Update.  They have over 200 internal services with a minimum of 1000 servers each.  This was an interesting talk (especially for this crowd) about how they do IT from a company with one of the largest infrastructures in the world.  Some of the highlights…

Numbers:

Over 200 internal services that use a minimum of 1000 servers.

Microsoft Live Search: 2 billion queries /month

MSN: 559 unique users, 10 billion+ page views/month

Windows Live:  1 billion+ Authentications/day

Hot Mail:  3.4 billion spam messages/day

Messenger:  8.2 billion text messages/day

Growth  - In the last 5 years GFS has grown

Increase in #Servers 15x (10,000/month added)

Egress Bandwidth  9x

Power 15x

# Data Centers 3x

And this is expected to increase over the next 5 years.

She stated (but asked not to be quoted on this) that they estimate that they are running at an average of 19% utilization today and that by increasing that to 40% by 2011 could save Microsoft $2Billion.

She also talked openly about using Microsoft Technology.  Two years ago they couldn't use Microsoft tools to manage all this and had to roll their own.  But today they are moving to System Center and Virtualization, and will be dog-fooding the pre-beta code a unheard-of scale.  Some examples:

SCOM 2007 pilot now:

1000 production servers

Starting with SP1  RC

Migrating 39,000 MOM 2005 agents next

Expect improvements in scalability

SCCM

Migrating from home-grown tool to SCCM to collect asset info & patch compliance first

Virtualization

Moving some services to Hyper-V with goal of increasing those services to 30% utilization

Vendors

There were plenty of vendors, and if I was really a System Center kind of guy they might have been more interesting.  But here are a few highlights from my tour.

Acresso

This used to be Macrovision [EDIT: Not Marcomedia], the InstallShield guys, but was spun off.  News here is that in addition to generating MSIs they announced last year that they had an option to spit out files compatible with Citrix packages for AIE.  Now they are announcing the same capability for producing SoftGrid SFT files.  They claim it will be out "before SoftGrid 4.5 ships", which should be this summer.  Personally, I think that end-users get too interested in this sort of thing.  Ultimately, the customer has to customize the virtual application and you can't automate that away.  Still, it will be interesting to play with when we can get our hands on it.

SCCM Experts

They have a Self-Service Portal Solution to looks neat.  I asked them if they could handle self-service for SoftGrid apps (since Microsoft dropped the Softricity ZeroTouch) and it turns out that they can.  In essence, when approved they just assign the user in AD.

Splunk

A European based company, they have primarily been in the Linux space but have now added Microsoft capabilities.  They provide "Search Data for IT".  Basically, this product scans things like configuration files, log files, and event manager events, and index them.  This allows an IT admin to do an efficient search against this data.  Unlike a "google" search, this search technology is "time based", meaning that it understands time relationships such as "at this time the configuration file looked like this".

AVICODE

Have a .net monitor product that monitors .net usage at the CLR level.  I like the concept except that it is monitoring for errors instead of performance and can't tell you what to do with the errors.  The CA/Wiley guys on the other side of the hall do this for .Net and Java as well.

Special Operations Software

Quite honestly, I couldn't figure these guys out.  They do something with Group Policy Expansion.

Lakeside Software

A familiar name around Citrix (and briForum) they are present here as their products fit into the landscape.  Nothing really new to report from them.

TriCerat

Another familiar name, also without a new product to show.  But an odd presence as they don't have anything that really ties into any part of System Center.  Still their booth seemed to have plenty of traffic.  When I asked them why they indicated that the Lockdown and Desktop products were of interest to the attendees.

Closing Thoughts

Microsoft has come a long way in the Management Space.  Ultimately, large enterprises using Microsoft Windows are almost forced to invest in System Center someday because otherwise they end up with a clutter of incompatible vendor products.  Not that System Center is complete or completely compatible.  I personally find SCCM and SCOM to be poorly implemented, in that they require way too many resources for the benefit provided.  This is something that can be addressed by Microsoft, but as long as customers continue to ask for more features over improved performance it will not happen.

In a way, it is interesting how Microsoft lumps all forms of virtualization under the "management" banner.  Is virtualization a product or is it just a different means to manage our stuff?  David Greshler had a good line in one session about how with virtualization we are "turning servers into data".  But we will need more than System Center to manage all that data.

Read More on Tim Mangan

Application Service Providers  (ASPs) provide customers with a hosted application solution.  Although many may have thought that the ASP market died in 2001, it has managed to hang in there and is surviving if not thriving in some cases.  I have worked with three of them in the last year alone.  Often, these folks have settled on a niche for servicing smaller companies.  Their customers tend to be totally focused on a non-technical business that needs to use some computer technology.  As a result, these customers tend to have less than one full time IP person and do not want to hire one.  So the ASP makes an attractive solution to those kind of customers because they have IT professionals to ensure that the aspects of IT administration (picking the right software, applying best practices on IT administration, ensuring backups happen) are applied to their company.  Ultimately, it is the sharing of the professional IT staff that makes the ASP attractive to the customer.

ASP Today

To be cost effective these ASPs today use Citrix and/or native Terminal Services (TS) so that the hardware and operating system can also be shared. The customer uses a PC on their site and remotely accesses their hosted desktop using the RDP or ICA protocols.  The hosted site, in addition to the hosted desktop provides the applications as well as back-end file stores and databases.   Microsoft also helps enable this market with attractive pricing (Server Provider Licensing Agreement) for things like hosted Office and exchange. 

Complicating the setup is the need to ensure complete security to prevent any possible cross-customer data breech.  Active Directory (AD) is used within the hosting site, even if not in use at the customer site.  The tendency is for the ASP to build a single domain and create an  Organizational Unit for each customer. The ASP often has to add in Application Virtualization to eliminate application conflict.  These application conflicts may be due to application-TS issues (apps that are designed for the single user desktop), application interference (for example dll version conflicts between two applications), or, or Multi-tenancy issues.  

That last one, multi-tenancy, means application issues that arise when you want to host more than one company on the same terminal server.  For example, each customer might use Quick Books and each need to be able to see only their data.  The same goes for SQL or Oracle based applications.

ASP Tomorrow?

I am wondering if a Virtual Desktop Infrastructure (VDI) based approach might also emerge as a solution for this space.  A VDI based solution would use Virtual Machine technology to host a virtual windows image for each user within the customer. To keep things simple and offer the customer a best quality user experience, I imagine that these these windows instances may be personalized by the end user, however they won't be administrators on the windows instance.  The ASP is responsible for maintaining the image and applications so letting the user try to install whatever they want would be a really bad idea.

Using an approach of pooled standard images that are less customizable might be an option, but it wouldn't make much sense to me.  To start with, the ASP would need to build a custom image per customer, with that customers installed applications and configured to the customers back-end resources.  Starting with such an image and then cloning to private images per user within the customer would not be much more work, and yet would vastly improve the user experience.

I can imagine that the existing ASP could offer the customer either option.  In offering both options to the customer, they would be offering either a shared hosted desktop experience, or an individual hosted desktop experience. The VDI desktop should probably be priced at a premium by the customer when compared to a Terminal Server based desktop due to the improved (perceived?) user experience.

From the ASP perspective providing both options , VDI has mostly increased costs for VDI.  Each VDI instance needs a client OS, rather than sharing the TS OS.  Scalability per the hardware will also be lower for VDI.  I have heard other experts on this give factors for this scalability of anywhere from 2.5x1 to 10x1 (meaning that for the same hardware, or at least hardware dollars, you can host N users via Terminal Services for every one user that you could support using VDI).  Additional options like OS streaming and disk differencing might seem attractive as well, but not if they add to the infrastructure costs or make things more complicated.

On the flip side, the ASP should have a much simpler infrastructure with VDI.  For example, Application Virtualization may not be needed for these customers via VDI as all application-TS issues and multi-tenancy issues go away, and the likelihood of multi-application conflicts are greatly redu