Continue reading »]]> I have recently been invited to present in a VMware User Group (VMUG), which I quickly answered with a big and resounding Yes! I really love to present and touch base with customers as that helps me to keep it real, instead of being a lab rat type of engineer/architect.

Mostly I look forward to present and talk about new VMware endeavors in the EUC space, things such as MDM, MAM, the new HTML5 capabilities in Horizon View or what’s coming with the next release of Horizon Mobile. All fresh and cool stuff!

When I proposed one of my fresh new sessions I was asked to present “Getting VDI Storage Right”. My immediate reaction was to say that this was an old topic that every VDI administrator and architect had already mastered and that there was no need to present on such subject. I am sure I am in right mind when I say that because at the end of the day I have been writing about storage performance and architectures for VDI for over 3 years now.

Well, I decided to respect the VMUG board request and put together a session deck on the subject. When the VMUG day comes, with a packed room, I open my session with a rant about presenting on such topic but I kept on going. Why am I writing about it then? Well, at some point during my session I had a slide with the following bullet:

• Who in the room knows what the average & peak IOPS are for your VDI environment?

That’s when I went perplex! There was only a single hand up. I then decided to ask how many were actually currently doing VDI in some form. More than half of the room had their hands up.

It seems to me that all those hundreds of storage architecture and performance blog posts written by me and the blogger community have not been yet assimilated by organizations doing or planning to do VDI, or they are not hitting the target audience. That honestly makes me sad because I also think that to some extent this is laziness from storage administrators and VDI architects, but on the other end I also think that most storage vendors are not prepared to provide customers with proper assessment and guidance, whilst are more interested in completing the sale as quick as possible.

I work extensively with storage vendors and I can attest that many of them will say their solution is best for VDI but they have absolutely no clue about architecting such solution even thou they may have a joint architecture with VMware, Citrix or another VDI vendor.

I worked for EMC and they used to have a EUC/VDI specialist team to help customers. It appears to me this group has been thinned a little bit.  Atlantis Computing has a very good grasp on VDI as this is their bread and butter. Nutanix will say their solution scale linearly, like Atlantis, and they have recently on boarded some subject matter experts.

Other storage vendors have reached out to me in search for VDI expertise, and some of them are even using a personalized version of my VDI calculator on the field to help customers.

I honestly believed this whole storage conundrum discussion was over, but it is clearly not. With Microsoft announcing project Mohoro set to deliver Windows 8 “Desktop-as-a-Service” the whole discussion has been  reignited. How do we scale storage to accommodate for hundreds of thousands of desktops?

On the other side of the spectrum I plead you to make sure you are working with storage vendors that are indeed interested in helping you with your VDI solution and have the expertise to put you in the right track avoiding storage bottlenecks in the future.

A IDC, Goldman Sachs research from Dec, 2012 demonstrated that in the enterprise space from 95% market domination in 2005 Microsoft market share has fallen to just 20% by 2012. However, Windows is not going away anytime soon and Microsoft project Mohoro is a clear indication that the “Desktop-as-a-Service” is only warming up. I would love to know how Microsoft is architecting their own storage/compute solution for Mohoro.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> After publishing What’s New in VMware Horizon View 5.2 (Beyond Marketing) I received the same question multiple times from different customers and colleagues in regards to the following excerpt from my article.

A new Print Redirection codebase has been introduced that supports Windows 8, and also fixes a number of outstanding issues.

In the following table you can see what has been introduced with the changes in Horizon View 5.2. The changes are based on the functionalities of ThinPrint Server Engine 8.6.

• SpeedCache allow print jobs to be compressed by avoiding the repetitive sending of the redundant image data.
• Finishing support let users control the full range of features of multi-functional printers.
• Currently there are two forms of cloud printing – virtual and mobile printing. While only part of the printing process in virtual printing is relocated to the cloud, with mobile printing the entire printing process takes place in the cloud. Read more about it here.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> If you are wondering why my blog posting frequency has slowed down in the last couple weeks – worry not!

Many consulting professionals and architects asked for a portable – offline – version of my  well-known VDI Calculator. I’m happy to say that the new tool finally coming and I’ve been busy trying to reach code complete.

This new version is built on Java framework and will run on Windows, Mac and Android devices. There will be a special version specifically built for Android devices using native gesturing capabilities. No, no iOS thus far.

Keep in tune and don’t worry about the less than optimal user interface on the screenshot below – it will look great by GA timeframe, and it will also look great on Android devices.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> If you are attending EMC World 2013 in Las Vegas next week and is interested in how to effectively deliver Desktop-as-a-Service using EMC technologies you should attend my session with EMC’s Matt Cowger.

In this session we will discuss what DaaS really mean to organizations and what are the associated challenges - infrastructure resiliency, licensing costs, IO Profiles, self-service, storage design and architecture. From a VMware perspective I will also touch on different technologies available Today that make DaaS a reality.

Here is our session, make sure you register and attend!

VMware: Desktops-as-a-Service with VMware View & EMC

VMware View 5.1 provides an ideal platform for delivering desktops as a managed service for enterprise customers of all sizes. For many years, service providers have been daunted by technology and licensing challenges involved in building a viable service. VMware View can provide the foundation for building a service offering for delivering desktops from the cloud. Storage infrastructure is one of the backbones of a DaaS offering, however new storage technologies are rapidly transforming deployment models.

Now, if your organization is embracing BYOD, PYOD, Consumerization and Mobile workforce you should attend Scott Davis (VMware EUC CTO) session to find out how VMware can help you manage this new and diverse mobile workforce.

VMware: End-User Computing Vision for the Post-PC Era and VMware EUC Product Update

This session features VMware’s CTO for End User Computing discussing VMware’s vision for the future and the myriad of industry developments forming it. Multiple devices, application platforms, mobile devices, Bring Your Own Device initiatives, Cloud computing, and next generation applications are combining to a profound effect on IT. Learn about VMware latest EUC products including View, Mirage, and the Horizon Suite, DaaS, and a sneak peek and demos of technologies percolating in VMware’s labs to help IT manage and secure the Post PC era environment.

See you there!!

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> USB 3.0 is capable of transfer speeds of up to 5Gbps (gigabits per second); that’s a little over 10 times faster than USB 2.0′s 480Mbps. In practice, it won’t always be this much faster — mileage may vary depending on hardware configuration — but it will always be much faster than USB 2.0 (USB 2.0 vs USB 3.0)

USB 3.0 devices are not supported in Horizon View 5.2 or earlier. The instructions below are The steps outlined below are TOTALLY UNSUPPORTED BY VMWARE. I recommend testing in development environment. If you decide to test or implement you are doing it on your own risk.

In order to enable USB 3.0 for Horizon View Windows soft-clients you will need to create a Windows registry key which enables Horizon View client to see USB 3.0 devices.

[HKEY_CURRENT_USER\Software\VMware, Inc.\VMware VDM\Client\USB]
“AllowUSB3″=”true”

For Horizon View Client for Mac OS X you will need to use the shell (/Applications/Utilities/Terminal.app) to edit the settings. For more information on how to edit settings for Mac OS X refer to the documentation here.

#defaults write com.vmware.viewusb AllowUSB3 true

Confirm  with:

#defaults read com.vmware.viewusb

Depending on the USB 3.0 chipset you may have issues with Windows 7, Vista and XP Guest OSes.

To validate that USB3 functionality is enabled you should look for the following item in the log files

[vmware-view-usbd] DevFltr: Device Speed = Super

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> The VMworld Public Session Voting is now Open to anyone and allows VMware to poll the general public on which sessions they want to see presented at the event.

This year I decided submit only co-hosted sessions with people who are specialist in their product areas. People like Mac Binesh, Director for Horizon Performance; John Dodge, Tech Marketing DIrector; and Narasimha Krishnakumar, Horizon View Product Manager.

Any session with those guys is certainty of getting out of the session with a list of things to research or act on.

I am also  part of a very special panel discussion on Enterprise Mobility Management with no one less than Jack Madden from Techtarget and BrianMadden.com.

The not so good part is that even as a VMware employee I still need your vote to get sessions approved (well, I guess that’s better and fairer for everyone).

Please go ahead and click HERE to start your voting; and bellow you will find a list os sessions I am being part of. Thanks for your vote!

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]>

If you are in the San Francisco Bay Are make sure you attend the Silicon Valley VMUG User Conference. The event is scheduled for Wednesday, May 1st 2013 in the Santa Clara Convention Center. This event is almost a mini VMworld with many attendees and exhibitors expected.

I am honored host a session entitled “Getting Storage Right for your VDI Environment“. As you can imagine there’s a lot to be discussed about it.

The event is also filled with sessions from well-known community members. VMware’s own Paul Strong (CTO, Global Field & Customer Initiatives) is doing the morning keynote.

Click here for Agenda and Click here for Registration

Join your fellow VMware users for this full-day event including in-depth technical sessions, demonstrations and exhibits. This is a great opportunity to meet with your peers to discuss virtualization trends, best practices, and the latest technology!

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> Yesterday I published a blog post entitled “Why you should move your VDI to Windows 8” comparing different versions of Windows during IDLE and Login process. One of my readers (John Hardman) suggested that I also should measure disk IO commands issued from multiple Windows OS during boot time.

What I found in the additional test is another pretty compelling reason to move VDI environments from Windows 7 to Windows 8. In this test Windows 8 excelled when compared to Windows 7. During the boot process there were 54% less write IOs and 39% less read IOs at peak time in favor of Windows 8.

The average IO consumption, for both read and write, were approximately 50% lower when compared to Windows 7.

Windows 2K8 also demonstrated interesting results with lower peak and average IO for reads than any other Windows version. For write IOs it didn’t do so well, but still better than Windows 7.

Most VDI solutions do not have constant boot storms, but those that use floating pools and require constant desktop refresh could hugely benefit from the lower IO overhead provided by Windows 8 during boot time.

It seems that Windows 7 is actually the most demanding Windows version in regards to disk IO operations.

The storage performance conundrum may not be an issue for newer storage architectures, such as those using storage class memory devices, that can drive an enormous amount of IOPs at very low latency. However, it could also mean that architects are able to increase host density without increasing latency.

I did not run a production working set using system testing tools such as View Planner or Login VSI to completely validate the theory. If you have done so and is willing to share, please do!

Here are the results for the previous tests but I would recommend you going over to the article for better explanation “Why you should move your VDI to Windows 8“.

IDLE

1st Login

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> I believe I may have found a pretty compelling reason to migrate your VDI deployment to Windows 8. While running some experiments I collected some statistics for Windows 7, Windows 8 and Windows 2K8 Server running as virtual desktops.

At this point we should all be familiar with the challenges associated with the number and latency of IOs in VDI environments, therefore I am not getting getting into details in this article.

In short, IOps are the number of commands sent by the GuestOS to the storage, and these commands can often overload storage solutions when they are not properly sized for the VDI workloads. Write IOs normally have a higher cost than read IOs due to RAID striping that occurs for every write IO. IOPs and RAID together define the storage layout and the numbers of disks required. The RAID type will set the numbers of spindles required to support the VDI workload based on the amount of IOPS and Read/Write ratios — especially Write IOs given that RAID adds write penalty depending  on the RAID type chosen.

VDI workloads are not only made of boot and login storms. Applications, logging and other components also generate both read and write IOs.

If you want to learn more on IOPS please refer to Get hold of VDI IOPs, Read/Write Ratios and Storage Tiering. Also read VMware View 4.5 Storage Tiering explained and Sizing Storage for VMware View Linked Clones and Real Life example of Storage Sizing for View Linked Clones.

Anyway, back to the tests.

I measured the average and maximum number of IOPS requested by different GuestOS at two different stages. The first stage (below) is when the desktops have been IDLE for over 12 hours.  The graph below demonstrate that the maximum number of IOPS for both Windows 7 and Windows 8 are roughly comparable, with one Windows 8 anomaly for the desktop “Win8-02″. I am not sure what caused the anomaly.

In this first test both Windows 7 instances presented much lower write IO average when desktops are IDLE, however not as low as Windows 2K8. Windows 2K8 has the lowest average IO consumption of all three Windows versions.

On the second stage of the test I measured the IO consumption during the 1st user login. In this test Windows 8 excelled when compared to Windows 2K8 and Windows 7 on both, read and write IOs. Windows 8 had a write IOs peak 46% lower than Windows 7. On the average write IO the number was 20% lower in favor of Windows 8, demonstrating much more granular control over the IOPS overtime.

Before I give you my final conclusions I have to state here that this is not a scientific test and I beg you to run your own validation tests.

Microsoft has without doubt improved Windows file system performance, reducing the amount of IOs and data required to and from the storage. If steady state workloads in Windows 8 present the same benefits  and performance improvements demonstrated during the login process it could drastically help to reduce VDI storage costs since nowadays most storage costs are driven by performance, not by capacity.

The storage performance conundrum may not be an issue for newer storage architectures, such as those using storage class memory devices, that can drive an enormous amount of IOPs at very low latency. However, it could also mean that architects are able to increase host density without increasing latency.

I did not run a production working set using system testing tools such as View Planner or Login VSI to completely validate the theory. If you have done so and is willing to share, please do!

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> The year of 2012 was an important year for enterprise mobility management, where many start-up companies and large organizations put their stakes on the ground competing for a share in this $180 million market that is increasing at a compound annual growth rate of 15% to 20% during the next few years, according to industry analysts.

Early adopters of the technologies available have already recognized how transformative and fast paced the enterprise mobility management market really is. Some analysts and enterprise mobility experts are already paving a way out of MDM (Mobile Device Management) saying it’s a too heavy-handed approach to secure mobile applications and data, where IT literally take control of employee-owned mobile devices. To this point, there are lighter-handed MDM options available on the market that offer a good degree of control without being too intrusive.

The market answer is to move to a mobile app centric approach, the MAM (Mobile Application Management), where a number of different technologies and approaches are used to secure mobile apps and enforce polices and security. Some of the technologies in use by MAM are app wrapping, app secure VPN and even SDK driven app frameworks where mobile apps are developed on top of a framework that manipulate and handle security.

MAM usually supports custom-built enterprise mobile apps that have been built specifically to use the MAM vendor’s API; or in a more light-handed approach it will require the app to be wrapped with the MAM technology.

According to a Dec, 2012 research from the Yankee Group, employee-owned tablets used for work purposes has grown 16% from 2011 to 2012. This demonstrates not only that mobility in general is full steam ahead, but also that employees want to use mobile technologies to be more productive.

To get work done employees will look for apps that make them more productive, and if those are not available in the corporate app store they will likely go outside to find a new app the suits their ‘business’ needs. This has happened before with file-synchronization engines and is currently happening with note-taking apps.

If the app is not available in the corporate app store it’s possible that organizational content is created or moved to cloud services. On the offensive, vendors are exploring this market, enabling content sharing with added security features for corporate customers. However, if an employee wants to use a different service, or if tomorrow there’s a better app that will make them more productive, nothing will stop them from moving data or creating content using non-approved corporate apps and services.

Just like MDM, MAM is focusing on the security aspects of who is using the app, leaving a security gap that delineates the usage boundaries for the data. These boundaries reflect what and how employees are allowed to access and share data, independent of application or device.

A 2012 Forrester research suggests that 70% of the corporations want to use mobility to allow document viewing and employee access to portals, email and calendar. More than 52% of the organizations surveyed have implemented or are expanding video and photo application activities. The research goes into more detail, but the point I want to make is that line-of-business mobile applications are not high priority in most organization’s agendas.

Even if LOB mobile apps are responsible for a huge market, we need to ask the following question – What fraction of LOB mobile apps will require security around it to prevent data leakage? In many cases a simple photo taken or a screen-grab would circumvent the security mechanisms. However, we need to remember that MDM and MAM are not just about security as they also embrace life-cycle and deployment of mobile apps.

There are LOB mobile apps being used today for data input in select market verticals, and they are being adopted at fast pace. Some well-known use cases are: delivery and drivers (acknowledgements, routing and GPS), insurance (photos and forms), road-warriors (ordering and search) and finance (market watch and real-time statistics). A large amount of the devices carried by road-warriors are corporate-owned and many of them are specially reinforced devices to cope with heavy-duty operation. If devices are corporate-owned, then both MDM and MAM are practical solutions.

The same Forrester research demonstrates that 28% of the organizations that answered to the survey already allow employees to use any non-harmful application or device and they DO provide some support for the item, a 13% increase from 2011 to 2012.

I personally like the idea that ANY mobile app should be fully embraced by organizations, enabling the “True BYOD”.

Organizations need to worry about is who is accessing the data and what can be done with it. An app downloaded from iTunes, Google Play or Ubuntu store could have access to corporate data, but data could be geo-tagged not allowing it do be consumed from unauthorized areas, data could be meta tagged not allowing it to be consumed under specific conditions or by specific devices, data could be tagged with mechanics of how it can be shared across apps etc.

Allowing employees to use with their own mobile apps with content management doesn’t eliminate risks associated with content being intentionally moved or created outside the enterprise, but it gives options to employees to do the right thing.

Enterprises feel they must control consumerization and users, however unless they fully embrace the idea of content tagging and accepting that any mobile app can be used by employees to make them more productive we will never reach the BYOD nirvana.

Nothing I am writing here is new and all that has been talked and discussed before. We as an industry are not quite there yet, but as Aaron Levie (BOX CEO and Founder), mentioned in a tweet: “The product that wins is the one that bridges customers to the future, not the one that required a giant leap.”

The opinions expressed anywhere on this article or this blog are my personal opinions. Content published here is not read or approved and does not necessarily reflect the views and opinions of my employer.

Continue reading »]]> This is just a quick note to say that Nutanix is sponsoring premiere sessions of 2013, Star Trek Into Darkness all over the country (USA). To register to one of the sessions just click on the link here.  I have already registered for Santa Clara, CA. Here is the movie trailer.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> VMware Horizon Mirage is a layered image management solution that separates the PC into logical layers, which are owned and managed by either IT or the end user. Mirage enables enterprise IT administrators to centrally control the base and application layers guarantying a consistent view of the image across all endpoints. Through the Mirage Management Console, the administrator configures and manages clients, base images and app layers, performs operations such as updates and restores, and monitors the system operation through the dashboard and event logs.

In a SDDC environment automation is one the most critical pieces to ensure all parts work together. Despite Mirage provides all management capabilities via a full-featured dashboard, the lack of an supported API hinders the ability to promote datacenter automation.

So, I have 2 good news for Mirage administrators.

Good News N°1 – There is a web API in the Mirage Server. It’s hidden, clunky and undocumented - but it’s there.

Good News N°2 – I have created a VMware Mirage PowerShell Automation Module and I am sharing it with the broader community.

Important – the Mirage API approach is TOTALLY UNSUPPORTED BY VMWARE. I recommend testing in a development environment. If you decide to test or implement you are doing it on your own risk.

I am sharing this module with the borader community to enable administrators to create automation for Horizon Mirage taks via PowerShell. The cmdlets can be executed manually or be part of a larger PowerShell script, or even be triggered by automation and workflow tools like VCAC, vCenter Orchestrator or Puppet.

This version of automation module was created for and tested with Mirage V3.7_SP1 and no forward or backward compatibility is guaranteed. I know for fact that some of the cmdlets will not work with Mirage 4.0 and need to be updated given the Mirage API has changed since.

I am hoping that sharing this module with the community will enable administrators to understand how to implement it and build upon what I have created. If you update this module or add new functions and cmdlets I would appreciate you sharing back with the community.

The module contains the following functions:

• Add-MirageCollection
• Assign-MirageBaseImage
• Centralize-MirageEndpoint
• ConfirmRestore-MirageEndPoint
• Connect-Mirage
• DeCentralize-MirageEndpoint
• Get-MirageCVDBaseImage
• Get-MirageEndPoint
• Get-MirageEndpointSnapshot
• Remove-MirageCollection
• Remove-MirageEndpoint
• Restart-MirageEndpoint
• Restore-MirageEndPointSnapshot
• Resume-MirageEndpointNetworkOperations
• Suspend-MirageEndpointNetworkOperations

The functions (cmdlets) are self-explanatory, but additional information about how to use it can be found in the module header. To load the module, open a PowerShell session and type “Import-Module -Name mirage_module.psm1”.

Download the module here.

Here are some examples of how to execute the cmdlets after the module has been loaded.

Connect-Mirage -mgmtHost "x.x.x.x" -username "administrator" -password "password" -DLLpath "C:\Program Files\VMware\Wanova\"
 Get-MirageCollection -collection "collectioname"
 Add-MirageCollection -collection "collectioname" -filter "mirage01"
 Remove-MirageCollection -collection "collectioname"
 Centralize-MirageEndpoint -Endpoint "endpointname" -BaseLayer "baseline"
 DeCentralize-MirageEndpoint -Endpoint "endpointname"
 Get-MirageEndPoint -ContainerType:Pending
 Remove-MirageEndpoint -Endpoint "endpointname"
 Get-MirageSnapshot -Endpoint "endpointname"
 ConfirmRestore-MirageEndPoint -Endpoint "endpointname"
 Suspend-MirageNetworkOperations -Endpoint "endpointname"
 Resume-MirageNetworkOperations -Endpoint "endpointname"
 Restart-MirageEndpoint -Endpoint "endpointname"
 Assign-MirageBaseImage -Endpoint "endpointname" -BaseLayer "baseimagename" -MajorVersion 1 -MinorVersion 2
 Get-MirageCVDBaseImage

In order to support different version of VMware Horizon Mirage you will need to create a new DLL. To create a new DLL to make the module work with different versions of Mirage one should follow the following steps:

1. Replace Management Server configuration (“baseAddresses”) – replace “localhost” in “ <http://localhost:8443/WanovaAd… with the actual machine address – required for the next step to work
2. Create stub client code from WSDL: SvcUtil.exe
   http://<Management address>:8443/WanovaAdmin?wsdl – creates LicensedAdminApi.cs
   The resulting LicensedAdminApi.cs file will contain the service contract (the methods, resulting from the WSDL) and the data contracts (the data portion, coming from the XSD) for your service.
3. Compile the client: csc.exe /t:library LicensedAdminApi.cs – creates LicensedAdminApi.dll

If you cannot find csc.exe in your environment please follow this instructions to detect what .NET Framework versions and service packs are installed. http://stackoverflow.com/questions/199080/how-to-detect-what-net-framework-versions-and-service-packs-are-installed

Please note that even with a new WSDL/DLL the PowerShell module would need to be fully tested for compatibility. The reason being is that the Mirage WEB API is constantly changing and is not supported for remote interaction.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> It was 2010 when I first published in my blog the online VDI calculator for View and for XenDesktop. In April 2011 I stopped updating and supporting the XenDesktop version of the calculator, but it is still available in my blog.

The calculator has since been trough multiple upgrade cycles, and new functionalities have been added according to the VMware View product evolution. However, many people also use the calculator to size VDI deployments for other brokering platforms such as XenDesktop and vWorkspace, or other hypervisors, such as KVM and Hyper-V.

Today, the calculator is pretty much a standard VDI sizing tool for many professionals, including professional services organizations from the largest IT consulting companies in the world. I’ve been told that by many and I also track where the hits are coming from.

I keep track of the utilization, and one thing I was able to observe through the numbers is how the VDI market is still growing. In 2011 the daily average number of unique visitor was 91, hiking up to 134 in 2012, and in 2013 is already at 154 unique visitors per day.

The uptake on the calculator usage could be due to number of different reasons, including Google Page Ranking or people getting to know the tool. However, I know for fact that Google has been raking it 1st for a long time whenever you search for “VDI Calculator” and 2nd when searching for “Virtual Desktop Calculator”

I believe organizations are starting to look at VDI and Desktop-as-a-Service as an effective way to provide secure, manageable and easy access to applications and internal organizational resources, enabling the remote workforce, rather than placing focus on BYOD and cost reduction.

A recent article by Forrester’s analyst, David Johnson, describes what may be happening in the market and how and when VDI or DaaS is making sense to organizations. (HAS VDI PEAKED? A CHANGE IN THE ADOPTION DRIVERS SHEDS NEW LIGHT, AND NEW LIFE).

In the meantime the number of unique views on the calculator continues to rise, demonstrating that the uptake on VDI or DaaS is not anywhere near over.

The free VDI calculator can be found here.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> When VMware View 5.1 was release I wrote about the changes to storage and memory footprint when the 3D software rendering option was enabled.

• VMware View 5.0 3D Reverse Engineered
• vSphere 5.0 New .vswp file & Storage Tax on VDI

Just to recap:

ESxi 5.0 implements a second .vswp file for every virtual machine created either with hardware version 7 or 8. This second .vswp file is dedicated to memory overhead and will be used when the host is under resource constraint. When the virtual machine is created the memory overhead is defined, however the VM and the VMkernel will not use the whole reserved memory until required.

With Horizon View 5.2 and vSphere 5.1 new changes have been implemented, modifying the amount of storage footprint and memory allocated for the video .vswp file. The new values are more aggressive than used to be with vSphere 5.0.

Couple things to highlight before getting into the numbers:

• Horizon View 5.2 does not allow selection of 24-bit or 32-bit color palette. 32-bit is the default option.
• The resolution mode has been narrowed in Horizon View 5.2, only allowing administrators to select between 1680×1050, 1920×1200 and 2560×1600. 1920×1200 is the default option.

3D Software Rendering Disabled

• Memory footprint per VM in MB.

• Storage footprint per VM in MB

3D Software Rendering Enabled

When the 3D software rending is enabled for a desktop pool the amount of storage utilized for the .vSWAP file is not dictated by the resolution and number of displays, but rather by the amount of video RAM assigned for display. In previous releases of VMware View when 3D software rendering was enabled it would automatically add 256MB to the .vswp file – but that mathematics does not work anymore. I performed some reverse engineering to identify the most common cases.

• Memory footprint per VM in MB

When you create a VM using vCenter you will see the following information: “Extra 64MB plus 16MB per display”.

• Storage footprint per VM in MB

Now, you might be asking yourself. Do I need to compute all that into my VDI solutions design? The answer is yes, you do. The simply fact of adding 512MB video ram will increase the VM storage footprint in 1.9GB. However, you don’t have to be concerned about all this calculation because I have included in a the release of my VDI calculator (http://myvirtualcloud.net/?page_id=1076).

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> I have just release the Horizon View VDI Calculator v.3.1 with important updates to the graphics stack in order to correctly support the new 3D software and hardware rendering.

What’s New in Release 3.1

- Added support for new 3D functionalities in Horizon View 5.2
- Added support for 64, 128, 256 and 512MB Video RAM
- Removed 24 or 32 bit color option
- Video resolutions removed: 800×600, 1024×768, 1280×1024, 1600×1200, 1152×864, 1280×800, 1440×900

Along with the changes to the 3D support there are important storage and memory footprint overheads to be observed. I am writing another article to explain what is changing with vSphere 5.1 and Horizon View 5.2 when 3D support is enabled for a virtual desktop.

Check out the VDI calculator training video. You can find the video at http://myvirtualcloud.net/?p=2551

Check out the manual to help you to better use the VDI Flash Calculator. The manual can be found athttp://myvirtualcloud.net/?p=1927.

The VDI calculator can be found at http://myvirtualcloud.net/?page_id=1076. Alternatively, go to my blog homepage and select View Calculator on the top bar.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> Gabrie van Zanten at gabesvirtualworld.com published an updated to his PowerShell script that enables Horizon View to utilize dynamic IP addressing for PCoIP connections using whatsmyip.com. This pretty cool for people installing Horizon View at home for lab purposes.

Update: I have checked the script to work with VMware View 4.6, VMware View 5.0, VMware View 5.1 and VMware View Horizon 5.2. There was an update needed for the URL used to get the external IP address. Now using http://ifconfig.me/ip as URL. In Horizon View 5.2 there is a new URL, called the Blast External URL, to enable HTML5 for View. This URL can use a FQDN instead of IP so you don’t have to change it.

Check out his article here.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> Services Providers and large customers with requirement for multi-tenant Desktop-as-a-Service frequently ask me how would be possible to integrate Horizon View and vCloud Director and get both technologies to work together. As of today, the official VMware answer is – this is not supported.

In this article I demonstrate how multi-tenant implementations with View and vCD is achievable with some automation. However, let’s first look at why this is not supported today.

Most attempts to integrate View with VCD focus in placing View Connection Servers inside a vCD tenant organization. This works well for VMs and VApps created by and through vCD. However, View only communicates with vCenter Server, not vCD. For this reason, unless a dedicated vCenter server is placed in each vCD tenant organization with nested ESX’s you will not be able to have a completely independent environment. In this article I don’t plan to discuss vCD architecture. However, placing additional vCenters and nested ESX’s in a vCD tenant organization is not a scalable solution nor will excel in performance.

In addition to that, for a true multi-tenant environment you would need dedicated Active Directory domain servers hosted within each tenant organization. This AD server may be an extension of a customer network trough a VPN connection.

When Horizon View creates virtual desktops it will only communicate and place them in vCenter, not in VCD. Therefore, desktops will not be able to contact the correct AD server to be able to join the domain and execute other customization related tasks; nor will they be in a vCD organization. Therefore, vCD does not see the desktop nor manage resources for it.

The picture below demonstrates a deployment with Horizon View running inside a vCD organization and how it will communicate to vCenter only.

VMware does not support the steps outlined below. I recommend testing in development environment. If you decide to test or implement you are doing it on your own risk.

Enter a new undocumented feature in Horizon View 5.2 “HoldCustomization”.  This feature was a request I placed with engineering team  to help me to achieve multiple different objective; being one of them out-of-band vCD integration for Horizon View. When using this feature make sure the desktop pool is created with provisioning disabled, and only enable provisioning after the pae-HoldCustomization attribute is set to 1. If you need help to modify the ADAM database, please refer to my article How to Access Windows 2K8 Server via PCoIP with VMware View  for instructions.

When a Horizon View desktop pool is enabled with HoldCustomization the desktops will not be automatically powered-on after creation – this applies to both full and linked-clones. Because desktops have not been powered on it is possible to execute certain tasks before the customization process starts and before the domain join fails while trying to contact the correct Active Directory server.

As an example, one of the tasks that could be executed at this point in time is an Import-CIVApp task in vCD via API (PowerShell, Java, .NET). I am not going to explain in details how the function work in this blog post. Here is a code example:

Import-CIVApp -VM (Get-VM -Name $vm.name) -NoCopy -OrgVdc $org -RunAsync -Confirm:$false | Wait-Task

Once the VM is imported into vCD it will be automatically configured with identity, placed in the correct organization and registered with vCNS (vShield Edge). Now you need to automate the addition of the vCD VM into a pre-configured org-network using vCD API. Here is a sample of how to do that with PowerShell:

Function Configure_VMNetwork {
Param(
 [parameter(Mandatory=$true)]
 [string]$org,
 [parameter(Mandatory=$true)]
 $vm,
 [parameter(Mandatory=$true)]
 [string]$newNetworkName,
 [parameter(Mandatory=$true)]
 [Int]$NetworkConnection,
 [parameter(Mandatory=$true)]
 $IpAddressAllocationMode #$IpAddressAllocationMode = Pool, DHCP
 )
 $vmext = (Get-CIVM -Org $org -Name $vm*).ExtensionData
 $NetworkConfig = $vmext.Section | where {$_ -is [VMware.VimAutomation.Cloud.Views.NetworkConnectionSection]}
 if (($NetworkConfig.NetworkConnection).Count -eq 1) {
 ($NetworkConfig.NetworkConnection)[$NetworkConnection].network = $newNetworkName
 ($NetworkConfig.NetworkConnection)[$NetworkConnection].IsConnected = $true
 ($NetworkConfig.NetworkConnection)[$NetworkConnection].IpAddressAllocationMode = $IpAddressAllocationMode
 ($NetworkConfig.NetworkConnection)[$NetworkConnection].NeedsCustomization = $true
 } else {
 ($NetworkConfig.NetworkConnection | sort-object NetworkConnectionIndex)[$NetworkConnection].network = $newNetworkName
 ($NetworkConfig.NetworkConnection | sort-object NetworkConnectionIndex)[$NetworkConnection].IsConnected = $true
 ($NetworkConfig.NetworkConnection | sort-object NetworkConnectionIndex)[$NetworkConnection].IpAddressAllocationMode = $IpAddressAllocationMode
 ($NetworkConfig.NetworkConnection | sort-object NetworkConnectionIndex)[$NetworkConnection].NeedsCustomization = $true
 }
 $NetworkConfig.updateServerData()
 Return ""
 }

At this point in time the VM is ready to be powered on via vCD (must be via vCD because vCD executes few run-time configurations).

When the VM is finally back online, now in vCD, the desktop customization will continue normally and the desktop will be added into the domain and available to be connected via View Connection Server.

At his point in time you have a fully contained View environment with Connection Servers, Security Servers, View Composer, Active Directory and the desktops running inside a vCD tenant organization and all View related tasks such as refresh, recompose and reset will continue to work as per usual.

It is important to observe an implication while using this approach – vCenter Server is shared amongst View deployments in different organizations, therefore it’s important to harden Horizon View to avoid attack from inside the organization. In order to do that the attacker would need to have passed AD credentials or RSA token authentication to be able to start trying to do something. However, I am just raising awareness of the shared vCenter.

Automating View, vCenter, vCD with pre-created vAPP templates to create a multi-tenant infrastructure for desktop-as-a-service is possible. This guide doesn’t show you how to do it, but rather gives you an idea how to accomplish it.

There is another method to achieve similar results, where desktops are created in vCD and then added to manual pools in View. I’ll cover that in a different article. However, the downside of that is that you loose desktop lifecycle and pool management capabilities provided by View; such as refresh on logoff, recompose and refresh operations.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> Horizon View 5.2 provides Microsoft Lync 2013 client support, including full support for UC VoIP and Video on both RDP and PCoIP. This new feature enable a tighter integration between Microsoft Lync and Office applications with full collaboration capabilities.

Microsoft Lync 2013 client is a Unified Communications (UC) application that is installed on a desktop.  Microsoft Lync client is used for presence, instant messaging (IM) and other functions, such as placing a Voice over Internet Protocol (VoIP) or Video chat call.

In order to enable the Lync 2013 client to pair up with the Lync VDI plugin running on the client endpoint, VMware implemented Microsoft’s dynamic virtual channels (DVC) inside PCoIP. The Microsoft DVC provides a communications path between the virtual machine (VM) and the client endpoint, and the Lync 2013 client uses the DVCs to communicate with the Lync VDI plugin running on the client endpoint. The graph bellow depicts how the solution is architected.

Using Horizon View 5.2 and Microsoft Lync 2013 together will enable the following features to be used:

• Presence
• Instant Messaging
• Desktop Sharing
• Application Sharing/PowerPoint sharing
• Whiteboards
• File Transfers
• Online Meetings
• Communication modes for Online Meetings are limited by peer-to-peer communication modes.  For example, if audio is not supported on the specified architecture, audio will not work in Online Meetings.
• Joining online meetings from Microsoft Outlook meeting reminder and/or meeting invitation is not supported.
• Office Integration
• VoIP
• Audio is supported only in a VDI environment. Audio is not supported in a session-based desktop delivery environment such as Microsoft RDS.
• Video Chat
• Multi-party video chat is not supported.

The solution works with Windows 7, Windows 8, Windows Embedded and Windows Thin PC. Before using Lync 2013 administrators must ensure that the client hardware meet the following specifications: CPU 1.5GHz, at least 2GB RAM and 4GB total storage.

Software Components	Description
Microsoft Lync 2013 Server Components	Backend server that the Microsoft Lync clients register with.
Microsoft Lync 2013 client	Microsoft Lync 2013 client application that is installed in VM.
Microsoft Lync VDI plugin	The Lync VDI plugin is installed on the client machine.  Lync VDI plugin that pairs with Lync 2013 client in VM.  Once Lync VDI plugin is paired with Lync 2013 client, all media processing is automatically redirected to the client device.
View 5.2 (Agent and Client)	Need to use the View 5.2 agent and corresponding View 5.2 client
VM Operating System	Supported on Windows 7 VMs only with Service Pack 1.
Client Operating System	Supported on Windows 7 and Windows 8 client OS.  Please note that Windows 8 clients is only offered in Tech Preview.

There are also some additional restrictions to be observed:

• PCoIP process running on the View client (32 & 64-bit) only operates in 32-bit mode.  Therefore only the 32-bit version of the Lync VDI 2013 plugin is supported.
• Client Machine – the 32-bit version of the Microsoft Lync VDI 2013 plugin will not work from a machine that has the 64-bit version of Office 2013. This is unsupported by Microsoft.
• RDP 8 is not supported.

I would like to acknowledge here the excellent work from Tony Huynh and his team to get these functionalities working. In my next post about Lync 2013 I will go a little deeper and explain how the communication works under the covers and provide some troubleshooting guidelines.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> View Pool Manager allow Horizon View administrators to easily manage users across desktop pools and security groups in large deployments.

One of the of the biggest challenges View administrators face in large deployments is the desktop pool maintenance. Commonly each desktop pool has an Active Directory security groups with the correct amount of users a given desktop pool will support. As employees leave or join the organization the security groups become messy, requiring maintenance to ensure that the number of users in each group correctly match the amount of desktops available in each desktop pool.

This little tool will help administrators to save time adding, removing and maintaining users across desktop pools and security groups while helping to ensure that each security group does not contain more users than the amount desktops in the desktop pool.

The tool allow administrators to bind an Active Directory connection and specify:

• Number of users per security group (based on View pool size)
• Floating or Persistent deployment model
• Source Security Groups (all users to be distributed to pools)
• Destination Security Groups (All groups where users may be assigned to)

Why is this tool awesome?

Because View administrators do not have to waste hours adding, removing and maintaining users, desktop pools and security groups while ensuring each security group does not hold more users than the amount desktops in a given pool.

Download the tool from VMware Fling website here.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> I have just release the Horizon View VDI Calculator v.3.0 with full support for Horizon View 5.2 and vCenter 5.1. This is a major release of the calculator with multiple enhancements to mostly support higher scalability with the new versions of View and vCenter.
 
What’s New in Release 3.0

- Added support for pools with 2,000 desktops
– Added support for new processors with 12 and 16 cores
– Added support for user-defined VM limit in vCenter
– Added support for user-defined vSphere cluster size
– Added support for 10,000 desktops with vCenter 5.1
– Multiple backend changes to support Horizon View 5.2
– Multiple backend changes to support vCenter 5.1
– Enabled high availability features as Default options
– Removed support for vCenter Server earlier than 5.0
– Removed support for View earlier than 5.0
– Removed cluster size dependency on storage protocol

vCenter VM Limit -  In previous versions of the calculator the limit was pre-set by the calculator. With vCenter 5.1 supporting up to 10,000 desktops you may now set how many VMs should run run per vCenter Server. Increasing or Reducing the number of VMs per vCenter will define the number of individual blocks per Horizon View pod.

Max Cluster Size – vCenter 5.1 and View 5.2 introduced support for up to 32 hosts per cluster with NFS and Block storage technologies. Based on your design and use-case you are now able to define the maximum number of hosts per cluster.
 

Support for new processors with 12 and 16 cores

What’s Next?

SEsparse – For the next small release I will be adding support for the new SESparse disk format. However, SESparse reclamation process is based on disk utilization and that will drastically vary according to use-cases and usage. To size for SEsparse today, if you know you are able to reclaim 2GB every time wipe/shrink operation is sun, than set the total disk size to a smaller number.

Check out the VDI calculator training video. You can find the video at http://myvirtualcloud.net/?p=2551

Check out the manual to help you to better use the VDI Flash Calculator. The manual can be found at http://myvirtualcloud.net/?p=1927.

The VDI calculator can be found at http://myvirtualcloud.net/?page_id=1076. Alternatively, go to my blog homepage and select View Calculator on the top bar.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> During many of my customer briefings, conference sessions  and discussions with friends I frequently notice that many of the terms used in EUC (End-User Computing) are yet unknown to many – so are the benefits of the multiple different technologies being created and implemented in enterprises today – things such as MDM, MAM, EMM etc.

In an effort to disseminate information about this emerging market that according to analysts is worth U$180B I recorded a series of videos covering the basics of EUC. These are short videos (about 2min) that cover a different part of the EUC spectrum in each episode. I’m calling it EUC 1:1

This fourth episode covers EMM (Enterprise Mobility Management) and Mobility Market. I hope you enjoy it!
 
(Watch it in HD)

If you cannot see the embed video click here.
 
This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> This week I received an interesting question from a colleague that I had also been asked in the past. However, in the past I was not able to present a good solution for the problem. Let me present the scenario.

A VMware View pool of persistent desktops was created and provided to developers using the Linked Cloning technology. Over time the developers deployed their own applications and created content in their virtual desktops. The storage administrator is now requiring the VMs to be moved around and distributed to different storage LUNs for maintenance and performance reasons.

Storage vMotion doesn’t work in this situation and is not supported with Linked Clones. VM cloning will also not work, failing with the following error message:

Another option is VMware View Rebalance operation, but that one will refresh and reset the virtual desktops to their pristine state. Learn more about Rebalance and other Linked Clone operations in my article VMware View 4.5 Linked Cloning Operations Explained (Part 2).

The VMware KB “Converting a linked clone virtual machine to a full clone virtual machine (1026753)” will also explain how to use vmkfstools to convert the linked clone VMs into full clone and than add it to a VMware View manual desktop pool.

However, my colleague didn’t want to convert Linked Clone VMs to Full Clones because they would lose manageability over the desktop, such as refresh and recompose operations. For this same reason V2V solutions were also not an option because we would have to let View Composer do it’s customization job before doing the migration.

The solution: Horizon Mirage.

Mirage keeps a complete virtual copy of each end point synchronized with the data center. The administrator can re-assign the virtual copy to a new end point (virtual or physical) and all user profile, user data, and user-installed applications will be restored to the new device – even if the new device is a different hardware make or model, or virtual instead of physical. Mirage is hardware and driver agnostic as it created different driver layers for different hardware models.

The administrator is now create a new Linked Clone VM in the correct LUN and use Horizon Mirage to migrate all the data from one linked clone to another linked clone VM.

Mirage also leverages de-duplication in storage – This means that data is only stored once no matter how many users or end-point device exists in a organization.  Additionally, all data is compressed before it is sent and  SSL can be enabled between all client-server communication.

If you want to learn more about Mirage read my article VMware Mirage Architecture and Use Cases.

Please note that at the time of writing this article Horizon Mirage integration is not officially supported by VMware.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> During many of my customer briefings, conference sessions  and discussions with friends I frequently notice that many of the terms used in EUC (End-User Computing) are yet unknown to many – so are the benefits of the multiple different technologies being created and implemented in enterprises today – thinks such as MDM, MAM, EMM etc.

In an effort to disseminate information about this emerging market that according to analysts is worth U$180B I recorded a series of videos covering the basics of EUC. These are short videos (about 2min) that cover a different part of the EUC spectrum in each episode. I’m calling it EUC 1:1

This third episode covers MIM (Mobile Information Management). I hope you enjoy it!

(Watch it in HD)

If you cannot see the embed video click here.
 
This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> One of the most common requests I get from customer and partners is – When will VMware View support Windows 2K8 as Guest OS?

As most of you probably know Microsoft does not provide SPLA licensing model for Windows 7 to allow customers and service providers to create Desktop-as-a-Service offerings using Enterprise grade VDI solutions such as Horizon View. However, Windows Server 2K8 does provides SPLA licensing model and if those customers and service providers are able to deliver Windows Server 2K8 combined with the new Horizon View Multi-VLAN feature they could effectively service a large part of the DaaS market without incurring into Microsoft licensing issues.

Please note that I am not inferring that Horizon View with Windows 2K8 is effectively Desktop-as-a-Service, like some of the vendors out there. I personally believe DaaS is much more than simply delivering Win2K8 desktops, and that it should embrace key important concepts such as Mobility, Portability, Automated Provisioning, Identify management, Federation, multi-tenancy at all layers, chargeback, and an easy-to-use self-service web portal where user can consume services. This is subject for another article. However, the solution below, despite unsupported, will effectively put customers and service providers in alignment with Microsoft licensing.

If you are an avid reader of my blog you will remember that I published back in 2012 a guide on how How to Access Windows 2K8 Server via PCoIP. That hack is currently being used by many customers, but it was not effectively easy to deploy because it required the administrator to hack into the VMware View database and manually add each desktop. Furthermore, Linked Clones and many other features such as printing and USB redirection were not working. But, you don’t need to go back to my old article because all you need is here and in a much simpler way.

I cooked up something TOTALLY UNSUPPORTED BY VMWARE that will allow customers and partners to provision VMware View desktops using Windows 2K8 Server, PCoIP, Linked Clones, ThinPrint, SmartCards, Persona Management and USB redirection; all in a natural and automated fashion.

Important Note: The techniques below has been only tested by me for the latest VMware View 5.1 release.

The only changes you will have to make to get the solution working is to the standard VMware View Agent 64-bit before you install it on your parent VM. VMware View Agent uses Microsoft Installer to run a series of validations to understand what components should or should not be installed and what conditions should follow. For Windows 2K8 specifically, VMware View agent will only allow Terminal Services Access Pools with the connection protocol being RDP. So, first we need to open and hack the View Agent.

You will need to download Orca. Orca MSI Editor allows you to edit the properties of any MSI file. With this tool, you can change the title and text within the installer and look at how and where the files are delivered. Once installed you can right-click any MSI and select Open with Orca. This tool used to be a part of Microsoft Developer Tools but is now retired and no longer supported.

Download Orca from here.

The steps outlined below are not supported by VMware. I recommend testing in development environment. If you decide to test or implement you are doing it on your own risk.

1. On a Windows 2K8 R2 Server with latest VMware Tools installed execute the installation VMware View 5.1 Agent package (VMware-viewagent-x86_64-5.1.2-928164). Do not click Install or Next.
2. Go to %temp% folder, typing %temp% at the Run prompt.
3. Find the folder containing the unpacked installation files and copy them to a temporary folder on your desktop.
4. Using the Orca utility open ‘VMware View Agent64.msi’ file existing in the temp folder.
5. Navigate to ‘Condition’ table.
6. Delete the conditions from PCoIP (0), SVIAgent (0), SmartCard (0), ThinPrint (0), and USB (0) features. Delete only the content of the condition, not dropping the entire row. Save the changes.
7. In ControlEvent table set:
1. Dialog CustomSetup, argument ConnectionServer, set condition to 0
2. Dialog CustomSetup, argument ReadyToInstall, set condition to 1
8. In InstallExecuteSequence Table set:
1. Action VM_AddLDAPMachineEntry_SD.2EDC2186_29E9_4662_86CD_26051967D0CA, set condition to 0
2. Action VM_AddLDAPMachineEntry.2EDC2186_29E9_4662_86CD_26051967D0CA, set condition to 0
9. In Registry Table set:
1. Registry vcManaged, set Value to #1
10. Save the VMware View Agent64.msi and copy all the files to a temporary folder in your Windows 2K8 VM.
11. Execute ‘VMware View Agent64.msi’ and go through the install normally.
12. This is a critical step to ensure VMware View Composer is will customize the desktop and bypass the OS version check. Replace C:\Program Files (x86)\Commom Files\VMware\View Composer Guest Agent\vmware-svi-ga.exe with the patched file you download from here.
13. Open the Windows Device Manager; go to ‘Display Adapters’ and select ‘Update Driver…’
    Select ‘Browse my computer for driver software’ and ‘Let me pick from a list of device drivers on my computer’. (optionally, if you see the VMware 3D driver available in ‘Display Adapters’ you may select it and skip step 13.
14. On the following screen select ‘Have Disk…’ and select the wddm_video driver at C:\Program Files\Common Files\VMware\Drivers\wddm_video\vm3d.inf. Without changing the video driver PCoIP will not work.
15. We are almost done! The next procedure will ensure that new View Composer internal disk goes online when attached to the Linked Clone VM. Open command prompt and type DISKPART.exe. Now type DISKPART > san policy=OnlineAll. Type exit.
16. Close the command prompt window shutdown the server.
17. Next you will need to change the VM OS type to Windows 7 (64-bit) so VMware View can recognize it as a valid Parent VM for View Composer.
18. Take a Snapshot of the VM and you are done!

From now on just create desktops pools the same way you always do. Use Automated Provisioning, Dedicated or Floating, Horizon View Composer linked clones. Here is my Windows 2K8 configuration:

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> Identifying what replica is assigned to what desktop pool can be somewhat challenging in Horizon View. However, in some cases administrators need to know the object relationship to properly plan desktop and pool migrations across hosts or datastores.

The PowerShell script bellow produces a output that demonstrate this relationship.
  

  
To execute save the script below as a .PS1 file in a View Connection Server with the VMware.VimAutomation.Core plug-in installed. The VMware.VimAutomation.Core plug-in is automatically installed with the VMware vSphere™ PowerCLI. Before executing the script you will need to change the vCenter server address at

 <p>Connect-VIServer -Server vsphere5.company.com

  
Here is the script:

Add-PSSnapin VMware.VimAutomation.Core

function Get-LinkedClone {
 #The following line is a fast replacement for: $vms = get-vm args[0] | get-view
 if( $args[0] -eq $null ) {
 # $vms = Get-View -ViewType VirtualMachine -Property Name,Summary,Config.Hardware.Device
 $vms = Get-View -ViewType VirtualMachine
 } else {
 $vms = Get-View -ViewType VirtualMachine -Property Name,Summary,Config.Hardware.Device, -Filter @{Name = $args[0]}
 }

 $linkedClones = @()
 foreach ($vm in $vms) {
 Write-Host "The name of the vm is " $vm.name
 $unshared = $vm.Summary.Storage.Unshared
 $committed = $vm.Summary.Storage.Committed
 $uncommitted=$vm.Summary.Storage.Uncommitted
 $provisioned= ($vm.Summary.Storage.Commited + $vm.Summary.Storage.Uncommitted)

 $ftInfo = $vm.Summary.Config.FtInfo

 # This identifies if it is thin provisioned.
 if ( ($unshared -ne $committed) -and (($ftInfo -eq $null) -or ($ftInfo.InstanceUuids.Length -le 1)) ){
 # then $vm is a linked clone.
 # Find $vm's base disks.
 $baseDisks = @()
 foreach ($d in $vm.Config.Hardware.Device) {
 $backing = $d.backing
 if ($backing -is [VMware.Vim.VirtualDeviceFileBackingInfo] -and $backing.parent -ne $null) {
 do {
 $backing = $backing.parent
 } until ($backing.parent -eq $null)
 $baseDisks += $backing.fileName
 }
 }

 $isReplica=$vm.name.Contains("replica-")

 if (($baseDisks -ne $null) -and -not $isReplica)
 {
 # $replica=$baseDisks[0].Substring($baseDisks[0].IndexOf("]") +2 , $baseDisks[0].IndexOf("/"))
 $slashpos=$baseDisks[0].IndexOf("/")
 $replicaDisk=$baseDisks[0].Substring(0, $slashpos)
 $rbracketpos=$replicaDisk.IndexOf("]") +1
 $replica=$replicaDisk.Substring($rbracketpos).Trim()

 #$VMName = $vm | select -expandproperty name
 #$VMUnshared = $vm | select -expandproperty storage | select -expandproperty perdatastoreusage | select -expandproperty Unshared
 #--- Get the disk space assigned

 Write-Host "the VM " $vm.name " has this many disks" $vm.Guest.disk.length
 $totalDiskSpace = 0
 $diskSpaceAssigned = 0

 for ($iCount = 1;$iCount -le $vm.Guest.Disk.Count; $iCount++)
 {
 $index=$iCount-1
 $diskSpaceAssigned = [math]::Round($vm.Guest.Disk[$index].Capacity / 1GB)
 #$Details | Add-Member -Name "Disk$($DiskNum)FreeSpace(GB)" -MemberType NoteProperty -Value ([math]::Round($disk.FreeSpace / 1GB))
 $totalDiskSpace=$totalDiskSpace + $diskSpaceAssigned
 }
 $GuestTotalDiskSpace = $totalDiskSpace
 $uncommittedGB= [math]::round($uncommitted/1GB,2)
 $VMUnsharedGB = [math]::round($unshared/1GB,2)
 # $VMUsed = $vm | select -expandproperty storage | select -expandproperty perdatastoreusage | select -expandproperty Committed
 $VMUsedGB = [math]::round($committed/1GB,2)
 $VMprovisionedGB=[math]::round($provisioned/1GB,2)
 $linkedClone = new-object PSObject
 $linkedClone | add-member -type NoteProperty -Name Name -Value $vm.name
 $linkedClone | add-member -type NoteProperty -Name Parent -Value $replica

 $linkedClone | add-member -type NoteProperty -Name DiskCount -Value $vm.Guest.Disk.Count

 #$linkedClone | add-member -type NoteProperty -Name BaseDisks -Value $baseDisks
 #$linkedClone | add-member -type NoteProperty -Name DiskProvisionedGB -Value $VMprovisionedGB

 $linkedClone | add-member -type NoteProperty -Name UniqueDiskUsedGB -Value $VMUnsharedGB
 $linkedClone | add-member -type NoteProperty -Name VMCombinedStorageGB -Value $VMUsedGB
 $linkedClone | add-member -type NoteProperty -Name GuestTotalDiskSpace -Value $GuestTotalDiskSpace

 $linkedClones += $linkedClone
 }
 #else { do nothing for VMs that are not linked clones }
 }
 else
 {

 Write-Host $vm.Name "Not a Linked Clone" -ForegroundColor 'Red'
 }
 $stan = $linkedClones | sort Name, DiskCount, VMCombinedStorageGB, UsedStorageGB, GuestTotalDiskSpace
 $stan | ConvertTo-Html > c:\temp\output.html
 }
}

Connect-VIServer -Server vsphere5.company.com
Get-LinkedClone</pre>

  
Thanks VMware EUC Senior Consultant Hans Bader for sharing the script with us. However, please not that I have not heavily tested the script and if you find any issues or want to improve it please share with us.
  
This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.  

Continue reading »]]> Eric Siebert (@ericsiebert) from vSphere-Land has just published results for the 2013 Best VMware and Virtualization Blog Election. My blog finished in 14th place, coming down from the 17th position last year and 39th place the year before.

Thanks to all my readers for the votes!

The announcement was made during a special vChat  recorded with special guests John Troyer, David Davis and Simon Seagrave. I would like here to acknowledge the amazing work Eric and his troop do every year to get this voting going. Finally, I would like to congratulate all bloggers out there for the amazing work we all do sharing knowledge and helping others. Thanks!!!

Here is the full 2013 top bloggers list.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> In my article Clearing up Space-Efficient Virtual Disk questions I explained the most important requirements and constraints when the new SESparse VMDK disk format with Horizon View 5.2.

Just in cased you missed it:

• This new feature requires vSphere 5.1 as well as hardware version 9 for the VMs.
• The Space Efficient format is only for for linked clone pools; and only for the OS disk.
• Horizon View automatically chooses the space efficient format when possible and there is no official way to change this behavior.
• The Space Efficient Format is not yet supported for Windows 8.
• The space reclamation process is kicked off automatically by Horizon View when the reclaimable space for the VM exceeds the threshold set by the administrator.
• The space reclamation statistics (reclaimed amount, reclaimed time) are displayed in the Pool Summary, and Desktop Summary.
• SESparse disks ARE COMPATIBILE with the Horizon View Storage Accelerator technology from View 5.1 / vSphere 5.

Despite many customers are already successfully using the new feature and are effectively saving on the overall storage capacity utilized, some customer asked me how to disable this feature and re-instate VMFS Sparse as the default disk type. For many VMware internal reasons SESparse is the Default behavior and cannot be changed from the Admin UI. However there are couple ways to disable it.

Please note: If the intuit is only to disable the shrink/wipe process, this can be done via Admin UI disabling storage reclamation or blocking the maintenance window for shrink/wipe operations.

• Create your Parent VM using Hardware version 8 instead of 9. Horizon View will not be able to assign SESparse on HW v8. Please note that some new Horizon View 5.2 features will not work with HW v8.

The steps outlined below are not supported by VMware. I recommend testing in development environment. If you decide to test or implement you are doing it on your own risk.

• If HW v8 is not an option for your deployment or if you are already using HW v9 in production you will need to change a parameter in Horizon View Active Directory Lightweight Directory Service (AD LDS) database.

1. Connect to one of your connection servers via RDP and execute the ADSI Edit tool or adsiedit.msc
2. Use the connection settings below to connect to the internal ADLDS (DC=vdi,DC=vmware,DC=int). Click OK.
3. Navigate to cn=common,ou=global,ou=properties.
4. Under cn=common,ou=global,ou=properties, create pae-NameValuePair (multi string)
5. Then set: "cs-disableSeSparse=true"

Done! All new VMs from now on will utilize VMFS Sparse disk format.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> An article by Jack McMichael (@jackmc4)

Starting in View 4.6, a state was added to VMware View to deal with a nagging issue for some administrators who wanted to ensure that a desktop was always clean and never previously used – the introduction of the ‘Already Used’ state. The issue is simple, a user logs into a non-persistent desktop and something horrible goes wrong. Perhaps a Windows blue screen or an abnormal crash of the View Agent, but ultimately something that left the virtual machine in an unusable state until rebooted.

Until the ‘Already Used’ state was created, there were situations where a VM would crash, automatically reboot, and then be put back into circulation for use as if it has never been used despite the fact that it had. This posed a problem that was resolved with the creation of ‘Already Used’. The idea was simple enough, once a VM has been logged into, toggle a flag to mark it as used and when the Pool is being evaluated, don’t allow the use of VMs in this state.

Note: The View Desktop can also go into a Already Used state if a virtual machine is powered on in another ESX host in the cluster in response to a HA event or if it failed before a power on. VMware KB1000590

In some larger environments where ongoing VM issues sometimes can and do exist, this created an issue of pools being filled with ‘Already Used’ VMs that needed to be manually deleted or otherwise dealt with.

With the release of View 5.1.2, I created a new pool attribute called ‘pae-DirtyVmPolicy’ that deals with this very issue. It has 3 options:

0 – The default value that will mark VMs as already used when logged in.
1 – A new value that will tell the pool to ignore any dirty VMs and just recycle them. This is valuable if your ultimate goal is to just have available VMs.
2 – A new value that will immediately remove or refresh the VM depending on the pool policy. This is the most destructive value, but ensures that you always have fresh VMs available in the pool.

If the pae-DirtyVmPolicy value is not present in the pool, the default behavior is automatically chosen. This is the preferred value since we want to catch problem VMs and diagnose them before deleting or refreshing them. Sometimes administrators just need to ensure that pool levels remain available, and so this new option gives them that ability. The attribute must be manually added in existing installations, but should be present in new installations, and should be put in each pool where the policy must be defined in OU=Server Groups.

Also refer to “Automating Desktop in ‘Already Used’ State in VMware View” and “Automating Desktop in ‘Already Used’ State in VMware View –Part 2”.  Another good article on the subject “Cleaning up “Already used” desktops in VMware View”.

This article was first published by Andre Leibovici (@andreleibovici) atmyvirtualcloud.net.

Continue reading »]]> I am happy to finally be able to scratch this item from my list of missing View features.

• Support for Multi-VLAN

In Horizon View 5.2 it is possible to configure an automated pool to use multiple VLANS or network labels. Administrators can assign multiple network labels to a linked-clone pool or an automated pool that contains full virtual machines.

Going back to the previous releases of View and understanding the problem.

When creating pools, each desktop inherits the network label of its parent’s template/snapshot. This network label specifies a portgroup/VLAN/DHCP IP address range. For large pools, this meant 1000s of desktops all forced into the same large VLAN, requiring enough IP addresses on the same subnet for each.

On the management side, administrators either had to accept this or implement workarounds to achieve industry standard VLAN sizing of /24 subnets of 256 IP addresses, either creating multiple smaller pools with different parent images, each with their own unique network label specifying the desired subnet size; or manually in vCenter or with the use of a script, modify the network label of each desktop VM to overwrite its parent’s network label after provisioning. This would have to be re-done after recompose/refresh or after modifying the pool size. This was one of the most common hurdles in implementing large pool sizes.

Understanding the new feature…

• Each pool can now be configured with a “network label spec”. This allows an administrator to specify, for each network adapter/NIC on the parent, a list of desired network labels to automatically assign to each newly provisioned desktop in the pool.
• Each label is given a maximum assignment count. After View has provisioned, assigned, and exhausted a network label to its configured capacity for a given NIC, it moves onto the next in the specification list.
• Horizon View remembers the per-pool count of desktop VM’s and their labels provisioned under this feature, so if the pool size is later changed, the configured network label capacity and assignments will be honored.
• This feature only assigns network labels to desktops upon first provisioning – previously created desktops are never modified.
• Network labels are treated similar to MAC addresses – refresh and recompose operations do not revert the child to its parent’s network label state.

Now, this new awesome feature can only be configured thought Horizon View PowerShell and does not have Admin UI support at this point in time.

Configuration is done in three steps:

1. Create a skeleton network label specification text file using parameters from a Cluster/Host and parent VM.
2. Edit the network label specification file to associate available network labels with available NICs, along with maximum capacity counts.
3. Create a new pool or update an existing pool using the network label specification file.

To understand how configure and implement  multi-VLAN support please refer to the Horizon View 5.2 Integration guide avaiable here.

Important Note on Upgrading Existing Pools

The PowerCLI to update an existing pool with the feature is easier to use than creating it from scratch, due to the ability to use Powershell’s piping functionality to populate the myriad other parameters. However, if the upgraded pool initially contained any desktop VMs not provisioned under the feature, the network label assigner will not count any labels on those desktops towards any maximum capacity counts.

Those original desktops can optionally be deleted after the feature is configured to have them re-provisioned and included in the count, or the administrator can configure the feature expecting them to not be included.
Likewise, VMs on the VC not created in the pool do not have their network labels counted against maximum capacity for this feature.

The PowerShell helper…

As you use the tool you will notice that creating  the configuration file with the correct parameters can be a daunting task due to the amount of parameters required.

Export-NetworkLabelSpecForLinkedClone -vc_id 1a2b3c4d-5e6f –clusterPath "/myresourcepool/host/Cluster1/" -parentVM “/myresourcepool/vm/Win7-Parent" -parentSnapshotPath  "/Win7-Parent/snapshot1"  -maxVMsPerNetworkLabel 244 -networkLabelConfigFile "C:/label.txt"

To make things easier for the administrators I created a PowerShell script that automatically gather all required parameters, including vd_id, full and relative paths to clusters and pools, and than automatically pipe them to pipe into the Export-NetworkLabelSpecForLinkedClone or Export-NetworkLabelSpecForFullClone.

Place the script in one of the connection servers an run it as follows:

>> Export-NetworkLabel -PoolName “pool_template_name” -Path “c:\label.txt”

There is no need to specify Full or Linked Clone as the script will automatically identify the pool type.

<#</pre>
<em id="__mceDel"> .SYNOPSIS
 Network Label Config File Automation for Horizon View 5.2</em>

Created by Andre Leibovici

THIS CODE MAY BE CITED, REPRODUCED OR DISTRIBUTED WITHOUT EXPRESS AUTHORIZATION.

.DESCRIPTION
 Automates the creation of NIC and Network Label Cofiguration File.
 Network COnfiguration File allow administrators to enable multi-VLAN pool deployments.
 Must be execute on a VMware View Connection server.

For more information on multi-VLAN support refer to VMware View Integration Guide.

.PARAMETER -PoolName Display Name an existing of template pool in Vmware View
 -Path Path to save the Configuration file. By Default c:\label.txt

.EXAMPLE
 Export-NetworkLabel -PoolName "pool_template_name" -Path "c:\label.txt"
 #>

Function Export-NetworkLabel {
 Param (
 [parameter(ValueFromPipeline=$false,Position=0,Mandatory=$false)]
 [string]$PoolName,
 [parameter(ValueFromPipeline=$false,Position=1,Mandatory=$false)]
 [string]$Path = "C:/label.txt"
 )

Add-PSSnapin "VMware.View.Broker" -ErrorAction:SilentlyContinue

$vc = Get-ViewVC -ErrorAction:SilentlyContinue
 $pool = Get-Pool -DisplayName $PoolName -ErrorAction:SilentlyContinue
 if (!$pool) {
 Write-Host "Pool not found"
 Break
 }
 $clusterPath = ($pool.resourcePoolPath.Split("/") | Select -first 4) -join "/"

if ($pool.desktopSource -eq "SVI") {
 Export-NetworkLabelSpecForLinkedClone -vc_id $vc.vc_id -clusterPath $clusterPath -parentVM $pool.parentVMPath -parentSnapshotPath $pool.parentVMSnapshotPath -maxVMsPerNetworkLabel 244 -networkLabelConfigFile $Path
 } elseif ($pool.desktopSource -eq "VC") {
 Export-NetworkLabelSpecForFullClone -Vc_id $vc.vc_id -ClusterPath $clusterPath -TemplatePath $pool.templatePath -MaxVMsPerNetworkLabel 244 -NetworkLabelConfigFile $Path
 }

Return ""
 }
 

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> During many of my customer briefings, conference sessions  and discussions with friends I frequently notice that many of the terms used in EUC (End-User Computing) are yet unknown to many – so are the benefits of the multiple different technologies being created and implemented in enterprises today – thinks such as MDM, MAM, EMM etc.

In an effort to disseminate information about this emerging IT market that according to analysts is worth U$180B I recorded a series of videos covering the basics of EUC. These are short videos that cover a different part of the EUC spectrum in each episode. I’m calling it EUC 1:1

This second episode covers MAM (Mobile Application Management). I hope you enjoy it!

(Watch it in HD)

If you cannot see the embed video click here.
 
This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> During many of my customer briefings, conference sessions  and discussions with friends I frequently notice that many of the terms used in EUC (End-User Computing) are yet unknown to many – so are the benefits of the multiple different technologies being created and implemented in enterprises today – thinks such as MDM, MAM, EMM etc.

In an effort to disseminate information about this emerging IT market that according to analysts is worth U$180B I recorded a series of videos covering the basics of EUC. These are short videos that cover a different part of the EUC spectrum in each episode. I’m calling it EUC 1:1

This very first episode covers MDM (Mobile Application Management). I hope you enjoy it!

(Watch it in HD)

If you cannot see the embed video click here.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> Horizon View 5.2 make use of the new Space-Efficient Sparse Virtual Disks available in vSphere 5.1. I previously wrote about it in Space-Efficient Sparse Virtual Disks and VMware View and announced the new Horizon View 5.2 feature at What’s New in VMware Horizon View 5.2 (Beyond Marketing).

“The new SE sparse disk implements a space reclaim feature to reclaim blocks that were previously used but now are unused on the guest OS. These are blocks that were previously written but currently are unaddressed in a file system/database due to file deletions, temporary files, and so on.
There are two steps involved in the space reclamation feature: The first step is the wipe operation that frees up a contiguous area of free space in the virtual machine disk (VMDK); the second step is the shrink, which unmaps or truncates that area of free space to enable the physical storage to be returned to the free pool.

Wipe
• Initiate a call to VMware Tools to scan the guest OS file system.
• Mark the unused blocks as free.
• Run the SCSI UNMAP command in the guest to instruct the virtual SCSI layer in the VMkernel to mark the blocks as free in the SE sparse disk.

Shrink
• SCSI device – VMware® ESXi™ issues an SCSI UNMAP command to the array.
• NFS device – ESXi issues an RPC call to TRUNCATE the file.

The wipe operation is initiated via an API call to VMware Tools. VMware Tools initiates a scan of the guest OS to find stranded space and mark the file system blocks as free. The first SCSI UNMAP operation is then run from within the guest OS, instructing the VMkernel as to which blocks can be reclaimed. The VMkernel captures these SCSI UNMAP commands and does not pass them through to the array. When the VMkernel detects which blocks are free, it uses its virtual SCSI layer to reorganize the SE sparse disk by moving blocks from the end of the disk to unallocated blocks at its beginning. This creates a contiguous area of free space within the VMDK. The shrink operation then sends either an SCSI UNMAP command (for SCSI disks) or an RPC TRUNCATE command (for NFS) to the array to free the space.”

Last week during VMware Partner-Exchange in my session EUC1222 that I co-presented with Robert Baesman (@rbaesman) we also briefly discussed the new disk format. Few questions came up from the audience and I would like to also address them here.

• This new feature requires vSphere 5.1 as well as hardware version 9 for the VMs.
• The Space Efficient format is only for for linked clone pools; and only for the OS disk.
• Horizon View automatically chooses the space efficient format when possible and there is no official way to change this behavior.
• The Space Efficient Format is not yet supported for Windows 8.
• The space reclamation process is kicked off automatically by Horizon View when the reclaimable space for the VM exceeds the threshold set by the administrator.
• The space reclamation statistics (reclaimed amount, reclaimed time) are displayed in the Pool Summary, and Desktop Summary.
• SESparse disks ARE COMPATIBILE with the Horizon View Storage Accelerator technology (Understanding CBRC) from View 5.1 / vSphere 5.

If you want to test the new Space-Efficient Sparse Virtual Disks feature in Horizon View 5.2 here is a quick walk-trough to configure and run execute the space reclamation process.

1. Ensure vSphere 5.1 is installed.
2. Create a Windows XP or Windows 7 VM in vSphere using the latest virtual hardware version, install the Tools & View Agent, and snapshot it to serve as a base image for a View Composer based View Desktop.
3. Install and Configure the View Connection Server appropriately  and ensure that space reclamation is enabled for vCenter.
4. Set up a space reclamation blackout time in Horizon View.
5. Create a new Automated, Dedicated, Horizon View Composer linked clone based pool and provision & entitle 1 or more desktops in the pool.
6. Ensure that space reclamation is enabled for the desktop pool.
7. Run a workload in a desktop in the pool that creates and deletes lots of data in the Windows file system, or manually copy and delete large files.
8. Observe growth of the VM’s disk over time in VC.  You can see this in VM in VC under Summary -> Resources -> Used Storage.  Make a note of how much it has grown.
9. Re-observe the VM’s disk consumption shrink following a reclamation.  You can see this event occur by either:
   (a) Waiting up to 1 hour for the automated process to kick in
   (b) Kicking the reclamation process of manually using the vdmadmin command

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]> Last week I published the VMware Horizon View 5.2 video (here) and due to it’s success I decided to also record and release a VMware Horizon Workspace video. The video has two parts, Horizon User Workspace and Horizon Administrative Workspace.

Please make sure you go to my article Introduction to the new VMware Horizon Suite to read about market background and understand all the new  features in the Horizon Suite solution.

Next you will find a VMware official video that demonstrate the proposition of VMware Horizon Workspace.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]>  

Back in November 2012 I blogged about an unsupported combination of VMware View features (View Storage Accelerator and View Storage Tiering). The question had originally been raised by a colleague, pointing to an article published on the End-User Computing Blog http://blogs.vmware.com/euc/2012/05/view-storage-accelerator-in-practice.html that clearly states “Use of View Storage Accelerator is not supported when View Replica Tiering is enabled”.

The VMware View Tiering technology was created to allow View administrators to place replica disks on Solid State Disks. With the replica being serviced from high performing disk devices all Linked Clone desktops benefit from the IO performance and throughput, particularly during boot storms.

With the introduction of VMware View Storage Accelerator (find more about it here) in View 5.1 and having the most common data blocks across all desktops in a host being serviced from RAM, the idea behind VMware View Tiering didn’t make sense anymore from a performance perspective and that’s the reason why both technologies have not been qualified or tested together by VMware.

Having the ability to use dedicated replica datastore via VMware View Tiering can be a huge storage capacity saving feature, in special for all the new all-flash based arrays where capacity utilization is far more critical than performance. When using Linked Clones with Dedicated replica datastore only a single replica is created per pool of desktops (up to 1000 desktops per pool). The picture below demonstrate the difference between the two models.

If you are interested in reading more about both technologies and how they complement each other please refer to my article View Storage Accelerator and View Storage Tiering [Unsupported].

So, I brought the subject up internally at VMware and demonstrated (over multiple meetings and conversations)  that our customers would benefit from both technologies being used together. I am happy to announce that VMware has qualified and tested the combination of both features – and this is now fully supported with with Horizon View 5.2.

I am looking forward to Nasimha Krishnakumar’s blog post making it official!

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.

Continue reading »]]>

Eric Siebert has officially opened the pools for this year’s vSphere-Land Top bloggers List.

If you like VDI,
If you like EUC,
If you like this blog,
If you like my articles,
If you like Horizon View,
if you like Horizon Suite,   
If you like VMware products,
Please consider voting for this blog.   

In 2010 the blog was amongst the top 40.
In 2011 the blog was amongst the top 20.
Let’s aim for top 15 this year!

Click on the button below to initiate voting

Continue reading »]]> Giving continuity to my VMware View Limits and Maximums series (VMware View 4.0 | VMware View 5.0) I am now releasing an updated version for Horizon View 5.2.

As an administrator or architect you should always make sure that your design sizing’s are within the product’s Maximums and Limits in order to be entitled to VMware Support.

Horizon View 5.2 introduced many new features and improvements, including a more scalable architecture that now supports 10,000 desktops using a single vCenter Server (I’ll soon post an article about ups and down for single vCenter architecture).

The scalability improvements introduced in Horizon View 5.2 come from four very important features: Multi-Network Support, Accelerated View Admin Performance, Space Efficient Disk Utilization and support for 32 host clusters when using NFS or VMFS. Read more about the new Horizon View 5.2 features in What’s New in Horizon View 5.2 (beyond Marketing).

Limits

The limits may vary according to the releases in use. The limits in this post are specific to Horizon View 5.2 and vCenter Server 5.1. The comparisons are against limits published with VMware View 5.1

· 32 Hosts per Cluster when used with VMFS – changed from 8 Hosts per Cluster

· 32 Hosts per Cluster when used with NFS – did not change

· 16 VM’s per CPU core – did not change

· 1,000 VMs per View Composer desktop pool or replica  – did not change

· 140 VMs per LUN with VAAI support – did not change
Without VAAI support the recommended number is still 64 VMs per LUN. This limit comes from the number of SCSI LUN reservations caused by VM metadata updates. With VAAI the reservation happens at the VMDK level.

· 10,000 VMs per vCenter – changed from 2,000 VMs per vCenter
VMware is now officially supporting 10,000 virtual desktops per vCenter Server, installable or appliance. This is a major architectural change and should drastically reduce View pod footprint and enables easier backup and DR for the entire solution.

· 1000 VMs per host – did not change
This limit is established by vSphere 5, not VMware View.

Maximum Number of Connections

· 1 Connection Server with Direct connection, RDP, Tunneled or PCoIP, 2,000 – did not change

· 7 Connection Servers (5+2 spares) Direct connection, RDP or PCoIP, 10,000 – did not change

· 1 Connection Server with PCoIP Secure Gateway, 2,000 – did not change

· Total HTML 5 connections per pod, 256 (To be Confirmed) – New Limit
This is a new limit for connections coming via HTML5 protocol.

History

I find amusing to identify how and when each component of the overall solution was improved or upgraded. The table below demonstrate when each component was upgraded and what is the new limit.

This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net.