A little while ago VMware released Horizon View 6.1, which for the first time allowed Linux VDI desktops among many other new improvements. All customer that have Horizon 6 Enterprise Edition, or VMware Workspace Suite are entitled to use Horizon Linux Desktops. Supported Linux distributions include Ubuntu, RHEL, CentOS, and NeoKylin. Importantly VMware also supports 3D GPU Hardware Acceleration with these operating systems leveraging NVIDIA. At the moment however there is no automatic way to install the Horizon View Linux Agent on the desktops out of the box, so you have to automate it if you want to use this at scale. After I had built my first Linux desktop and started using it I liked it so much I wanted to provision more desktops, and see how fast I could do it. This article will take you through some of the steps and the scripts I used to provision 400 Linux desktops from template and register them as managed machines with Horizion View in just 11 minutes.

First before we look at the scripts there are some pre-requisites. You’ll need to have a working VMware Horizon View 6.1 environment, VMware PowerCLI installed (latest version recommended), a Linux desktop based on one of the supported distributions, the VMware Horizon View Agent for Linux, and a Linux customization specification created in vCenter.

In order for the new desktops that you provision to be able to automatically register with the Horizon View Connection Server you need to place some first boot scripts into the template of your Linux Desktop, and you need to ensure that the Horizon View Agent is installed on the template VM. In the scripts below the directory for the Horizon View Agent is /root/VMware-viewagent-linux-x86_64-6.1.1-2772438.

In the template, which I have called LinuxDT-Template, the first script is placed in /etc/rc.d. This script is called at startup and will check for the existence of the first boot script, and if it exists will be executed.

/etc/rc.d/rc.local:

#!/bin/sh
#
# This script will be executed *after* all the other init scripts.
# You can put your own initialization stuff in here if you don't
# want to do the full Sys V style init stuff.

touch /var/lock/subsys/local
FBSCRIPT="/root/firstbootview.sh"
if [ -f $FBSCRIPT ]; then
echo "Firstboot script exists, executing..."
$FBSCRIPT
fi

The next script is the first boot script itself. This will check to see if the Horizon View Agent is already installed and if so will clean that out. It will also check that the script is not being run on template itself. This script will install and configure the Horizon View Agent on the Linux Desktop and register the desktop with Horizon View. It checks to make sure the hostname of the VM matches the DNS records, else the registration process will not be successful.

/root/firstbootview.sh

#!/bin/bash
VIEWAGENTDIR="/etc/vmware"
VIEWAGENTINSTALL="/root/VMware-viewagent-linux-x86_64-6.1.1-2772438"
VIEWMANAGER="nxviewmgr.nxmgmt.local"
DOMAINNAME="nxmgmt.local"
USERNAME="viewmgr"
PW="YourPass"
HOSTNAME=`hostname`
LOOKUP=`nslookup $HOSTNAME | grep $HOSTNAME | cut -c 7-`

echo "Checking if Horizon View Agent was previously installed"

if [ -d $VIEWAGENTDIR ]; then
 echo "$VIEWAGENTDIR exists, uninstaling"
 /usr/lib/vmware/viewagent/bin/uninstall_viewagent.sh >> /root/viewagentinstall.log
 rm -rf $VIEWAGENTDIR
fi

while [ "$LOOKUP" != "$HOSTNAME.$DOMAINNAME" ]; do
 sleep 5
 LOOKUP=`nslookup $HOSTNAME | grep $HOSTNAME | cut -c 7-`
done

echo "Installing VMware Horizon View Agent for Linux..."
cd $VIEWAGENTINSTALL
echo "Working directory" `pwd`
$VIEWAGENTINSTALL/install_viewagent.sh -b $VIEWMANAGER -d $DOMAINNAME -u $USERNAME -p $PW >>/root/viewagentinstall.log
cat /root/viewagentinstall.log

if [ "$HOSTNAME" != "LinuxDT-Template" ]; then
 echo "Firstboot script complete, self destructing in 5 seconds..."
 sleep 5
 echo "Deleting $0 ..."
 rm -rf $0
 reboot
fi

Now that you have the Linux Desktop Template created and the first boot scripts set up, along with the other pre-requisites, you can proceed to deploying as many Linux VDI desktops as you wish. Each will be registered with Horizon View, and then you will be able to add them to a desktop pool. The sample PowerCLI script below will rapidly provision as many desktops as you like. It has been written to be multi-threaded, and will use all of the CPU on the client system it is run from. Also because it runs all clone tasks in parallel, it will also place a fairly high load on vCenter. So it is recommended you test this first in a non-production environment and decide if it’s safe to use in production before you go provisioning all of the desktops on any existing systems. As with all the scripts in this post, use them at your own risk, they are provided for informational and educational purposes only.

Clone-LinuxDT.ps1:

# PowerCLI to create VMs from existing vSphere VM
# Version 1.1
# Original Author: Magnus Andersson RTS
# Updated by: Michael Webster
#
# Specify vCenter Server, vCenter Server username and vCenter Server user password
$vCenter="YourvCenter"
$vCenterUser="YourUsername"
$vCenterUserPassword="YourPassword"
#
# Specify number of VMs you want to create
$vm_count = 400
#
#Specify where to start the VM Numbering Sequence
$start_count = 11
#
# Specify the VM you want to clone
$clone = "LinuxDT-Template"
#
# Specify the Customization Specification to use
$customspecification = "Linux"
#
# Specify the datastore or datastore cluster placement
$ds = "Erebor"
#
# Specify vCenter Server Virtual Machine & Templates folder
$Folder = "NXPERFENG"
#
# Specify the vSphere Cluster
$Cluster = "NXNZTest2-3000"
#
# Specify the VM name to the left of the ñ sign
$VM_prefix = "LinuxDT"
#
# End of user input parameters
#_______________________________________________________
#
write-host "Connecting to vCenter Server $vCenter" -foreground green
Connect-viserver $vCenter -user $vCenterUser -password $vCenterUserPassword -WarningAction 0
$start_count..(($vm_count+$start_count)-1) | foreach {
$y="{0:D2}" -f $_
$VM_name= $VM_prefix + $y
$ToD = (Get-Date).ToLongTimeString()
Write-Host "Script Stated Executing at" $ToD
$ESXi=Get-Cluster $Cluster | Get-VMHost -state connected | Get-Random
write-host "Creation of VM $VM_name initiated" -foreground green start-process powershell -NoNewWindow -ArgumentList "Add-PSSnapin VM*; Connect-viserver $vCenter -user $vCenterUser -password $vCenterUserPassword -WarningAction 0 | out-null; New-VM -Name $VM_Name -VM $clone -VMHost $ESXi -Datastore $ds -Location $Folder -OSCustomizationSpec $customspecification | out-null; Start-VM $VM_Name | out-null; write-host 'Creation of VM $VM_name initiated' (GetDate).ToLongTimeString() -foreground white;"
}

I used the script above to successfully provision 400 Linux VDI desktops in 11 minutes. In order to do this you not only need a decent client system to run the script (I used a Windows VDI Desktop running on a Nutanix cluster), you also need a very good back end infrastructure to stand up all of the cloned desktops. They will all be provisioned, customized and booted essentially in parallel. There will be a lot of IO and a lot of CPU usage across the system. It would be normal for this to tax the CPU’s on the hosts to 100% utilization for a sustained period.

Part of the reason I was able to do all of this in 11 minutes was because I was using a Nutanix 4 node NX3000 cluster (albeit a 3 year old cluster) and ESXi 6.0. So on a newer more modern cluster I probably would have been able to do the same task in less than 10 minutes. Remember, this is provisioning all desktops from template, customizing them, powering them on, and registering them with VMware Horizon View, for 400 desktops, in 11 minutes. If I doubled the number of desktops and doubled the number of Nutanix hosts in the cluster (to say 8), the provisioning time wouldn’t take twice as long, provided I have enough horsepower on vCenter and on the system executing the clone tasks, it would take the same amount of time. This is because the Nutanix platform scales linearly. Once you know how many VM’s you get per node, every node you add, you get exactly the same, consistent, predictable performance.

Here is a copy of the LinuxVMCustomizationSpec.xml file you can modify and import into vCenter. This is pretty simple. Uses DHCP for IP addressing. It has been configured to use the Google DNS Servers. So you will need to change that to your DNS servers.

LinuxVMCustomizationSpec.xml

<ConfigRoot>
 <_type>vim.CustomizationSpecItem</_type>
 <info>
 <_type>vim.CustomizationSpecInfo</_type>
 <changeVersion>1414648162</changeVersion>
 <description/>
 <lastUpdateTime>2014-10-30T05:49:22Z</lastUpdateTime>
 <name>LinuxVM</name>
 <type>Linux</type>
 </info>
 <spec>
 <_type>vim.vm.customization.Specification</_type>
 <globalIPSettings>
 <_type>vim.vm.customization.GlobalIPSettings</_type>
 <dnsServerList>
 <_length>2</_length>
 <_type>string[]</_type>
 <e id="0">8.8.8.8</e>
 <e id="1">8.8.1.1</e>
 </dnsServerList>
 <dnsSuffixList>
 <_length>1</_length>
 <_type>string[]</_type>
 <e id="0">nxmgmt.local</e>
 </dnsSuffixList>
 </globalIPSettings>
 <identity>
 <_type>vim.vm.customization.LinuxPrep</_type>
 <domain>nxmgmt.local</domain>
 <hostName>
 <_type>vim.vm.customization.VirtualMachineNameGenerator</_type>
 </hostName>
 <hwClockUTC>true</hwClockUTC>
 <timeZone>Pacific/Auckland</timeZone>
 </identity>
 <nicSettingMap>
 <_length>1</_length>
 <_type>vim.vm.customization.AdapterMapping[]</_type>
 <e id="0">
 <_type>vim.vm.customization.AdapterMapping</_type>
 <adapter>
 <_type>vim.vm.customization.IPSettings</_type>
 <ip>
 <_type>vim.vm.customization.DhcpIpGenerator</_type>
 </ip>
 </adapter>
 </e>
 </nicSettingMap>
 <options>
 <_type>vim.vm.customization.LinuxOptions</_type>
 </options>
 </spec>
</ConfigRoot>

Now if you were running these scripts in a test environment and planning to run the tests over and over again you will quickly realize that Horizon View allows duplicate entries in it’s database. You’ll also quickly realize that there is no easy way to quickly remove lots and lots of registered machines from Horizon View. You’ll also notice, if you research it, that there is currently no way to do anything about either of these problems with the Horizon View PowerCLI CMDLets. So where does that leave us? Well I’m glad you asked. In this case you will have to edit the Horizon View LDAP directory directly to remove the machines after they have been deleted. You can do this also when you decommission large numbers of VDI VM’s. But it is certainly advisable to have a backup before doing anything like this. The PowerShell script below uses LDIFDE to interrogate the Horizon View LDAP Directory and search for any machines that are showing as Agent Unreachable. This means they are off or not longer can be contacted (probably as you’ve deleted them). Any agents that fall into that category will be removed form the Horizon View LDAP Directory and therefore be removed from the Registered Machines in View Manager.

RemoveUnreachableDesktops.ps1

$desktopcn = @()

$removedesktops=(Get-DesktopPhysicalMachine -State "Agent unreachable").displayName
$removedesktops | foreach {
   write-host "Searching for dn of $_"
   ldifde -s localhost -d "dc=vdi,dc=vmware,dc=int" -r "(description=$_)" -l dn -f "$_.txt"
   write-host "Writing output to $_.txt"
   $getcn = (Get-Content "$_.txt")[0]
   $desktopcn += "$getcn"
   Remove-Item "$_.txt"
   write-host "Desktop DN's to Remove = $getdn"
}

$desktopcn | foreach {
   $output = @()
   write-host "Adding items to remove file..."
   write-host "$_"
   write-host "changetype: delete"
   $output += "$_"
   $output += "changetype: delete`n"
   $output | Out-File -Filepath removedesktops.txt -Append
   write-output `n | Out-File -Filepath removedesktops.txt -Append
}

write-host "About to remove all of the desktops..."
pause
ldifde -i -f removedesktops.txt -s localhost
Remove-Item removedesktops.txt

I’m no scripting expert, but I’ve used all of these scripts above in my lab environment. They could have probably been written more efficiently and more elegantly. But I just wrote them in a brief period of time to do a job and that’s it. So please feel free to improve upon them and if you do I’d greatly appreciate it if you could post any updates here.

Final Word

The combination of VMware Horizon View, VMware vSphere 6.0 and Nutanix web-scale converged infrastructure is an unparalleled combination for running any desktop workload (and almost all non-desktop workload as well). In this example I have used the new functionality in Horizon View 6.1 to provision Linux VDI Desktops rapidly. Due to the smart cloning features of the Nutanix platform the clones didn’t consume any additional storage space. This was not dedupe, this was data avoidance. No point creating unnecessary data in the first place. This is a demonstration of what you can do with a smart modern infrastructure platform to move your business forward, do non-disruptive upgrades in your lunch time, and with self healing capabilities, spend more time with your friends and family. If you like the sound of that, get in touch with any of the Nutanix team and ask us about it.
—

This post first appeared on the Long White Virtual Clouds blog at longwhiteclouds.com. By Michael Webster +. Copyright © 2012 – 2015 – IT Solutions 2000 Ltd and Michael Webster +. All rights reserved. Not to be reproduced for commercial purposes without written permission.

One of the little known and infrequently used features of vSphere since version 4.1 is the ability to connect a USB device to an ESXi host and then mount that device to a VM, and still allow vMotion to work without any problems. This is usually used for USB dongles required for software licensing, but can be used with a number of other devices. More often these days the USB connectivity is being used from a VMware Horizon View client to connect a USB device to a desktop. But if you’re 9000 miles away from your desktop, and you’ve been asked to connect a console cable to a physical network switch in the same location and run some debug commands, how can you do that? Well I figured it out and it makes for a good story.

I’ve been having a few weird things go on in my network with my home lab and from time to time the engineers I’m working with on this problem have asked me to run debug commands, which can only be run via the console to the switch. Usually this is no drama as I’m sitting right next to the switch and I have my console cable handy. However this week I’m in Boston, MA, and my lab environment is in Auckland, New Zealand (282ms RTT). So connecting up the console cable presenting a bit of a problem from 9000 miles away. Especially when I usually connect the console into my laptop, and in this case I have it with me. At least my VMware Horizon View virtual desktop was usable, almost to the point of not noticing the distance at all, even when using graphics intensive applications.

I figured that if I could get the console cable connected to the switch, I could connect it to a USB port on the back of one of my Nutanix ESXi hosts using a RS232 Serial to USB converter. From there I should be able to connect the virtual serial port to one of my virtual desktops. My first problem was getting all this connected.

Fortunately my 7yr old son Sebastian is pretty clever and often watches me as I pull my systems apart and put them back together. So I decided I would call him on FaceTime and walk him through the process of connecting up the console cable to the switch, and then into the back of one of my ESXi hosts. This was relatively easy, and using the FaceTime call I was able to verify that it had connected properly. As a reward I’ve bought Sebastian a T-Shirt from Cheers in Boston. I’m sure in the future his rates will go up. When I’m not around the house Sebastian takes the role of home tech support.

Next challenge was to find out which host the USB device was connected into, and then to connect it into my VDI desktop. After a quick search I found that using the command lsusb would tell me what devices were connected into my hosts. To save a bit of time I used a for loop to check all my hosts for the devices as follows (only works on Nutanix from one of the CVM’s):

for i in `hostips`; do echo “Host: $i”; ssh root@$i “lsusb”; done

That was easy. Identified which host the USB device is connected to. Now the fun part. Getting it connected to my VDI desktop, while I’m running on the VDI desktop. This is actually very easy. To start with all I had to do was vMotion my VDI desktop (while I’m accessing it) over to the host with the device connected. After that it’s just a matter of installing a USB controller on my desktop (in this case Windows 8.1), which is a simple hot add operation. I had to use the EHCI controller as xHCI didn’t work for this particular device. Within a couple of seconds Windows had detected and installed the correct driver for the USB controller.

Next step, connect the USB device. This turned out to be a little harder. I connected the device and it popped up in Windows, but it couldn’t find the right device driver. After a bit of surfing around I found the right driver and got it installed. Wala, I had a new USB virtual serial port inside my Windows VDI desktop, and could not connect that into Putty to use to access the switch. All of this was done line, using hot add features of vSphere and Windows, and all while I was accessing and vMotioning around the desktop I was connected to. This is how it looked in the virtual machine configuration once I had it completed:

USB Device support is different across the different versions of vSphere and dependant on whether you want to connect a local device directly connected to the ESXi host or via a client, including a View desktop. More information on USB device connections to ESXi and what is supported can be found in VMware KB 1022290.

Final Word

I was successfully able to get onto the console of my switch via Putty using this USB virtual serial port, connected into my Windows 8.1 VDI desktop, via VMware Horizion View from Boston, MA, to Auckland, NZ over a distance of 9000 miles (26 hr flight with layover time included). All without leaving my hotel room. I could collect the data and upload it to the engineers that are working on the problem with me. All while online using the same virtual desktop, including vMotioning it around my hosts, and all without losing connection to the USB device. A big thanks to my 7yr old son Sebastian, as I would not have been able to do this without his help.

—

This post first appeared on the Long White Virtual Clouds blog at longwhiteclouds.com. By Michael Webster +. Copyright © 2012 – 2014 – IT Solutions 2000 Ltd and Michael Webster +. All rights reserved. Not to be reproduced for commercial purposes without written permission.

I wanted to share a brief story about an excellent VDI experience I had recently with VMware Horizon View while on a business trip to the USA. I think this story demonstrates the power of virtualization, modern mobility technology, and the productivity that really can break down barriers and provide massive business value. Before we start the story you might like to check out my home lab environment. This is where my VMware View virtual desktop is located (Windows 8.1). My environment is connected via a home VDSL service to the internet and uses dyndns so I can access it when I’m away from home. This environment is located in Auckland, New Zealand. This story started while on a journey of over 8000 miles and starting at 38,000 feet above the US, somewhere between San Francisco and New York.

On this trip I flew Air New Zealand (@flyairnz), which is my usual airline and part of the Star Alliance network. When I get to San Francisco they connected with United (@united), which is also part of Star Alliance, for my connecting flight to New York (Newark International – EWR). Fortunately for me there was satellite based WIFI service available on my United flight. This allowed me to get a lot of work done on the 6 hours or so it took to get to my destination. But that isn’t the main reason for the story.

I don’t remember exactly how far it was into the flight, but I got an email from the Nutanix (@nutanix) engineers I was working with at HQ in San Jose, California on some new features of an upcoming release. We had started working through some things before I left Auckland on my 12 hour flight to the US, while I was at the Airport lounge. In the email they wanted to set up a webex so that we could continue working on my lab system, which as you know is back in Auckland, and only connected by VDSL. To put this in perspective, on a good day, the latency to San Francisco or San Jose is 198ms round trip from my home lab / home office. As you may know, your virtual desktop experience can be limited by high latency if your display protocol isn’t great at handling the latency.

I have previously run video demonstrations using my cell phone as a VMware View client, using the PC Over IP protocol, while in Singapore, back to my home lab datacenter in Auckland. I’m not claiming it was even 20 frames per second, there is some noticeable jitter, but this is a 3G cell phone connection over thousands of miles back to a desktop connected at the end of a VDSL line at the bottom of the world. It was usable and you could tell what was going on, and with sound that was for the most part in sync with the video. If that works, I thought to myself, maybe I could do something similar while on board my flight from San Francisco to New York.

While on board the United flight the WIFI is connect to satellite. So the signal has to bounce up to the orbiting satellite, probably 22,000 miles or so above the earths surface, before returning to a ground station where it connects to the rest of the internet. It would then have to travel the approximately 8,000 miles back to my office in Auckland. So we’re probably talking a round trip in excess of 30,000 miles and probably a good part of almost a second in round trip latency. I didn’t bother to do a speedtest while on the plane, perhaps I should have, it would have made things more interesting. This was going to be a lot higher latency than my 3G cell connection from Singapore to New Zealand. Would it even be usable? Obviously it was, else I wouldn’t be writing this.

I logged into my VMware View desktop while on the plane. From my desktop in Auckland I set up a Webex session back to the engineers in HQ in San Jose. I was impressed by how usable it was and although there was a bit of a lag, I could still type quite well even into SSH sessions, review PDF documents and word documents. This was important as we were using SSH as the means to access a number of my lab systems, and were referring to some PDF’s. I was able to give the engineers in San Jose control of my keyboard and mouse and watch them in pretty much real time. We were able to be as productive using this VMware View connection from an air plane as if I was sitting at my laptop in my home office.

At one point I tried using YouTube to watch a recent movie trailer, just because I could. It was usable and I could tell what was going on, but it was only a frame or two per second. Definitely not an HD experience (although it was an HD feed), but considering the distance, latency and limitations, it was perfectly fine. I wouldn’t recommend watching videos in this way, but using productivity apps, emails, using Webex, viewing a presentation, using SSH, managing your environment using vCenter and the like, is perfectly doable.

If you have WIFI on a plane you can be just about as productive as if you were still in the office, provided you don’t have the person in front putting their seat all the way down so you can’t work on your laptop. One of the joys of flying economy only as a company policy. Fortunately this isn’t something that happened to me on this trip. By the time I had landed in Newark we had got everything that we’d needed form the Webex, and it was one less thing for me to worry about.

Ultimately the VMware Horizon View PCoIP protocol really made this possible. If the protocol didn’t do a good job of handling low bandwidth high latency links, then I would have been stuck. Fortunately, even with the extreme latency, I was able to get done what I needed to do, and saved a lot of time in the process.

Final Word

If this is what can be delivered by technology today, imagine what we will be able to do in the future. Imagine the power that this can deliver to business all over the world. An additional few hours of productivity on a flight could add up to a lot of value for a business. Of course the benefits of virtual desktops are not just available to big businesses, but also small business. Anyone can afford to set up a virtual desktop infrastructure. You can start small and grow to whatever scale you need to. If you choose to do it with Nutanix and need anything from 150 virtual desktops to over 200,000 virtual desktops, then Nutanix will guarantee the performance and the service levels if you go with the VDI Assurance Program. I would encourage everyone to give VDI a try. But even if you don’t want full VDI desktops, now with VMware Horizon View 6.0 you can do application presentation. It would definitely be worth taking a look. As always comments and feedback are welcome.

—

This post appeared on the Long White Virtual Clouds blog at longwhiteclouds.com, by Michael Webster +. Copyright © 2014 – IT Solutions 2000 Ltd and Michael Webster +. All rights reserved. Not to be reproduced for commercial purposes without written permission.

In a previous article, Nutanix Disruption as a Service Serves Up VDI Assurance World First, I wrote about the Nutanix VDI Assurance program that allows customers to pay for VDI Infrastructure on a per desktop basis (perpetual or term based), based on certain pre-defined user profiles, and guarantees the performance and service levels (takes away the risk). It included in that article an architecture diagram that would allow up to 10,000 Power User desktops, or 20,000 Task Workers. But like a lot of people I always like to do more, and I got some questions about how I’d scale up to even larger numbers. This article will answer those questions. Below I present an architecture that is initially sized for 20K Power Users or 40K Task Workers, but can scale to 200K+ Power Users, just by adding in more of the standard building block components. I’ve also included a diagram of how this might look logically in a multi-site scenario.

The diagrams below assume the standard Nutanix definition of a Power User or Task Worker. The platform is based on the Nutanix VDI Assurance Model NX-3060 node type (256GB RAM). Each rack is a modular unit or building block and you can just keep adding racks. This is enabled by the underlying network architecture being a leaf-spine design, and the web-scale architecture and linearly scalability of the Nutanix Virtual Computing Platform. Information on leaf-spine network architecture can be found from Cisco and Arista and others.

The Nutanix Virtual Computing platform delivers Power Users desktops at <6w per desktop, and Task Workers <3w per desktop at any scale. This is much more power efficient than competing solutions. The density of 2.5K Power Users or 5K Task Workers per rack includes resources for N+1 resiliency per VDI cluster. When looking at the design, remember that this includes all server compute, storage, and networking components for the entire solution. There are no separate racks for storage or network equipment (except the uplinks to the WAN, which are not shown).

What makes the VDI Assurance model from Nutanix so simple is that you don’t need to worry about this detail. Nutanix takes care of it for you. I just drew this pretty picture to get you interested. You simply have to know the number and profile of the users you need to host and Nutanix will give you the right infrastructure to run them with guaranteed service levels. If it’s not performing then Nutanix will fix, which may mean deploying more hardware, at no additional cost. You pay per desktop in packs of desktops for a perpetual or for a term (1/3/5 yrs). That’s it. Uncompromisingly simple.

Note these diagrams are of my creation (based on previous example diagrams and good work by Steven Poitras – http://stevenpoitras.com/, author of the Nutanix Bible) and your actual deployment may be different to this based on Nutanix VDI Assurance model. This is for informational purposes only, but I did spend a lot of time calculating the numbers to make sure the design would work. This is still one component and an over simplification. These designs could be deployed with your favourite VDI solution, so you can choose between either VMware Horizon View or Citrix XenDesktop.

VDI for 10K Power Users, 20K Task Workers

This is the design diagram from my previous article.

VDI for 20K to 200K+ Power Users, 40K to 400K+ Task Workers

If you click on the image it will be expanded and easier to read. This can easily be expanded to 72 racks by adding line cards to the spine switches using the MLAG approach. Using ECMP (for above 72 racks) you could add an additional 36 racks, for a total of 108 racks without any significant modifications to the design or it’s building blocks.  You could keep going, but that would require some modifications to the design. Although I have not tested this design in the real world, I have calculated all the various components of the solution based on the Nutanix VDI Assurance model for VDI user profiles and nodes. With the Nutanix VDI Assurance model though you really just have to know how many you need, and then Nutanix will do the rest.

This is by no means the only way to achieve this result. This is just one of many possible architecture options.  But I think this shows the power of the Nutanix Virtual Computing Platform to deliver on the VDI use case while consuming an efficient power footprint and a very efficient datacenter footprint. There is more to a VDI design than a single diagram, but this should get the creative juices and imagination flowing.

Taking VDI to Mult-Site for DR and Even More Scale

So how do you provide DR and multi-site resilience and scale to a VDI design based on the Nutanix Virtual Computing Platform? Good question. Because VDI is a business critical application, especially when deployed at scale, you need to make sure you have headroom to handle failure and disaster scenarios. Here is a logical diagram of how this might look. You can click on the image to make it bigger.

Because the Nutanix Virtual Computing Platform has data replication built in, you can easily protect your master images and replicated them to one or more sites. You can quickly clone images using the Nutanix VAAI integration, and you can quickly deploy, refresh and power on desktops. Data Deduplication, shadow clones, and the unique data locality features of the Nutanix platform make sure your desktops always receive optimal performance. With the Nutanix solution you get web-scale, cloud economics, high performance, simplicity and choice, at your place, and at your pace.

Final Word

The above solution delivers power consumption for power user desktops for < 5W per desktop and < 2.5W for task workers. Take a look at the Nutanix VDI Assurance program, go through the Nutanix Product Info and Tech Papers, and check out the Case Studies. As always your feedback and comments are appreciated. Let me know your thoughts on scalable, large scale VDI design. If you want to see how the Nutanix solution compares to other reference architectures check out Battle RA Royale: More VDI For Less Moolah.

—

This post appeared on the Long White Virtual Clouds blog at longwhiteclouds.com, by Michael Webster +. Copyright © 2014 – IT Solutions 2000 Ltd and Michael Webster +. All rights reserved. Not to be reproduced for commercial purposes without written permission.

I normally bring you information about business critical apps such as Oracle, SAP, Java, SQL Server and the like, and the infrastructure configurations and performance that support them. But if you look at it any Virtual Desktop Infrastructure that is widely deployed to a larger percentage of an organisation, it is a business critical app. I’ve been involved in architecting and troubleshooting a few VDI environments over the last few years, one as big as 6,000 seats using Horizon View, and if we go back a bi,t a much larger environment using Citrix XenApp. I’ve seen many of the challenges, and I’ve seen what it takes to succeed. This is why today’s announcement from Nutanix is exciting, in that it brings all the benefits of VDI to organisations of any size, without all the risk and complexity that might have come along with it. Let’s take a look at the announcement and what an environment and infrastructure to support 10,000 Power Users’ VDI desktops might look like on Nutanix.

Nutanix today announced a major new initiative, coming hot on the heals of the recent patent announcement, that is set to disrupt the VDI market and make it uncompromisingly simple for any organisation to benefit from VDI. In a world first Nutanix launched its VDI Assurance program that takes the guess work and risk out of running a VDI environment of any size. Starting from as little as 150 VDI desktops you can scale up to any number, 10,000+. Now customers have the option of paying per VDI Desktop for their own Desktop as a Service infrastructure on premises, rather than being forced to go to a cloud provider. Purchase options don’t just include perpetual ownership, but also term based ownership of 1, 3 or 5 years. Nutanix termed this “the industry’s most comprehensive VDI solution” in their press release. The Nutanix offering is available for both VMware Horizon View and Citrix XenDesktop environments.

Traditionally VDI projects have struggled to succeed, due to the complexity and cost of legacy solutions, when trying to deliver an acceptable and high performance user experience. By Nutanix taking away the risk and guaranteeing that sufficient hardware will be implemented to meet the end user experience, anyone can implement a VDI solution with confidence. All a customer has to do is select the number and type of user profile they wish to support and Nutanix will take care of the infrastructure at a consistent and predictable price per desktop. This brings the transparency and benefits of Desktop as a Service Cloud offerings to Customers to run on premises. Disrupting and changing for the better traditional VDI business models, not just technology. To find out more about how you can deploy VDI with assurance in your environment in just three easy steps visit the Nutanix per Desktop VDI and Assurance page.

So what might an environment for 10,000 VDI Power Users look like? The image below has an example of what it might look like. This is based on the Power User Profile Nutanix described in the  Nutanix per Desktop VDI and Assurance program and includes all compute, storage, network and management infrastructure to run the complete solution.

As you can see from the above diagram you can achieve 10,000 Power Users in 4 racks, with the small management infrastructure footprint in a 5th rack and plenty of room for expansion. But this isn’t the limit by any means. Due to the way that Nutanix infrastructure linearly scales out you can keep growing the environment to meet whatever your business needs are.  You can start from as little as 150 VDI users and scale up as much as you like without reducing performance, or running out of capacity, and all without changing your architecture or design significantly. Now this really is uncompromisingly simple VDI. Note: This diagram is my creation and the assumptions I used to create it may be different to the Nutanix VDI Assurance Program architecture, your architecture and requirements might be different, this is an example and provided for informational purposes only without any warranty of any kind.

Final Word

The assumptions I used to create the diagram above is a Power User Desktop each consisting of 2 vCPU, 4GB RAM, 60 IOPS and 80GB raw storage (specifications according to the Nutanix VDI Assurance program). This works out to 56 persistent desktops per VDI host (23 hosts) plus 1 host failover and maintenance, 1288 desktops per VDI Cluster, 2576 desktops per rack.  I based the diagram on the Nutanix 3060 platform with an N+1 cluster resiliency architecture with 24 nodes per VDI cluster and 48 nodes per Nutanix cluster. There are 2 VDI Clusters per rack and one Nutanix cluster per rack. There is a vCenter Server managing each rack, which is located in the management cluster. The networking assumes a leaf-spine architecture with L2 leaf top of rack switches per rack connected to each of the L3 spine switches, two in this example.

—

This post appeared on the Long White Virtual Clouds blog at longwhiteclouds.com, by Michael Webster +. Copyright © 2014 – IT Solutions 2000 Ltd and Michael Webster +. All rights reserved. Not to be reproduced for commercial purposes without written permission.

Today Nutanix (@nutanix) delivered major hardware updates to it’s radically simple Google like infrastructure platform for the masses. The updates included integration with new GPU and Teradici APEX encoding offload cards for the most graphic intensive desktops, while at the same time providing increased VM density and lower TCO across a new range of hardware options. This announcement came just in time for the VMware Horizon View 5.3 GA release, which was also today. The new Nutanix 7110 platform breaks the final barrier to delivering all applications to all virtual desktop users and powers video and graphics-rich applications with workstation-level performance. This is all achieved while maintaing simplicity customers have come to know and love, and without losing any availability and manageability normally associated with dedicated hardware required for workstation class CAD/CAM and 3D desktops and other high end graphics use cases. The Nutanix 7110 platform and VMware Horizon View 5.3 are a powerful combination for All Virtual Desktops, but wait there’s more.

From the Nutanix Press Release – “The NX-7110 integrates NVIDIA GRID and Teradici PCoIP technologies enabling users to work with the same graphics applications that they use every day, but via a desktop that is delivered virtually. Combining NVIDIA’s GRID and the Teradici PCoIP Hardware Accelerator solutions with the scalability and performance of Nutanix’s virtual computing technology enables enterprises to tackle virtual desktop deployments of unprecedented size and scope.” Full Nutanix Press Release.

When combined with VMware View 5.3 and leveraging the new virtual Shared Graphics Acceleration (vSGA) feature (and vSphere 5.5) high performance graphics intensive workloads are no longer tied to a physical server and can migrate freely while still achieving their requirements for GPU intensive tasks. For workloads that need passthrough (and don’t require vMotion) virtual Dedicated Graphics Acceleration (vDGA) is now fully supported. This greatly improves the manageability, maintainability and availability of your organizations critical desktop infrastructure. Importantly existing Nutanix environments can dynamically deploy NX-7110 appliances into a unified cluster that is centrally managed, while maintaining graphics intensive users in a separate desktop pool. With View 5.3’s full support for the View Composer API for Array Integration (VCAI), which was previously a tech preview, deploying and recomposing desktops gets even faster.

In addition to the features of VMware Horizon View 5.3 mentioned above the release adds support for Windows 2008 R2 to be used as a desktop, which is described in KB 2057605, and Windows 8.1. Windows 2008 R2 is a great option if you are a service provider and want to offer desktop as a service to end users in a multi-tenanted environment, primarily because it gets arounds restrictions in the Microsoft SPLA licensing. Full VMware Horizon View 5.3 Release Notes.

Not only did Nutanix release a great distributed Google like infrastructure platform for graphics intensive workloads they also launched updates to the NX-3000 and NX-6000 series of systems that boosts VM density and enhances performance.  The new platforms now include Intel Ivy-bridge processors, more cores, higher clock speed options, more RAM and more importantly higher VM density in the same amount of space and power footprint. Further information about the launch including links to solutions briefs and white papers can be found here.

Final Word

Virtualized GPU’s are not just the realm of high performance graphics intensive virtual desktops, they are also greatly beneficial and supported for use with Linux Guest VM’s as of vSphere 5.5. This should not be confused with Linux Virtual Desktops, which are not supported, but HPC server workloads. This opens up some incredibly good use cases for high performance computing (HPC) clusters, that can benefit from GPU’s to assist with embarrassingly parallel operations.  So not only is the NX-7110 platform a great virtual desktop platform, because of the unique mix of high performance compute, graphics power and local storage access combined with a distributed architecture, availability and simplicity it could also be a great platform for large scale HPC environments for research facilities, institutions and universities. Just like the one I wrote about in my article Virtual Beats Physical for HPC Monte-Carlo Grid Performance. I bet you hadn’t considered that. As always your feedback and comments are welcome.

—

This post first appeared on the Long White Virtual Clouds blog at longwhiteclouds.com, by Michael Webster +. Copyright © 2013 – IT Solutions 2000 Ltd and Michael Webster +. All rights reserved. Not to be reproduced for commercial purposes without written permission.

The earthquakes yesterday in Wellington, the capital city of New Zealand, brings back vivid memories of the devastating earthquakes of Christchurch and Japan a couple of years ago, which both countries are still recovering from. The earthquakes in Wellington were quite severe, with the biggest quake recorded at 6.5 with others of 5.7, 5.8 and over a hundred aftershocks. Luckily there was no loss of life in the Wellington quakes, but there was a bit of damage and the CBD was shut down for a day while assessments were completed on the buildings and infrastructure. I spoke to a few friends and customers after the quakes to see if they were ok. Some of the customers I know where already up and running in their DR sites, or had access to their infrastructures remotely from wherever they were. What yesterday’s events highlight is the need to have your BCP and DR plans in place and tested before disaster strikes. This article will cover briefly some things you might want to consider when designing your BCP and DR Plans, that can make them easier to test, simpler to execute, more reliable and easy to audit, based on the events in Wellington.

So that we’re all on the same page I’d like to define BCP and DR. Business Continuity Planning (BCP) is the process of defining how your business will operate after a disaster or significant business impacting event, potentially  in the absence of IT. A business continuity plan provides a roadmap for continuing operations under adverse conditions. Disaster Recovery is the process of preparing for the recovery or continuing operations of vital IT systems during a natural or human induced disaster. Your DR plans should be based on and support your BCP Plans. The most common disasters that will impact IT and business operations are human induced, but the quakes yesterday highlight the need to have these plans in place before the earth moves. You should plan for the worst.

Once you have your BCP and DR plans in place you need to know they will work and you need to regularly test them. In terms of making DR simple and testable one of the most important tools you can use is VMware Site Recovery Manager and also VMware’s Horizon View. The first thing to consider is that if you can’t test or haven’t tested your DR then you can’t trust it. If you can’t test it you can’t trust it. So your DR and BCP plans must be testable.

VMware Site Recovery Manager provides DR run book automation and allows for completely auditable testing of your DR plans in a way that is non-disruptive to production. This allows you to test your DR much more often and provide confidence that it will work. When disaster strikes you can’t guarantee that your IT operations team will be available, Site Recovery Manager provides a simple way for even non-technical users to recover critical business systems when the worst happens. VMware’s Horizon View allows remove access to virtual desktops from anywhere and supports soft phone integration or integration into call centre systems. So essentially you could have your employees running remotely just as if they were local to the office. Provided the systems are available in your datacenters.

Learning from Disaster

So what can we learn from the experience in Wellington yesterday? Here are some things worth considering.

Even a minor natural disaster could result in large parts of a city or geographic area being cordoned off and unaccessible for a number of days. In Wellington CBD many streets are cordoned off and unaccessible due to building damage. It’s likely that the cordons will be removed quickly, but that might not always be the case. If you don’t have a strategy to provide remote access to your systems then you may suffer a more severe business impact. This is where VMware’s Horizon View can come in. Some of the customers I spoke to in Wellington were remotely accessing their systems from virtual desktops. They had a DR office as well as access from home for staff. Many human-induced events can cause cordons to be put in place such as bomb scares, virus outbreaks, construction activities etc. Even thought the damage in Wellington was not that severe the cleanup in some of the buildings will still take days.

Although the quake itself didn’t cause widespread severe building damage broken sprinkler systems and broken water mains has caused flooding and water damage, including destroying computers in some office buildings. Broken sprinkler systems and water mains in your datacenter and in your office could destroy your systems even if the other disaster events do not. If your business is located close to the coast in an area prone to Tsunami’s then you probably don’t want your power generators or datacenter to be located below ground.

It’s what you don’t know or don’t plan for that will hurt you. If you’ve assumed physical access to a building as part of your DR process, then that won’t work if the building is cordoned off for a number of days. If you plan for the worst and use technology to simplify your recovery process then you will be able to better adapt to situations you haven’t anticipated.

You need to take a risk based approach to your BCP and DR planning. By this I mean the strategies you employ to provide for business operations need to be cost effective and justifiable and weighted to the likely business impact and probability that a particular event will occur. Plan for the likely events that might impact your particular locations. Not everyone is in an earthquake prone location, but you may be prone to other events. Make sure your organisation has sufficient business interruption insurance.

Final Word

Even minor disasters can cause a big impact and mean days of cleanup and days before you can get access back to your office. Even if there isn’t major building damage resulting from an earthquake there can be water damage and cordons that can still severely disrupt operations. If you can’t test it you can’t trust it. Have your DR plan in place and tested, you never know when you’re going to need it. VMware technology can help make the recovery process much quicker, more reliable, simple, auditable, testable.

—

This post first appeared on the Long White Virtual Clouds blog at longwhiteclouds.com, by Michael Webster +. Copyright © 2013 – IT Solutions 2000 Ltd and Michael Webster +. All rights reserved. Not to be reproduced for commercial purposes without written permission.