My demo equipment at TechEd

Hi there,

I spoke to multiple people being totally excited about my demo equipment at TechEd, and was asked couple times if I can blog this. So here we go.

The hardware I used is a Lenovo Tablet X220T – thanks Lenovo! It’s a great hardware: I love to work with tablet pc’s, and I’ve worked with it for years, however a hardware I had before died on me (was my personal one, but I used it as main device reducing my work laptop for demo and test). I always preferred Lenovos Keyboards and their solid business laptops. And the tablet is great – I like being able to work in a train, plane, wherever, having all input options (mouse+keyboard, touch or stylus) and select whether to write a long text using keyboard, reviewing, sketching out or handwriting annotations using a pen or simply touching the things I want to open or select. I want to work the way I prefer, not thinking about input but being able just to do it. I POWERED (had to say this in caps) the Lenovo with Windows 8, which rocks!! It’s also able to handle 16GB of RAM, which is great for Hyper-V in Windows 8. There’s a new version out, X230T, with USB 3, mSATA for Broadband or additional disk (cool, mSATA Disk for the OS, traditional and bigger HDD for the data and VMs in a convertible tablet form factor with 16GB), and a optional battery where they claim up to 18 hours uptime. I cannot wait to get my hands on one of these, and if it’s as satisfying as I believe it’s definitely shopping time for me. To get back to my presentation – I’ve done multiple with the same hardware and since I currently have only one internal HDD I’m using an external SATA-Drive with PCI-Express-Adapter to speed up if I need more power for my VMs (did this at The Experts Conference in April).

The new default installation of Windows Server 2012 is the Server Core option. You are also able to switch back and forth – or in between. The options are Server Core, Full Server, or in between with Server Core with Management GUI. The last option has Server Manager and the management interfaces but still lacks Explorer (Shell, Start Menu, File Browsing), Internet Explorer and many other things. It’s new to Server Core in Windows Server 2012 to decide whenever to either install or uninstall Management or full GUI. Additionally, Server Core offers now the possibility to uninstall binaries which are not needed. Back in the early days of Windows there were no unused binaries on disk, however it was always hard to struggle when you were installing a new component on an existing server because you were asked for the CD and had to insert the right language version and also the right distribution media (e.g. it depended whether it was a volume licence media, off the shelf media, MSDN or TechNet,…). With some version – IIRC it was Vista/2008 – this was changed and all the compresed binaries for all components (roles and features) were copied onto the system, even when they were not used but so that if anyone was going to install a component later (s)he wasn’t asked for the media.

Today, Virtualization and packing multiple machines on one host is critical, especially when we talk about cloud computing. So in Windows Server 2012 we are able to install or uninstall roles or features, but we are also able to uninstall the features and remove the binaries from the system, allowing us a smaller footprint of the operating system installation. However, when you remove a feature, you can still install it, but you need to ensure that the machine is able to connect to Windows Update or you need to provide the install.wim-file, installation media or an installed server to pull the binaries from.

I’ve created a base image (I always do this, then creating differential disks to create individual machines, gives higher performance with multiple machines, less disk space and easy creation of new machines).

So since it’s the new installation default, and I think that is a great way to go (reducing systems to what they are supposed to do), I used Server Core as only operating system option for my demos at TechEd. I decided to strip down the base image as much as possible, and was running a Powershell command to remove all binaries which are not used right after the installation:

(get-WindowsOptionalFeature -online) | %{ if ($_.State -eq ‘Disabled’)


    disable-WindowsOptionalFeature -FeatureName $_.FeatureName -online –remove


I needed the management tools somewhere, so I put the Remote Server Administration Tools (RSAT) for Windows 8 on the host operating system. There is some configuration needed when you remotely want to install a Server Core as first domain controller, since the client and the server are obviously not on the same domain. However, you can do this (enable remote management on the server, configure the client to trust the server using Windows Remote Management and HTTPS,..), for some things you’ll have to fall back to the commandline in Server Core (Server Manager allowed me to install the binaries, however was unable to promote the DC, I had to do this with dcpromo /promotes …). But I always had to right-click and configured the account used for Management. This is not the experience I want for the attendees of my session.

So I decided to join the host to the domain of the virtual machine on the same host. Risky? Not really. The default configuration will start the VM when it was started when powering down. But it’s taking a bit longer than the host, apparently. Also, cached credentials allow me to log on without a running DC. So when I was logging on to quickly, I didn’t get a kerberos ticket and was unable to access the server. But [WIN]+[L] to log Windows and then logging back on is a workaround in this case, and I made sure before my session that I was able to start Server Manager and work remotely against my machine.

Joining the machine was a bit tricky. I tried to avoid mangling DNS. On the conference net my client is getting its IP-Settings via DHCP, but trying to keep the server on DHCP was a hazzle since I needed to reconfigure the trusted hosts. So my DC needed a static IP, and I felt I want this different from the conference net. So the client was basically on two different subnets. But he needs full DNS to the DC in order to join the domain and in order to work. LMHosts and Hosts are no option, since you can’t configure SRV-Records there (what the client is looking for in an Active Directory domain). So one option was to configure the client (=host) to use only DNS-Services of the DC. But the DC was not able to forward requests, remember, it’s on a separate network. And I didn’t need internet connectivity for the DC, but for the client (since I allowed questions via Twitter in my session).

So I thought it would be cool if I’d be able to use conditional forwarding on the client. Conditional Forwarding is a DNS-Server feature introduced in Windows Server 2003, where you can configure that certain DNS namespaces are not resolved via the standard forwarder but via another specific one. Conditional Forwarding (and Stub Zones) are frequently used within companies when they have multiple DNS namespaces.

Conditional Forwarding on the Client brought me to Direct Access. In DA you are able to configure, on the client, which IP-Adresses and which DNS-namespaces should be resolved against a corporate DNS-Server instead of using Internet DNS-Services. Direct Access is much more, but I just needed this piece. So I configured the name resolution policy table to forward requests against the virtual subnet or against my virtual DNS-namespace to the virtual DC, and everything worked like a charm.

I think this setup is really cool. I was able to demo almost everything without logging into the virtual machine, by just using the RSAT-Tools from the host. The host was able to connect to the internet and to the virtual world and knew where to go with every request. I was able to receive twitter questions right on stage and answered them in the session (and also online after the session later). And with the Windows 8 tablet, I was able to highlight areas using the pen, using touch to advance slides or to bring in the twitter application on a split screen – twitter to the right and presentation to the left – without leaving the current topic but also showing attendees what questions got in and that we are really answering them on stage. Switching to the demo consoles was also easy doing touch. And keyboard/mouse for demoing the server and typing in commands in Powershell or CMD.

It was a great success at TechEd US, and I will repeat the same setup and strategies at TechEd Europe in about a week.


Leave a Reply