Thursday, July 2, 2015

Cloud Consistency with Azure Resource Manager - Finally available!

I am very glad to announce the following:

1)      I am renewed as a Cloud & Datacenter MVP – for the fifth time!
2)      As a courtesy, we are releasing our newest whitepaper “Cloud Consistency with Azure Resource Manager

I have been doing a lot of engineering the last year, and Azure Resource Manager is one of the technologies I consider as a game changer and let us finally be able to achieve what we have always wanted, without knowing it was this we really wanted.
An idempodent and declarative way to describe our cloud resources, regardless of location and resource type.
Together with Desired-State Configuration, this is one of my big bets as we move forward.

I really hope that you will enjoy this whitepaper while we are all waiting for Microsoft Azure Stack, that will bring the ARM capabilities on-prem.

If you want to see more of Azure Resource Manager and how to model your cloud resources and applications, I would like to invite you to System Center Universe in Basel in August, where I will be giving several deep dive sessions on the topic:

You can download the whitepaper from TechNet Gallery by following this URL:

BTW: Here's the look on my daughters face when witnessing the capabilities of Azure Resource Manager

Thank you!

Tuesday, June 2, 2015

Announcing a new whitepaper!

Hi everyone!

It has been quite quiet here on this blog since last month, but there’s of course some reasons for that.
I would like to use this opportunity to give you a heads-up on an upcoming whitepaper that I have been working on together with a few other subject matter experts.

This blog post is not about the specific whitepaper itself, but the goal is rather to give you an explanation of why we are having this approach – putting a lot of effort into a whitepaper instead of publishing books.

I have personally been authoring books myself, and also together with other authors. The experience was interesting to say the least, and also required a lot of my time. Not just to do the research, testing and writing, but also to meet the deadlines, engage with reviewers and much more.
In short, the flexibility you have to modify – or even change the subject, is very very limited when working with books.

The limited flexibility is a showstopper in a business where drastic changes (as in new features and releases) happens at a much faster cadence than ever before.
In order to be able to adopt, learn and apply all what’s happening, – writing whitepapers seems like a better idea than doing books.

At least this is what we think. When discussing this with some of our peers, we often get questions around royalties etc. to be honest, you will never ever get rich by writing a book, unless you are writing some fiction about some magic wizard with glasses, or a girl describing her fantasies of a rich man.

So jokes aside, we do this because of the following reasons:

·         We enjoy doing it

This is not a secret at all. Of course we will spend some massive amount of time on these projects, and probably our significant others would have a grin every now and then. But we enjoy so much, that it is worth the risk and potential penalty we might get.

·         For our own learning and knowledge

Let us be honest. We dive deep into this to learn it by heart. There’s no secret that the technology we will cover will be our bread and butter, so we better know what we are doing.

·         To share it with the community

Do it once –and do it right. We spend a lot of our time in forums, conferences, etc and engage with the community. Being able to point towards a rather comprehensive guide that many can benefit from, instead of supporting 1:1 is beneficial for all of us

·         Recognition

If you do something good and useful, I can ensure you that many people – regardless whether they know you or not, will appreciate it and give you credits. We’ve heard several times from our previous whitepaper (Hybrid Cloud with NVGRE (Cloud OS) ) that it helped peers, IT-pro’s, engineers, students and CxO’s to make a real difference. This is probably worth the effort all alone.

So let me introduce you to the upcoming whitepaper that will hit the internet very shortly:

“Cloud Consistency with Azure Resource Manager”

This whitepaper will focus on cloud consistency using Azure Resource Manager in both the public cloud with Azure, as well as the private and hosted clouds with Azure Stack.

I won’t disclose more about the content, structure or the initial thoughts right now, but I encourage you to stay tuned and download it once it is available on the TechNet Gallery.

Thanks for reading!

Friday, May 8, 2015

Microsoft Azure Stack with a strong ARM

How did God manage to create the world in only 6 days?
-          He had no legacy!

With that, I would like to explain what the new Microsoft Azure Stack is all about.

As many of you already know, we have all been part of a journey over the last couple of years where Microsoft is aiming for consistency across their clouds, covering private, service provider and public.
Microsoft Azure has been the leading star and it is quite clear with a “mobile first, cloud first” strategy that they are putting all their effort into the cloud, and later, make bits and bytes available for on-prem where it make sense.
Regarding consistency, I would like to point out that we have had “Windows Azure Services for Windows Server” (v1) and “Windows Azure Pack” (v2) – that brought the tenant experience on-prem with portals and common API’s.

Let us stop there for a bit.
The API’s we got on-prem as part of the service management APIs was common to the ones we had in Azure, but they weren’t consistent nor identical.
If you’ve ever played around with the Azure Powershell module, you have probably noticed that we had different cmdlets when targeting an Azure Pack endpoint compared to Microsoft Azure.

For the portal experience, we got 2 portals. One portal for the Service Provider – where the admin could configure the underlying resource providers, create hosting plans and define settings and quotas through the admin API. These hosting plans were made available to the tenants in the tenant portal with subscriptions, where that portal – was accessing the resources through the tenant API.

The underlying resource providers were different REST APIs that could contain several different resource types. Take the VM Cloud resource provider for example, that is a combination of System Center Virtual Machine Manager and System Center Service Provider Foundation.

Let us stop here as well, and reflect of what we have just read.

1)      So far, we have had a common set of APIs between Azure Pack and Azure
2)      On-prem, we are relying on System Center in order to bring IaaS into Azure Pack

With cloud consistency in mind, it is about time to point out that to move forward, we have to get the exact same APIs on-prem as we have in Microsoft Azure.
Second, we all know that there’s no System Center components that are managing the Hyper-Scale cloud in Azure

Let us take a closer look at the architecture of Microsoft Azure Stack

Starting at the top, we can see that we have the same – consistent browser experience.
The user facing services consists of hubs, a portal shell site and  RP extensions for both admins (service provider) and tenant. This shows that we won’t have two different portals as we have in Azure Pack today, but things are differentiated through the extensions.

These components are all living on top of something called “Azure Resource Manager”, which is where all the fun and consistency for real is born.
Previously in Azure, we were accessing the Service Management API when interacting with our cloud services.
Now, this has changed and Azure Resource Manager is the new, consistent and powerful API that will be managing all the underlying resource providers, regardless of clouds.

Azure Resource Manager introduces an entirely new way of thinking about your cloud resources.
A challenge with both Azure Pack and the former Azure portal was that once we had several components that made up an application, it was really hard to manage the life-cycle of it. This has drastically changed with ARM, where we can now imagining a complex service, such as a SharePoint farm – containing many different tiers, instances, scripts, applications. With ARM, we can use a template that will create a resource group (a logical group that will let you control RBAC, life-cycle, billing etc on the entire group of resources, but you can also specify this at a lower level on the resources itself) with the resources you need to support the service.
Also, the ARM itself is idempotent, which means it has a declarative approach. You can already start to imagine how powerful this will be.

In the context of the architecture of Azure Stack as we are looking at right now, this means we can:

1)      Create an Azure Gallery Template (.json)
a.       Deploy the template to Microsoft Azure
b.      Deploy the template to Microsoft Azure Stack

It is time to take a break and put a smile on your face.

Now, let us explain the architecture a bit further.

Under the Azure Resource Manager, we will have several Core Management Resource Providers as well as Service Resource Providers.

The Core Management Resource Providers consists of Authorization – which is where all the RBAC settings and policies are living. All the services will also share the same Gallery now, instead of having separate galleries for Web, VMs etc as we have in Azure Pack today. Also, all the events, monitoring and usage related settings are living in these core management resource providers. One of the benefits here is that third parties can now plug in their resource providers and harness the existing architecture of these core RPs.

Further, we have currently Compute, Network and Storage as Service Resource Providers.

If we compare this with what we already have in Azure Pack today through our VM Cloud Resource Provider, we have all of this through a single resource provider (SCVMM/SCSPF) that basically provides us with everything we need to deliver IaaS.
I assume that you have read the entire blog post till now, and as I wrote in the beginning, there’s no System Center components that are managing Microsoft Azure today.

So why do we have 3 different resource providers in Azure Stack for compute, network and storage, when we could potentially have everything from the same RP?

In order to leverage the beauty of a cloud, we need the opportunity to have a loosely coupled infrastructure – where the resources and different units can scale separately and independent of each other.

Here’s an example of how you can take advantage of this:

1)      You want to deploy an advanced application to an Azure/Azure Stack cloud, so you create a base template containing the common artifacts, such as image, OS settings etc
2)      Further, you create a separate template for the NIC settings and the storage settings
3)      As part of the deployment, you create references and eventually some “depends-on” between these templates so that everything will be deployed within the same Azure Resource Group (that shares the same common life-cycle, billing, RBAC etc)
4)      Next, you might want to change – or eventually replace some of the components in this resource group. As an example, let us say that you put some effort into the NIC configuration. You can then delete the VM (from the Compute RP) itself, but keep the NIC (in the Network RP).

This gives us much more flexibility compared to what we are used to.


So, Microsoft is for real bringing Azure services to your datacenters now, as part of the 2016 wave that will be shipped next year. The solution is called “Microsoft Azure Stack” and won’t “require” System Center – but you can use System Center if you want for managing purposes etc., which is probably a very good idea.

It is an entirely new product for you datacenter – which is a cloud-optimized application platform, using Azure-based compute, network and storage services

In the next couple of weeks, I will write more about the underlying resource providers and also how to leverage the ARM capabilities. 

Stay tuned for more info around Azure Stack and Azure Resource Manager.

Monday, May 4, 2015

Azure Site Recovery: Generation 2 VM support

For almost a year ago, Microsoft announced the preview of a cloud service that has turned out to be the leading star when it comes to Hybrid Cloud scenarios, out of the box from Microsoft.

Microsoft Azure Site Recovery let customers extend their datacenter solutions to the cloud to ensure business continuity and availability on-demand.
The solution itself is state of the art and covers many different scenarios – and can rather be seen as their “umbrella” when it comes to availability and recovery in the cloud, as it has several different offerings in different flavors under its wings. 

Besides supporting DR protection of VMware and Physical computers (newly announced), Azure Site Recovery is considered as mandatory for organizations that need DR for their Hyper-V environments, regardless whether the cloud or a secondary location on-prem is the actual DR target.

Just recently, Microsoft announced support for protecting Generation 2 Virtual Machines to Azure.
This is fantastic good news and shows that the journey towards cloud consistency is established for sure.

Let me add some context before we look into the details.

I’ve been working with the brilliant Azure Site Recovery Product Group at Microsoft for a long time now, and I have to admit that these guys are outstanding. Not only do they ship extremely good quality of code, but they also listen to feedback. And when I say listen, they actually engage with you and really tries to understand your concern. In the end of the day, we are all on the same team, working towards the best experience and solution possible.

During TechEd in Barcelona, I was co-presenting “Microsoft Azure Site Recovery: Leveraging Azure as your Disaster Recovery Site” ( ) together with Manoj, and this is when our real discussion started.
Using Azure as the secondary site for DR scenarios makes perfect sense and many customers would like to take benefit from this as soon as possible. However, we often saw that these customers had deployed their virtual machines as Generation 2 VMs – which wasn’t suited for the Azure platform. This was a blocker and the amount of Gen2 VMs were increasing every day.

Earlier in January this year, I made a community survey around the topic and the result was very clear:

Yes – people would love to use Azure as their secondary site, if there was support of Generation 2 VMs in the Cloud.

I am glad to say that the Product Group listened and now we can start to protect workloads on Gen2 VMs too.
But, how does this work?

When you enable a VM for protection, the data is sent to an endpoint in Azure, and nothing special has happened so far.

However, the ASR service will perform a conversion in the service at the time of failover to Gen1.


Let me explain further.

In case of a disaster where you need to perform a failover to Azure, the VM(s) is converted and started as Gen1, running in Azure.
The ASR backend services used during failover has the conversion logic. At failover time, backend service reads Gen2 OS disk and convert the disk to Gen1 OS disk (hence the requirements of the OS disk in Azure).
If you need/want/have to failback to your on-prem Hyper-V environment, the VM will of course be converted back to Gen2.

For more details – check out the official blog post by one of the PM’s, Anoob Backer

Thursday, April 30, 2015

VM Checkpoints in Windows Azure Pack

Fresh from the factory, Update Rollup 6 has been released and shipped by Microsoft.

This isn’t a blog post that will point out all the bug fixes and the amazing work all of the teams has been doing, but rather point you towards a highly requested feature, that finally made its way to the tenant portal in Windows Azure Pack.

With Update Rollup 6, we now supports creation and restore of Hyper-V checkpoints on virtual machines, provided by the VM Cloud Resource Provider.

Tenants that have deployed virtual machines may now create checkpoints and restore them on their own, without any interaction from the cloud provider.

Let us have a closer look at how this actually works, how to configure it and what additional steps you might want to take as part of this implementation.

Enabling create, view and restore of virtual machine checkpoints at the Hosting Plan level

Once the UR6 is installed for WAP and the underlying resource provider, you will notice some changes in the admin portal.

First, navigate to a Hosting Plan of yours – that contains the VM Cloud Resource Provider.
When you scroll down, you can see that we have some settings related to checkpoints.

Create, view and restore virtual machine checkpoints – will let the tenants that has subscriptions based on this hosting plan, be able to perform these actions on their virtual machines.

View and restore virtual machine checkpoints – let the tenants view and restore virtual machine checkpoints, but not create them. This can for example be performed by the cloud provider on an agreed schedule.

When you enable either of these options, an update job is taking place at the plan level and communicates the changes back to VMM, ensuring that the tenants will have permissions to take these actions in the tenant portal once it has completed.

If we switch over to the tenant portal, we can see that when we drill into one of the existing VMs (click on the VMàDashboard) we have some new actions available.

If you would manage checkpoints for your VM Roles, you can of course do that too, but you then have to drill into each specific instance, as the VM role potentially can have multiple instances when supporting scale-out.

To create a new checkpoint, simply click on Checkpoint and type the name of the checkpoint and eventually a description.

If we switch back to the fabric and VMM, we can see that a VMM job has completed with details about the checkpoint process for this specific tenant, with the name and description we typed.

If we would like to perform the same operation again, creating an additional checkpoint on the same virtual machine, we get a message telling us that the existing checkpoint will be deleted.

This is because that the current checkpoint integration in WAP will only keep one checkpoint, and avoid the scenario where you could potentially have a long chain of differential disks.

When we create the second checkpoint, we can switch back to VMM to see what’s actually happening:

First, a new checkpoint is created.
Second, the previous checkpoint is deleted.

When we explore the checkpoints settings on the VM itself afterwards, we see that we only have the latest checkpoint listed.

Regarding the restore process, we can also perform this from the same view in the tenant portal.
Once you click on the restore button, the tenant portal will show you the metadata of the available checkpoint, such as name, description and when it was created. Once you click the confirm button, the restore process will start in VMM.

Now what?

If you are familiar with how checkpoints in Hyper-V works, then you know that each static disk will  be either .vhd or .vhdx – depending on the format you are using (.vhdx was introduced with Windows Server 2012 and should be the preferred format, but Azure is still using .vhd).
Once you create a checkpoint, a new disk (.avhd or .avhdx) will be created– a differential disk, containing all the new write operations, while read operations will occur on both the parent disk (vhdx) and the newly created differential disk. 

To summarize, this might not be an ideal situation when it comes to performance, life-cycle management and storage optimization.

Since we don’t have any action in the tenant portal to perform a delete operation, this can be scary in some scenarios.
The fact that the VM will always run on a checkpoint once a checkpoint is created, means you will always be able to restore to your latest checkpoint from the portal.

In order to solve this challenge, we can leverage the integration of Service Management Automation in Azure Pack.
One of the best things with Azure Pack and the VM Cloud resource provider, is that we can extend it and create valued added solutions and services by linking certain actions happening in the tenant portal, to automated tasks that are executed by a SMA runbook in the backend.

The following screenshot shows that there’s an event related to creation of VMM Checkpoints performed by the tenant, which can easily be linked to a runbook.

Here’s an example of a runbook that will check for checkpoints created on VMs belonging to a specific VMM Cloud that is used in a Hosting Plan in WAP. If there’s any checkpoints that exists, they will be deleted and the VMs will have their disks merged back to a static disk (.vhd/.vhdx).
Wokflow to check for - and eventually delete old VM checkpoints

 workflow delete-scvmcheckpoint  
    # Connection to access VMM server. 
    $VmmConnection = Get-AutomationConnection -Name 'SCVMM'  
    $VmmServerName = $VmmConnection.ComputerName  

 # Import VMM module. 
Import-Module virtualmachinemanager  
# Connect to VMM server. 
Get-SCVMMServer -ComputerName $Using:VmmServerName  

$vms = Get-SCVirtualMachine | Where-Object {$_.Cloud -like "*Copenhagen IaaS*" -and  $_.VMCheckpoints }
       foreach ($vm in $vms)
Get-SCVMCheckpoint -VM $vm | Remove-SCVMCheckpoint -RunAsynchronously

}-PSComputerName $VmmServerName -PSCredential $VmmCredential 

This simple code can so be added to a schedule that will execute this runbook on a daily basis – as an example, ensuring that no VMs in the cloud will run on a checkpoint on a long term.

Thanks for reading!

Wednesday, April 15, 2015

Why I am investing in DSC

In order to get a good grasp on something new, like a technology, it is always important to find a use case.

Once you have a use case, I can ensure you that the learning process is much more interesting, fun – and perhaps easier too.

That is what I did when I went deep into Desired State Configuration. I found a use case.
My use case was to leverage DSC as part of VM Roles in Azure Pack in a way that would be valid for the future too.

Here comes some reasons for my decision. 

Powershell has been around for some time now, and one of the best benefits by learning and using the shell is the amount of work you are able to do, combining modules, components, technologies and much more through the same API. Considering that everything that MS builds and do – regardless of cloud, will be accessible and manageable through Powershell in addition to other options, ensures that this is a real no-brainer.

With Windows Management Framework 4.0, we also got Powershell Desired State Configuration added to our table.
Powershell Desired State Configuration is Microsoft’s way to implement an idempotent configuration that ensures that the “desired state” will be reached by applying the entire configuration, regardless of the current state.

-          But, what does this really mean? Aren’t we able to do everything using native Powershell scripts already?

That is correct. There’s no “limits” by using Powershell natively today.
However, with native Powershell scripts you are responsible for building all the error handling and logic into your scripts. And as you probably know, that can be both very time consuming and challenging.

Desired State Configuration handles this automatically for you, letting you make and deploy any incremental changes to your configuration over time without risking the system to be put in a bad state.
If you have any configuration drift? Depending on how the Local Configuration Manager is configured – the engine that’s responsible for applying the configuration and follow the instructions, the system can heal itself by enforcing the desired state.

Think of Powershell Desired State Configuration as a contract between you and your nodes (manageable objects).

In order to create and deliver this “contract”, Desired State Configuration is based on CIM – and use WinRM for communicating. CIM uses a language called Manaed Object Format – often referred to as “MOF”. Powershell Desired State Configuration is a way to create and distribute MOF files that can be applied to systems supporting this standard.

The way it’s applied to the node(s) is either through “Push” or “Pull”.

(The difference between Push and Pull is out of scope right now and deserves a dedicated blog post later on. I promise).

To put it short, the Pull mechanism requires some infrastructure in order to work, where the node(s) are talking to the Pull server – either through SMB, Http or Https.

The Push method is pretty straight forward and what you can start using right out of the box. DSC requires that WinRM listeners are configured so that the CIM can push the configuration to the remote systems.

Here’s an example of how a Powershell DSC Configuration looks like:

configuration DNS
    node kndsc006
        WindowsFeature DNS
            Name = "DNS"
            Ensure = "Present"
            IncludeAllSubFeature = $true


Start-DscConfiguration -wait -force -Verbose .\DNS

As you can see, the format here is quite easy to read.
We can easily see that we will install (Ensure = "Present") DNS (Name = "DNS") on the target node (kndsc006). 

Actually, it is so easy to read that Powershell newbies like me are able to manage J

Hopefully this gave you some more context about the “why”, but we are not done yet.

In Azure today, we are able to leverage DSC as part of the VM extension, meaning we can create – upload – and apply our DSC configuration to Azure IaaS virtual machines. The method of applying the config for these VMs are “Push”.

As you probably know, we don’t have the exact same capabilities on-prem in order to leverage DSC as part of Azure Pack. However, we are able to simulate the same experience at some extent, by using the combination of DSC, SMA and VM Roles ( )

Moving forward, we know that the consistency across clouds will be as near as 1:1 with the introduction of Azure Resource Manager that will introduce us for a complete new way to interact with our cloud services – regardless of location. Also worth to note, the Azure Resource Manager itself will be idempotent.

What about your existing DSC scripts?
Exactly, that is the main point here. These configurations will be valid using Azure Resource Manager too J

So in essence, you invest in DSC now and use it both for Azure Pack (VM Roles + SMA) and Azure (VM Extension), and later on you can reuse the investment you’ve made into the era of Azure Resource Manager.

Hopefully this gave you some inspiration to start learning Desired State Configuration, available in Windows Management Framework 4.0 – but also available in 5.0 (which is in Preview).
Please note that everything you do in Azure when using the DSC VM Extension there is based on the 5.0 version.

Monday, March 16, 2015

Application Modeling with VM Roles, DSC and SMA

Earlier this year, I started to go deep into DSC to learn more about the concept, possibilities and most important, how we can improve what we already have and know, using this new approach of modeling.

For more information and as an introduction to this blog post, you can read my former blog post on the subject:

Desired State Configuration is very interesting indeed – and to fully embrace it you need to be comfortable with Powershell. Having that said, Desired State Configuration can give you some of what you are requiring today, but not everything.

Let me spend some minutes trying to explain what I am actually saying here.

If you want to use DSC as your primary engine, the standard solution to configure and deploy applications and services across clouds throughout the life cycle, there is nothing there to stop you from doing so.
However, given the fact that in many situations, you won’t be the individual who’s ordering the application, server and dependencies, it is important that we can make this available in a world full of tenants with a demand for self-servicing.

Looking back at how we used to do things before to handle the life-cycle management of applications and infra, I think it is fair to say it was something like this (in context of System Center):

1)      We deployed a Virtual Machine based on a VM Template using SCVMM
We either
a)      Manually installed and configured applications and services within the guest post VM deployment
b)      Used SCCM to install agents, letting the admin interact with the OS to install and configure applications using a central management solution
2)      If we wanted to provide monitoring, we then used SCOM to roll out the agents to our servers and configured them to report to their management group
3)      Finally yet importantly, we also wanted to be secure and have a reliable set of data. That’s why we also added backup agents to our servers using SCDPM

In total, we are talking about 4 agents here (SCVMM, SCCM, SCOM and SCDPM).
That is a lot.

Also note that I didn’t specify any version of System Center, so this was probably even before we started to talk about Private Clouds (introduced with System Center 2012).

And that’s the next topic, all of this in the context of cloud computing.

If we take a walk down the memorial lane, we can see some of Microsoft’s least proud moments, all the attempts in order to bring the private cloud a fully functional self-service portal.

-        We’ve had several self-service portals for VMM that later was replaced by different solutions, such as Cloud Service Process Pack and App Controller
-        Cloud Service Process Pack – which was introduced with SC 2012 – where all the components were merged into a single license, giving you out-of-the-box functionality related to IaaS.
The solution was one of the worst we have seen, and the complexity to implement it was beyond what we have seen ever since.
-        AppController was based on Silverlight and gave us the “single-pane of glass” vision for cloud management. With a connector to Azure subscriptions (IaaS) and to private and service provider clouds (using SPF), you could deploy and control your services and virtual machines using this console

Although it is common knowledge that AppController will be removed in vNext of System Center ( ), AppController introduced us to a very interesting thing: self-service of service templates.

The concept of service templates was introduced in System Center 2012 – Virtual Machine Manager, and if we go back to my list of actions we needed to perform, we could say that service templates at some point would replace the need of SCCM.
Service Templates was an extension to the VM template. It gave us the possibility to design, configure and deploy multi-tier applications – and deploy it to our private clouds.
However, I have to admit that back then; we did not see much adoption of service templates. Actually, we did not see some serious adoption before Microsoft started to push some pre-configured service templates on their own, and that happened last year – at the same time as their Gallery Items for Azure Pack was released.

To summarize, the service template concept (which was based on XML) gave the application owners and the fabric administrators a chance to interact to standardize and deploy complex applications into the private clouds, using AppController. In the same sentence there we found AppController (Silverlight) and XML.

If we quickly turn to our “final destination”, Microsoft Azure, we can see that those technologies aren’t the big bet in any circumstances.

VM Roles are replacing service templates in the private cloud through Windows Azure Pack.

A VM Role is based on JSON – and define a virtual machine resource that tenants can instantiate and scale according to their requirements.

We have in essence two JSON files. One for the resource definition (RESDEF) and one for the resource extension (RESEXT).
The resource definition describes the virtual machine hardware and instantiation restrictions, while the resource extension definition describes how a resource should be provisioned.

In order to support user input in a user friendly way, we also have a third JSON file – the view definition (VIEWDEF), which provides the Azure Pack details about how to let the user customize the creation of a VM Role.

These files are contained in a package, along with other files (custom resources, logo’s etc) that describe the entire VM Role.

You might ask yourself why I am introducing you to something you already know very well, or why I am starting to endorse JSON. The answer lays in the clouds.

If you have every played around with the Azure preview portal, you have access to the Azure Resource Manager.
ARM introduced an entirely new way of thinking about you resources. Instead of creating and managing individual resources, you are defining a resource model of your service – to create a resource group with different resources that are logically managed throughout the entire life cycle.

-        And guess what?

The Azure Resource Manager Templates is based on JSON, which describes the resources and associated deployment parameters.

So to give you a short summary so far:

Service Templates was great when it came with SCVMM 2012. However, based on XML and AppController for self-service, it wasn’t flexible enough, nor designed for the cloud.

Because of a huge focus on consistency as part of the Cloud OS vision by Microsoft, Windows Azure Pack was brought on-premises and should help organizations to adopt the cloud at a faster cadence. We then got VM Roles that should be more aligned with the public cloud (Microsoft Azure), compared to service templates.

So we might (so far) end up with a conclusion that VM Roles is here to stay, and if you are focusing too much on service templates today, you need to reconsider that investment.

The good, the bad and the ugly

So far, the blog post has been describing something similar to a journey. Nevertheless, we have not reached the final destination yet.

I promised you a blog post about DSC, SMA and VM Roles, but so far, you have only heard about the VM Roles.
Before we proceed, we need to be completely honest about the VM Roles to understand the requirement of engineering here. To better understand what I am talking about, I am comparing a VM Role with a stand-alone VM based on a VM Template:

As you can see, the VM Role gives us very much more compared to a stand-alone VM from a VM template. A VM Role is our preferred choice when we want to deploy applications in a similar way as a service template, but only as single tiers. We can also service the VM Role and scale it on demand.

A VM on the other hand, lacks all these fancy features. We can purely base a stand-alone VM on a VM Template, giving us a pre-defined HW template in VMM with some limited settings at the OS level.
However, please note that the VM supports probably the most important things for any production scenarios: backup and DR.
That is correct. If you use backup and DR together with a VM Role, you will end up in a scenario where you have orphaned objects in Azure Pack. This will effectively break the relationship between the VM Role (CloudService in VMM) and its members. There is currently no way to recover from that scenario.

This got me thinking.

How can we leverage the best from both worlds? Using VM Role as the engine that drives and creates the complexity here, supplemented by SMA and Desired State Configuration to perform the in-guest operations into normal VM templates?

I ran through the scenario with a fellow MVP, Stanislav Zhelyazkov and he nodded and agreed. “-This seems to be the right thing to do moving forward, you have my blessing” he said.

The workflow

This is where it all makes sense. To combine the beauty of VM Roles, DSC and SMA to achieve the following scenario:

1)      A tenant logs on to the tenant portal. The subscription includes the VM Cloud resource provider where the cloud administrator has added one or more VM Roles.
2)      The VM Role Gallery shows these VM Roles and provides the tenant with instructions on how to model and deploy the application.
3)      The tenant provides some input during the VM Role wizard and the VM Role deployment starts
4)      In the background, a parent runbook (SMA) that is linked to the event in the portal kicks in, and based on the VM Role the tenant chose, it will invoke the correct child runbook.
5)      The child runbook will deploy the (stand-alone) VMs necessary for the application specified in the VM Role, join them to the proper domain (if specified) and automatically add them to the tenant subscription.
6)      Once the stand-alone VMs are started, the VM Role resource extension kicks in (which is the DSC configuration, using push) that based on the parameters and inputs from the tenant is able to deploy and model the application entirely.
7)      Once the entire operation has completed, the child runbook will clean-up the VM Role and remove it from the subscription

In a nutshell, we have achieved the following with this example:

1)      We have successfully been able to deploy and model our applications using the extension available in VM Roles, where we are using Desired State Configuration to handle everything within the guests (instead of normal powershell scripts).
2)      We are combining the process in WAP with SMA Runbooks to handle everything outside of the VM Role and the VMs.
3)      We are guaranteed a supported life-cycle management of our tenant workloads

Here you can see some screenshots from a VM Role that will deploy Windows Azure Pack on 6 stand-alone VMs, combining DSC and SMA.

In an upcoming blog post, we will start to have a look at the actual code being used, the challenges and workarounds.

I hope that this blog post showed you some interesting things about application modeling with VM Roles, SMA and DSC, and that the times are a-changing compared to what we used to do in this space.