Azure Stack integrated systems – how to bring Azure to your data center

Coming up, whether you’re an enterprise
or a service provider; we’re going to take a look at Azure Stack
integrated systems. The unique combination of optimized hardware
software and services that delivers Azure with Azure Stack to your own data center. And as you take on the role as you own cloud operator. We’ll explain the Azure Stack
integrated systems approach and demonstrate the end-to-end IT
experience including, planning the size of your system
and setting yourself up for scale, configuring the system with your hardware partner, provisioning resources
to your developers, plus ongoing management and monitoring. Microsoft Mechanics I’m joined today by Vijay Tewari from the Azure Stack engineering team, welcome. Thanks for having me on the show! So we’ve covered Azure Stack in past shows. This really gives you a hybrid platform so that you can use the same services,
automation, tools and portal as Azure public cloud. But your apps are actually running
on-premises in your own data centers. One area we haven’t really explored in detail is
really how Azure Stack is delivered and integrated within your own data center. So I’m excited to see the direction that you’re taking now with Azure Stack integrated systems. But can you explain more about the approach that
you’re taking here? Absolutely as you mentioned the
premise of Azure Stack is to bring the core attributes we run in the public cloud of Azure into your data center. For example, if you want to move to the cloud and you’re held back due to regulatory compliance or other technical or business reasons, this opens up new options for you to be able to move to the cloud
yet operate Azure in your data center. We deliver Azure Stack as a
purposeful integrated systems that combines software, services, and hardware. And these are highly optimized systems that benefit from learning from the public cloud, but the agility at scale to provide a
pragmatic hybrid cloud solution to you. And these can be obtained from our partners like Dell, HP, and Lenovo. And Cisco coming later this year. And as with the cloud, there are no upfront costs. You only pay for the Azure services that you use. And we also give you a full
end-to-end support experience shared between us and our partners. Alright so let’s make this real because
if you’re watching at home I bet people are wondering what’s the experience to get all stuff up and running in my own data center. Sure so let’s walk through what it means to a day in the life of getting Azure Stack
running in your data center. The first step is to work with your hardware partner to plan the size of your system based on the needs of your apps and services. Do you have guidance in terms of the
size when I start planning for my system the initial kind of roll out? So you can decide upfront what is the size
of your scale unit. This is a group of servers that you can
start with as small as 4. And it can be as large as 12 servers. Soon you’ll be able to add
capacity to a scale unit by adding more servers. and then you’ll be able to add regions. Think of regions as a geographical area that you define which you will be able to add to
your Azure Stack instance. And then you’ll be able
to add multiple scale units per region. And as you can see just like Azure. Azure Stack is architected for scale and we are delivering incrementally in this area. All right so let’s put this in to context here. What do I get for say a 12
server set of Azure Stack servers? If you have a 12 server system with the right CPUs, with the right amount of memory and the right amount of storage you can get about 400 plus Azure D2 size VMs with the right amount of resiliency. Now a D2 size VM is two virtual CPUs and seven gigabytes of memory. So as you can see It’s a pretty significant amount of capacity. Plus it has redundancy built in. But once I’ve determined the size of the system what I do next so next? Next you’ll need to share your
required configuration so that the hardware partner can build a system
that’s operational in your environment. By completing a deployment worksheet with your information. Such as Azure active directory, your network settings and other information that’s required for customizing system for your environment. Okay so I’ve shared the specs out with
my hardware partner. How do I know the system is actually
ready to go and is usable? So our hardware partner ships the
hardware to your data center and along with it comes some engineers who will then deploy and integrate the
system into an environment. They will then validate it to make sure it is
operational in the environment before they hand over the system to you. Now Azure Stack is one system in
your sea of other systems. But it’s not an island. You can work with your system integrator
to make sure you can connect Azures Stack monitoring and identity subsystems to your centralized systems. Okay and once I’ve got all this stuff up and running how do I start to provision services out
to my customers, my developers in house and really manage that with my
organization? Right so let’s start by actually first
looking at an Azure portal. Now this is the Azure portal and as
a tenant I’m consuming resources from the Azure portal. And this is the Azure Stack portal. Now by virtue of the
consistency between Azure and Azure stack, you get the same functionality between
Azure and Azure Stack. Along with Azure Stack, we also ship an
administrative portal and that administrative portal allows you to
manage services in your data center. Azure Stack comes with a bunch of
foundational services out-of-the-box. Infrastructure as a service with compute,
storage with blobs, tables and queues and networking. And in your local gallery you
can select the services that you want to be able to expose. You can actually go to
Azure and you can select services from Azure that you want to
bring down into your data center and bring them down and expose them to your developers from your local gallery. You then need to go and set offers, plans and quotas which is how you encapsulate
the services that you want to be able to provide to your developers. And then you can set quotas as to how much of the services they can actually consume from
that particular offer and plan. So you really are a cloud operator here now. Once it’s up and running your developers start to consume resources How do I go about operating Azure Stack in general? So as you can see this is the Azure
Stack administrative portal and this gives you the health of
all your resources. As you can see I happen to have one region in this Azure Stack instance and I’m seeing a warning. Let’s click on the alert and see what
the alert is telling me. So the alert is showing me that a physical disk and its connectivity has been lost to that disk. And if you look at the details of that alert, it is clearly telling you a set of steps that you can
take to remediate that alert. It’s asking you to recede the disk. and if that doesn’t work to reach out to your hardware partner to actually replace the disk. Now by virtue of the resiliency that the system has, the system is absolutely
continuing to operate. So that’s one of the benefits that you get out of an
integrated system for Azure Stack. You said something interesting earlier. You actually only paying for what you use even though you’ve got capital and
infrastructure in your own data center. But how do I keep track of all the
resources and things that people are consuming? And ultimately, what I then as
the IT department need to pay for? So just like other Azure services, you could
watch this consumption in the Azure portal or use the API to get a sense of
what is the consumption in your environment. So how do we go about getting the
alert that you’re showing here to actually appear in my other data
center management operations tools? So all our administrative tasks are
actually available via an API or a PowerShell. That allows you to integrate this into your existing management tools in your data center like Nagios or Systems Center Operations manager. So here I’m actually showing an instance of
operations manager where I’ve imported the Azure Stack management back and you can see the alerts from an
Azure Stack instance being surfaced into your operations
manager environment. So we’ve covered monitoring performance with hardware and the support there, but what happens if there’s a software bug or an issue that we have to report? So it’s a simple model. So just like for the hardware from a
software perspective, if one of the core services is faulty you’ll get an alert in the environment and you’ll be given a
simple set of steps to be able to take to remediate that. However if that doesn’t work, then you can actually go to the a short portal and you can raise a new support request. I’m actually here in the Azure portal and as you can see I’ve selected my subscription and I’m selecting Azure stack to be able to
raise a new support ticket. Now the support ticket is received by us in the back-end and We’ll triage it between us and the partners and make sure that you get a seamless experience between us and the partner. Okay makes sense and I think I saw earlier there was a software update that was pending. How does that work given that we want keep that consistency between Azure Stack and Azure public cloud? As you can see here there is an alert
which refers to the fact that there is an update that’s
available for Azure Stack. Our updates can contain security fixes. They can contain fixes for issues
that were found in the product, or they can contain new capabilities. And how does the actual software
update process work? Now since we have a well understood and defined architecture between us and the hardware partner, we can orchestrate the application of the
updates in a manner that ensures there is minimal downtime for the workloads and the services that are running on Azure Stack. For example, when we are updating a host we can drain all the virtual machines that
are running on a particular physical host. Update that host and then rehydrate this host with the virtual machines that were moved to a different machine while the system
was being patched. Makes sense, but how often do I
expect updates to be rolling in and is there a way I can defer them? So our goal is to be able to ship updates as
frequently as is acquired by the customer. But as a customer you have the choice of being
able to defer an update for up to 6 months. So let’s switch gears a little
bit to disaster recovery. Since we’re maintaining its consistency
with Azure and Azure Stack, can I use Azure Stack to be a DR solution
for my Azure Stack instance? Yes, you can actually fail over your virtual machines from Azure Stack to Azure. By setting up this relationship you can replicate virtual machines from Azure Stack to Azure. And in the case of a failure of an Azure Stack instance you can then spin up those workloads in Azure and continue in the face of disaster. It’s great to see all the progress that you’ve made so far with Azure Stack, but what’s next? So there are three areas that you will
see us delivering on soon. The first is around expanding your
Azure Stack infrastructure So that for example you can have different regions followed by multiple scale units as I
had mentioned earlier. Secondly we are working on more
purpose-built monitoring and alerts, and intend to integrate it with Azure log
analytics and OMS. Lastly, we will integrate the automation
of firmware updates into our patching and update framwork. So thanks Vijay for taking us through
all the different practicalities of getting Azure Stack running in our data centers. But where can people go to learn more? You can learn more at the link below and work with your hardware partner to start planning your Azure Stack integrated system. This is really good stuff and also keep
watching Microsoft Mechanics for the latest technology updates across Microsoft. That’s all the time we have
for this show. Thanks for watching and good bye for now. Microsoft Mechanics


Leave a Reply

Your email address will not be published. Required fields are marked *