Start with 7 free days of training.

Gain instant access to our entire IT training library, free for your first week.
Train anytime on your desktop, tablet, or mobile devices.

Architecting Microsoft Azure Solutions

Ben Finkel covers the topics and material required to architect and organize enterprise implementations on the Microsoft Azure Platform. The 70-534: Architecting Microsoft Azure Solutions takes a high-level approach to the structure and organization of implementations on the Microsoft Azure Platform. This course also helps you prepare for the certification exam by covering each exam objective, as well as demonstrating practical applications for using these features....
Ben Finkel covers the topics and material required to architect and organize enterprise implementations on the Microsoft Azure Platform. The 70-534: Architecting Microsoft Azure Solutions takes a high-level approach to the structure and organization of implementations on the Microsoft Azure Platform. This course also helps you prepare for the certification exam by covering each exam objective, as well as demonstrating practical applications for using these features.

The 70-534 exam and this course cover the following general topics: Design Azure infrastructure, Securing Resources, Storage and Data Access Strategies, Advanced Application Development, Websites, and Disaster Recovery and Business Continuity.

Note: Microsoft recently updated many of its exams, including those associated with Azure. CBT Nuggets plans to update this course to align it with the new exam objectives and technology advances, but please note that this course is currently aligned with outdated exam objectives.

Recommended Experience Recommended Equipment
  • A desktop computer running Microsoft Windows version 7, 8, or 10
  • Microsoft Visual Studio 2013
Related Certifications
  • MCSA: Cloud Platform
  • MCSE: Cloud Platform and Infrastructure
Related Job Functions
  • Web development
  • Application engineer/team lead
  • Systems Manager
Ben Finkel has been a CBT Nuggets trainer since 2014. His areas of expertise include Google Developer Expert, Google Certified Trainer, Microsoft ASP.NET (WebForms and MVC), Data Analysis and Design, Relational Database Architecture, MS SQL Server, Microsoft C#.NET, Javascript, jQuery, Microsoft Visual Basic .NET & 6.0, .NET Reporting Services, MS Access, and Python.
 show less
1. Azure GFS Datacenters (12 min)
2. Azure Virtual Networks (19 min)
3. Designing Azure Compute (19 min)
4. Azure VPN and ExpressRoute (7 min)
5. Azure Load Balancing Services (9 min)
6. Azure Active Directory (12 min)
7. Hybrid Identities (12 min)
8. Azure Data Security (10 min)
9. Role-Based Access Control (11 min)
10. Designing Data Storage (14 min)
11. Azure Mobile Services (12 min)
12. Push Notifications (7 min)
13. Web APIs (10 min)
14. Hybrid Applications (8 min)
15. Azure Media Services (7 min)
16. Compute-Intensive Applications (10 min)
17. Long Running Applications (13 min)
18. Selecting Storage (15 min)
19. Architecting Azure Web Apps (8 min)
20. Deploying Azure Web Apps (11 min)
21. Business Continuity with Azure Web Apps (10 min)
22. SCCM Deployment with Azure (10 min)
23. Design a Monitoring Strategy (9 min)
24. Business Continuity/Disaster Recovery (10 min)
25. Azure Automation (10 min)
26. Azure Classic vs. Resource Manager (10 min)

Azure GFS Datacenters


Azure GFS datacenters. Hey, everyone. This is Ben Finkel. And in this Nugget, we're going to talk about the powerful datacenters that underlie everything Azure. These are the GFS datacenters where all the computers are that make Azure work. By now, you must have heard the term, the Cloud.


It's the latest buzzword in IT. The Cloud this, the Cloud that. Buzz, buzz, buzz. Cloud, Cloud, Cloud. But what is the Cloud? What makes it the Cloud? If you've been viewing Nuggets here at CBT Nuggets on any of the Cloud services that are available-- Azure or Google Cloud Platform or AWS-- then you understand and you're familiar with a lot of the services that you could take advantage of inside of the Cloud.


But the services don't make the Cloud. The services are a product of the Cloud. The Cloud itself is really a gigantic array of datacenters. These are datacenters that look just like-- at least at a high level-- a datacenter that you might have worked with on premise at your own local organization.


But I can promise you the scale, the scope of these datacenters for Microsoft Azure, is way beyond anything that you've worked with unless, of course, your job is to work at Microsoft Azure on their datacenters. Then maybe you've worked on them. But otherwise, these datacenters support the Microsoft Azure Cloud.


And they have to be dramatically different. They have to be operated dramatically different in order to support Microsoft Azure. First of all, they're available in 17 different regions. All over the globe, these are all networked together in all these different places.


You can get to these datacenters from 140 different countries and you get operate in 10 different languages and 24 currencies. And these datacenters not only underlie the Azure services that you and I can use, but they also underlie the Microsoft products and technologies that are available on the internet.


So these are the same datacenters that are being used to support things like Bing Search and Office 365. Microsoft Xbox, if you use Xbox, their video games run online inside of these Azure GFS datacenters. So these datacenters have to serve up an enormous volume of data to a wide variety of different products and services to different customers, both inside and outside of Microsoft themselves.


It's got to do this around the globe in real time with incredibly low latency and incredibly high reliability. If you have any kind of background in infrastructure, that ought to make you go wow a little bit. That's a big problem to solve. But the Azure GFS datacenters are here to solve that problem.


Now just a little side note for the purposes of the exam, you may see both the term GFS or the term MCIO, Microsoft Cloud Infrastructure and Operations. They had an internal name change of the department. It's no longer GFS. But depending on what documentation or the exam that you happen to be looking at when those questions or when the documentation was written, you might see either terminology.


Just understand that they are completely interchangeable. The first way that Azure GFS datacenters are probably different than the datacenters you've worked on is just the sheer volume and geographic dispersion of these datacenters. You can see here the current list of the 17 regions that Azure offers, and they're all over the globe.


They're everywhere. And, of course, that' because a datacenter really represents a physical building, a physical data center, where your services can run. And in order to provide the best service to your users, you want to run your services in a datacenter that's geographically close to them.


That will reduce the latency. So when you spin up a new virtual machine or a web app or an SQL database, you can choose what region you want to run it in. So you want to choose to run it in a region that's going to close your primary user base. And, of course, don't worry.


If your users are all over the globe, Azure provides ways to share your system across multiple different regions so that you can ensure that your users always have access to a datacenter that's relatively close to them, reducing the latency, improving the performance and the reliability of your product.


A win-win across the board. Now it's important to note that Azure is also available in China, but it's kind of separate and distinct from the rest of this. It's provided by a third party ISP. So the datacenter does not share data in and outside of China.


But if you need to run Azure in China, you can do that as well. Now all of these regions, all these datacenters are not actually the same. The services that you can get vary by region. And I've just clicked a little picture of the diagram here. But if you go to this URL-- and I'll bring it right up on screen-- you can actually see the services that are available per region.


Down on the left-hand side is each service. So if you want, let's say, D-series, you can see that you can get D-series in all of these different Eastern US locations, but not in these government locations in Virginia, for instance. So as you scroll down here, you've got all your different services down the left.


Across the top, you have your different regions. You can see what services are going to be available in what region. Another way that GFS datacenters are different than yours? Well, Azure datacenters are big. Some are as large as three cruise ships packed end to end and they're just spilling over with servers.


They have more servers than you can count stacked inside of them. And that introduces some unique complexities, some unique challenges, that Azure datacenters have to deal with. For one, hardware failures become the day-to-day norm. Because you have so many servers, you're going to frequently and often have servers that are failing, drives that are failing, equipment, power supplies, et cetera that is failing.


In a typical datacenter, what you're concerned with is your mean time between failures. You want to reduce the amount of times that you have a failure, and you want to make sure and keep your uptime as close to 100% as possible. But when you're dealing with this many servers and so many expectations of failure, what you're more interested in is your mean time to recover.


You want your services that are running on a piece of equipment that might fail to quickly and seamlessly migrate over to some clean-running equipment. They use something called commodity servers for this, which means that they are less likely to stay up and be reliable, but they're cheaper and easier to move services around between them.


So inside of an Azure GFS datacenter your meantime to recover, MTTR, becomes much more important than your mean time between failure, your MTBF. And this is unique or at least a little bit different than your typical datacenter. Now in order for this to happen, your workload migration as much as possible must be automated.


You can't count on people to pick up and move things-- services, drives, physical equipment, et cetera-- between different servers. It's all got to happen in an automated fashion. Software has to migrate off of bad equipment and onto functional equipment.


And that's a big part of what makes Azure so impressive is the software layer that runs on top of this infrastructure that keeps things running smoothly, that makes everything up here reliable, even though underneath is very tumultuous and there may be failures and things going wrong all the time.


You and I, as users of Azure, we don't really see that. Another thing to understand about GFS datacenters is that security is the highest priority operation inside of any datacenter. Think about it. This datacenter has got my data. It's got your data.


It's got Microsoft's data. It's got 1,000 other customers' data all inside of this great, big equipment-filled room. And that makes it a juicy target for cyber threats, as well as physical threats. Microsoft takes great care to ensure that the datacenters are as secure as they can possibly make them.


And this happens at a couple of different levels. First of all, the staff themselves, Microsoft staff, is carefully monitored to ensure that they don't inappropriately access the data or the information that's contained inside of those services. Physical security, of course, is provided.


Keep people from wandering into the datacenter that are allowed to be there. And threat management, software threat management, is provided as a base level of protection for all of your services that are running inside of the datacenter. So you can ensure that they are already, before information requests, et cetera even get to your virtual machine, even get to your web application, it's being scrubbed and monitored for potential threats.


You can read all about this and more at the Trust Center security page that I've just linked right here. This is a good URL if you need to understand exactly how secure the GFS datacenters are and you should want to go ahead and check that out, by the way, because security should be a big priority for us, as well as for Microsoft.


Now like I said, as your datacenters are really good at moving your services around off of failed hardware, but sometimes failures are more systemic or widespread than a single piece of hardware. And for that, we have update and fault domains. Update and fault domains are groups of machines, instances inside of Azure, that are separate and distinct from one another.


Usually on separate racks, separate power supplies, separate cooling systems, et cetera. And the reason for that is because when you have a widespread outage, either planned or unplanned, you know that at least you have some services running on a different system and are probably not affected by that outage.


So update domains are for planned updates to the service. Say they have to roll out a big update to the Microsoft Azure infrastructure software. Well, if all your machines were running on the infrastructure that was being updated, you would have an outage.


But as long as you spread your machines, your instances, cross update domains, well, they'll only ever update one domain and one time. So the machines that are running in the second update domain won't get updated. Fault domains work exactly the same, but they're for unplanned outages.


Let's say the power goes out to an entire row of racks at a datacenter. These things happen. They're ultimately run by people and problems happen. Well, as long as you have your machines spread across different fault domains, you know that you have a whole set of machines on a different row of racks that are probably not affected by this outage.


The likelihood of a problem happening across two different fault domains is very, very low. In fact, if it happens across two fault domains, probably the entire region is out. And at that point, you need cross-region replication. All you need to do to enable update and fault domains is simply have two or more virtual machines in an availability set.


It's really easy to use. You don't have to do anything other than create the availability set and then add your virtual machines to it. When you do that, they get automatically spread across these different domains. And as long as you have two or more VMS in that availability set, you qualify for 99.99% uptime SLA.


So as long as you have more than one machine in your set, they will get spread across multiple update and multiple fault domains. And we'll talk more in depth about that in a different Nugget. And that will get you your 99.9% uptime. Now I just mentioned, of course, the possibility that you have a region-wide outage.


Maybe a natural disaster takes out an entire datacenter. And in order to deal with that, your fault and update domains are not enough. You need to make sure that you have cross-region replication set up. So for your really important high-priority production systems, you can set up cross-region replication, which is exactly what it sounds like.


You're just going to replicate your information to a secondary region. You choose the primary region where your services are going to run from and your users are going to access. And for whatever reason that region goes down, that primary is lost, well, the system automatically fails over to the secondary region.


Your users are not afflicted by the outage. There's simply accessing the data, accessing the services, from the secondary region. Again, what's nice about this is once you set up the replication, once you've set up the cross-region backups, you don't have to do anything to failover.


In the event that there's a problem in the primary region, the secondary region will automatically spin up. It's like a hot backup. It will automatically spin up and start getting used by your users. They might see a little bit slower service because that secondary region is not as close to them as the primary region, but at least they won't be down.


They won't be affected by the outage. Now the last thing that you may see or you may have to deal with inside of a GFS datacenter is service throttling. And this is to ensure consistent delivery, not just to you, but to every customer, every client, inside of the datacenter.


Because it's a shared system, a shared datacenter, they need to throttle service and impose some quotas on how much you can use in order to grant everybody consistent access and get everyone consistent delivery. Now some of these quotas have soft limits that can be increased by spending more money.


Some of them have hard limits that can't be increased or can only be increased by contacting Microsoft Support directly. But for the most part, you're going to find that they are well above and beyond anything you need. And inside of the various Nuggets where we talk about specific services, we'll often talk about how you can set up or use those services in parallel in order to scale past some of those quotas and throttles limits.


So that concludes this Nugget on Azure GFS datacenters. Just to recap, the Azure GFS datacenters are located in 17 different region all over the world. They're available from 140 different countries and 10 different languages. The Azure GFS datacenters have to be run and managed differently than your typical on-premise datacenter in order to support the enormous infrastructure that Microsoft Azure supplies.


I hope this has been informative for you, and I'd like to thank you for viewing.

Azure Virtual Networks

Designing Azure Compute

Azure VPN and ExpressRoute

Azure Load Balancing Services

Azure Active Directory

Hybrid Identities

Azure Data Security

Role-Based Access Control

Designing Data Storage

Azure Mobile Services

Push Notifications

Web APIs

Hybrid Applications

Azure Media Services

Compute-Intensive Applications

Long Running Applications

Selecting Storage

Architecting Azure Web Apps

Deploying Azure Web Apps

Business Continuity with Azure Web Apps

SCCM Deployment with Azure

Design a Monitoring Strategy

Business Continuity/Disaster Recovery

Azure Automation

Azure Classic vs. Resource Manager

Please help us improve by sharing your feedback on training courses and videos. For customer service questions, please contact our support team. The views expressed in comments reflect those of the author and not of CBT Nuggets. We reserve the right to remove comments that do not adhere to our community standards.

comments powered by Disqus
Intermediate 4 hrs 26 videos


Training Features

Practice Exams
These practice tests help you review your knowledge and prepare you for exams.

Virtual Lab
Use a virtual environment to reinforce what you are learning and get hands-on experience.

Offline Training
Our iOS and Android mobile apps offer the ability to download videos and train anytime, anywhere offline.

Accountability Coaching
Develop and maintain a study plan with one-to-one assistance from coaches.

Supplemental Files
Files/materials that supplement the video training.

Speed Control
Play videos at a faster or slower pace.

Included in this course
Pick up where you left off watching a video.

Included in this course
Jot down information to refer back to at a later time.

Closed Captions
Follow what the trainers are saying with ease.
Ben Finkel
Nugget trainer since 2014