Why Run Databases in the Cloud?

Welcome in this Apress training, Azure SQL. This session is providing an introduction on why running databases in the cloud could be beneficial.

In this video, I will cover the following topics. I’ll start with a quick roadmap of the SQL Server Platform, the history, present, and future, followed by introducing SQL Server in an Azure Infrastructure as a Service model, after which I will do the same for Azure Platform as a Service concept. I will also highlight the basics of Azure SQL integrated security, and then ending with an overview of how Azure SQL can be deployed, taking high availability and disaster recovery in mind.

Now, when we look at a roadmap of SQL Server, it started about 25 years ago on large, expansive, physical SQL Server machines. And then about 15 years ago, we saw a move from larger dedicated physical servers to virtual machines running on top of VMware or Microsoft Hyper-V physical hosts.

This was the first huge optimization, as different workloads could run next to each other on the same physical machines. The virtual machines themselves were easier to migrate across physical machines, helping in disaster recovery and high availability.

The next wave, which obviously is still ongoing, is running or migrating your virtual machines from your on-prem data center into virtual machines running in a public cloud, like Azure AWS or Google Cloud. Now, besides running virtual machines in a public cloud environment, you could also move up to the next layer. And that’s running SQL Azure, the main topic of these course, using hosted databases in a public cloud, where you’re not responsible, not even deploying the underlying virtual machines anymore.

And then obviously, looking back at the physical dedicated SQL Server approach, in most scenarios, it means complex administration. We’re moving up to the cloud, moving from virtual machines to a full SQL Azure hosted database. It should lower the cost and make your administration a lot more efficient.

Now, let me zoom in a bit more on the SQL options when deploying it in SQL IaaS, Infrastructure as a Service. Overall, it’s a lot simpler than running the database servers in a virtual machine environment in your own data center. It’s easy to deploy, easy to manage, easy to migrate, since you have full control of the virtual machine.

Now, from a management perspective, it’s pointing at everything that Azure already has from an Infrastructure as a Service perspective, which basically means it’s running a virtual machine. You have virtual machine backup, disaster recovery, networking– everything.

Looking at a security perspective, when you deploy a virtual machine in Azure– and this is no different for running a SQL workload– you can define native Azure virtual network integration, where you’re going to define what communication from a network perspective is possible from and to your SQL Servers.

And then lastly, knowing that SQL Server is a business-critical workload, know that Azure IaaS has a lot of capabilities around disaster recovery built into the platform. And again, for SQL, it’s no different. If you want to build a SQL always on cluster, even across multiple Azure regions, then you can perfectly do that.

Now, let me do the same for SQL in a Platform as a Service or PaaS options, starting with, again, easy deployment. Lift and shift, where, again, it’s focused on the database workload. There are no virtual machines to manage, no virtual machines to deploy, to patch, and so on.

Similar to the IaaS services, it’s providing you a fully managed environment, where this time, it’s even built on top of the same Platform as a Service concept that you have in the overall Azure environment. Similar to the virtual machines, however, although we’re not really managing or deploying the virtual machines, SQL PaaS– and I’ll show you later on in a few demos in other sections in the course– SQL PaaS gives you full virtual network implementation, again, allowing you to lock down the connectivity to your database.

And then lastly, it should help you as an organization in really optimizing your business and the business workloads. Knowing that SQL is a typical business-critical application, Azure SQL gives you options to be more competitive, running your workloads more optimized, easier to migrate across versions– so quite a lot of benefits built into the platform.

Now, when we look at running SQL in Azure Virtual Machines, the architecture could look something similar to this one. We have the on-prem hybrid integration with Azure using ExpressRoute or Site-to-Site VPN. End users connecting to an Azure web app would be one option.

The web server could be on-prem but still connecting to an Azure SQL Virtual Machine backend. Out of the web app, there’s connectivity to the Azure SQL Virtual Machine virtual network. Users connect to the web app. Secured connectivity, everything’s encrypted– so a perfect design from a virtual machine infrastructure concept.

Now, switching to SQL instances in Platform as a Service, nothing’s really changing any more, besides we’re not running virtual machines in the cloud, but natively connecting to our SQL instances. A SQL DBA admin can securely connect using ExpressRoute or VPN to manage the SQL instances. And the end users connecting to the applications, a web app running on-prem or a web app running on Azure, by using the connection string– encrypted connectivity to the SQL instances.

And I may be switching back to a more advanced architecture, where high availability and disaster recovery are super important. Now, note that you can achieve that by using SQL virtual machine topology. Similar to building it in your on-prem SQL Always On cluster, you’re going to deploy multiple virtual machines in different Azure regions, allowing for high availability.

And then lastly, in the diagrams here, switching back to the PaaS option, know that Azure SQL gives you that built-in geo-replication, allowing you to replicate your database instances across multiple copies in the same region or ultimately across multiple regions.

And with that, we’re at the end of this section here, where I started with an introduction about SQL databases in the cloud and some kind of roadmap around the history, the present, and the future. Then, I covered what it takes to deploy SQL virtual machines in Azure, moving from IaaS to Platform as a Service.

Then next, I walked you through some security features and high availability diagrams, giving you confidence it is not all that different than running your SQL databases from an on-prem environment, virtual machine environment, and slightly moving to Azure SQL.

Now, stay tuned, as we have more detailed, more technical sections coming up. Now, thank you for being here, and have a nice day.