If you haven’t heard about Azure yet, I’d be surprised. Cloud based Platform applications like Windows Azure SQL Databases (WASDs) have been around for several years, and are becoming highly popular. WASDs (for example) don’t quite have all of the same features as Infrastructure based SQL Server. For those times when you need infrastructure based technologies, but still want to leverage the portability and ease of provisioning benefits of cloud based computing, Windows Azure Infrastructure as a Service (Iaas) may offer a solution. And IasS is… That rack you have in your datacenter with a bunch of blades that runs your virtual environment – you know the one. IaaS is that. Only bigger. Lots bigger. Massively big. The fact is, Windows Azure IaaS allows YOU to spin up any kind of virtual machine you wish – with a variety of different size options. These machines run in a remote datacenter, and come with full RDP access as well as Remote PowerShell script access enabled. Managing these machines is easy and done over standard Internet connections. Really? How does all of this work? If you’re familiar with Hyper-V or VMWare technologies, it’s very similar to that. Essentially, there are large datacenters based worldwide. When you provision a new machine – that is, create a new server – a VHD is created using one of the predefined image templates or one you define on your own. This machine is then housed in one of the Azure datacenter – should a disaster happen, The Azure infrastructure will handle moving it to another place that is not currently experiencing said disaster. So, as a SQL professional, what does this mean to me? Great question – Here are a few use cases to think about: If you’re building a new SQL environment, and want to have a place to test things, using Azure IaaS to temporarily build a DEV/TEST environment can be a quick way to get started. Azure IaaS machines are billed by the minute and only billed when powered on, so by managing the operational time of those environments costs can be easily managed. Azure IaaS can be used to extend your SQL environment. You can provision machines in the cloud to extend your AlwaysOn group. Joey D’antoni (b | t) talks about this as SQL Saturday. You can create virtual networks in IaaS to connect those machines that are in the cloud directly to your on premise network and use Azure as a remote datacenter. This can either be used in a DR plan, or to provide geocentric mirroring for users based worldwide. As Azure IaaS machines can be configured with different OSes like Windows or Linux, it can be a great option to run a Hadoop Cluster. This cluster can be easily scaled out during high processing times – like end of the month archival, and then scaled back during lower usage periods to help save costs What other options are there? I’d love to hear in the comments section at the bottom of this post.
How do I get started? Get started is pretty easy. First, you can get access to a free 30 day trial at http://www.windowsazure.com – before you can start creating any machines, you’ll need to make an account there. Once your account is provisioned, you are all set to begin provisioning new machines. We’ll take a look at two options for creating those machines in the next segment.