Learn about the different virtual machines SKUs and which virtual machines are best suited for specific workloads. Microsoft recommendations for SQL virtual machines are also outlined.
- [Narrator] Before we jump into and create virtual machines in Azure, we need to select the appropriate virtual machine for our deployment. Before we start diving into those SKUs, let's first review the guidelines for choosing a virtual machine. Our first step is to know the workload. And that may seem very simple. But not only do you need today's workload, you also have to understand what the future workload will be. Next, we need to review if our virtual machine can scale up. Not all virtual machines can scale up as far as we would like.
There are scaling limits within those SKUs. Always double-check that. And that's why we need to know what those future workloads could be. Next, double-check to make sure your virtual machine is available in the region that you would like to put it into. Not all virtual machines are available in all regions in Azure. Next, make sure you select the virtual machine to accommodate that future growth. And finally, what type of storage is required? For example, Microsoft recommends premium storage for your SQL virtual machines.
Now let's go ahead and review the different types of virtual machines that we can use in Azure. First, we have our general purpose virtual machines. We typically use these for test and dev or for small-to-medium databases. They're great for low-to-medium-traffic web servers. The general purpose virtual machine has a balanced CPU to memory ratio. And we'll find these virtual machines in the DSv2, the Dv2, the DS, the D, the Av2, and finally, the A0 to A7 SKUs.
Next, we have our cloud optimized virtual machines. We would use these for medium traffic web servers, for our network appliances, for our batch processes, and application servers. These virtual machines have a high CPU to memory ratio. And you'll find these in the Fs and F SKUs. Next, we have our memory optimized virtual machines. These would typically be used for our relational database servers or for workloads that require medium-to-large caches or in-memory analytics.
They have a high memory-to-core ratio. And these memory optimized virtual machines will be found in the GS, the G, the DSv2, D, DS, and Dv2 SKUs. We also have virtual machines that are optimized for storage. If you have applications that require high-disk throughput and I/O or big data, SQL, or NoSQL, then you'll be looking for a virtual machine in the Ls SKU. If you do a lot of video editing, then the GPU optimized VM may be your best solution.
We use these virtual machines for anything that has heavy graphic rendering or video editing. They include single or multiple GPUs, and you'll find these in the NV and the NC SKUs. And finally, we have the high-performance virtual machines. These are the fastest and most powerful CPU virtual machines that are available in the Azure data centers. They include RDMA which is high-throughput network interfaces. And you'll find these virtual machines in the H and the A8 through to 11 SKUs.
And finally, let's just quickly talk about SQL VM best practices. Microsoft recommends for your SQL VMs that if you're using Enterprise edition, use a DS3 or above virtual machine. If you happen to be using Standard or Web editions of SQL, then you can use a DS2 or above virtual machine. Microsoft does recommend that you use premium storage for your virtual machines. And best practice is not to use geo-redundant storage.
To recap, when it comes to planning for your virtual machines, the trick is to understand what your needs are now and what those needs will be in the future.
- Designing virtual machines
- Selecting appropriate VM SKUs
- Designing template deployment
- Deploying ARM templates via PowerShell and CLI
- Designing for availability
- Designing Azure Virtual Networks
- Azure VPN and ExpressRoute architecture and design