Azure VM Sizes for your SQL Server Workloads

Views 2018

With the introduction into the market by Microsoft the Azure data centers around the globe, this has changed the way we approach the consuming of resources for our workloads and our requirements for either hosting our application workloads in an On-Premises environment versus a Cloud environment. Now there is a lot of factors to take into account around this particular topic and that discussion can be for another day.

Here in Australia we are lucky enough to have access to 2 Azure data centers (as at the time of writing this) and the largest VM we can get our hands on is a GS5. This VM (and I have had the luxury or working with a client using a couple of these large VM’s) provides us with the following resource specs:

  • 32 vCores
  • 448 GB of Memory
  • Up to 64 Disks
    • Max I/O – 80,000
    • Max Throughput – 2 GB/s

Now to utilize this size server, comes with a nice price tag when we consider the number of core licenses that we need for SQL Server. What if your workload was Memory or Disk utilization hungry and not so much on the CPU side of things? The only way you could achieve the level of performance was to implement your solution on this Monster Server, with the associated costs.

This all now changes. Microsoft has announced the introduction of new versions of the following popular VM series:

  • DS
  • ES
  • GS
  • MS

These new versions now allow us to provision a VM (lets say a GS5) but constrain the vCPU count on the VM while maintaining the same resource allocation for Memory and Disk. The all new GS5-16 allows you to have the following:

  • 16 Cores
  • 448 GB of Memory
  • 64 Disks

While the new GS5-8 gives you:

  • 8 Cores
  • 448 GB of Memory
  • 64 Disks

 

So as you can see, if your workloads are not CPU bound, but you need more memory and more disk, then this is a great new option. Louis Vargas (Principal Program Manager at Microsoft) gives a great breakdown here.

Enjoy the new versions with the smaller Core Count based on your requirements while still providing you with maximum memory and disk resources for your SQL Server Workloads in the cloud.

 

Leave a Reply

Your email address will not be published.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

Warwick

I am a Microsoft Data Platform MVP as well as a Microsoft Certified Master working as the Principal Consultant here at SQL Masters Consulting. When I am not working with the SQL Server Stack I like to get away to the Snow and spend time Snowboarding.

Search