1 Answers
Very perceptive question, you’re spot on with that assumption. We can dig into that a little bit. Since you’ve got some background, I’ll use a bit of tech speak here. For anyone who might be reading this, and isn’t immediately familiar with the language, don’t worry. As long as you remember the concept of scaling up (bigger mugs) as opposed to scaling out (more mugs), you’re spot-on track.
This video mostly describes the concept of scaling, rather than the mechanism. Where in a traditional hypervisor, you can manage your resources for a Virtual Machine dynamically, such as adding an extra CPU, or a few extra GB of RAM, most cloud providers are a lot less flexible here.
Thinking of it from the Hypervisor Admin perspective, having a farm the scale of a Public Cloud Provider like Azure pretty much prevents that sort of thing from happening effectively. The Hypervisor needs to be able to allocate resources in pre-determined chunks, like "2 CPU’s and 4GB RAM". Adding 2GB of RAM would just be far too difficult in that model. For this reason, Cloud Providers like Azure offer resource configurations for your instances, something called "Sizes" or sometimes "Series" in Azure, or "Instance Types" in AWS and GCP. Lars covers this in a later lecture.
So, what happens if you need to scale up from one configuration to another? You’ve probably already guessed it; you need to use a new Virtual Machine, and set up the application on there.
It’s certainly not the cleanest experience, and as you can guess from the fact about instance sizes, there are limits to how well scaling up works in the Cloud. You can definitely get tools to make this easier, but by default, the best approach in the cloud is usually scaling out. But since many applications don’t support scaling out, scaling up remains an option.
Hope this helps. The further you dive into AZ-900, the more this will hopefully become clear. Likewise, you might find value in the AZ-103 course, covering Azure Administration around Infrastructure. Best of luck with your studies!