Putting it all together
We've covered a lot of foundation. This is important as we need to get the theory right so we know that we understand reality correctly, because it will match the theory. You should have the expected result or baseline.
The way you perform capacity management changes drastically once you take into account performance and availability. Let's consider an example to drive the point. We'll take storage, as it's the easiest example:
- Say your physical SAN array has 200 TB usable capacity.
- It supports 1000 VMs, which take up 100 TB.
- It has 100 TB usable capacity left. That's plenty of space.
- Assuming an average VM occupies 100 GB of space, you can fit almost 1000 VMs!
My question is this: how many additional running VMs can that array support?
1000? 500? 1?
You are right—it depends on the VM IOPS. Just because there is space does not mean the array can handle the workload.
Now, what if the existing 1000 VMs are already experiencing high latency? Users...
