A common question I get is what size datastore would you recommend? The answer it depends .. on your requirements. I am going to walk you through a sizing example to show how I would go about calculating a datastore based on requirements.
First, we need to define certain requirements that we will need.
Recovery Point Objective (RPO): How much data an organisation can afford to lose
Recovery Time Objective (RTO): How quickly an organisation wants to recover from an outage
We also need details on the backup solution an organisation is using, to be more specific, its data restoration speed.
Let me illustrate with an example:
A client has a tape backup solution LTO-5 which has restoration speed of 140 MB/s which equals approximately 500 GB/Hour
If the organisation has a maximum of RTO of 2 hours, and your restoration speed is 500GB/Hour, that means the maximum size of datastore should not exceed 1TB (500GB/Hour x 2Hours=1TB).
If you have to do a fully recovery within the 2 hours RTO, you will be able with your 500GB/hour restoration speed.
The above calculation is enough for basic datastore sizing. However .. there is an extra calculation if you wish to be efficient ..
We need to consider the requirements for the Virtual Machine size on this datastore. Lets assume requirements is to have 40GB VMs running on this datastore. How many 40GB VMs can you fit in 1TB datastore? 1TB Datastore ÷ 40GB VMs = 25 VMs.
We also need to allow some headroom for VM SWAP or snapshots. So recommendation is to allow 20% cushion.
Final size of datastore: 1TB x 80% = 800GB
Final number of VMs on datastore: 800GB ÷ 40GB = 20 VMs
I hope the above calculation help you answer the question: How large should be my datastore?
Thank you for reading.