There are multiple factors that may not be reflected in AWS Cost Calculator but this is a standard tooling for checking the cost of machines or other AWS services.
https://calculator.s3.amazonaws.com/index.html
I have done calculation for 4 Boot Nodes running in Sydney region.
The items has been explained on the screenshot above.
Note:
Regions can have different pricing for machines.
Instance size can be adjusted without any change any time we need. AWS best practices is always start with smal size. So we can start t2xLarge or t2x XLarge. The important parameters about them are : CPU, Memory and Network.Cloud Providers Give you a credit limit for them e.g. CPU Credit, and once you finished it, you will be slow down. You can have a look at them here: https://aws.amazon.com/ec2/instance-types/
Boot nodes with IP Address in chain spec needs to have a reservered public IP, in AWS its called EIP (Elastci IP Address), as long as they are attached to Running nodes - no cost - otherwise we need to pay for them. If we are going to use DNS (similar to Polkadot Chain spec) then we don't need them otherwise we need to use them.
I am looking to find a best practice for instance sizing, but if you know that from Substrate or Polkdato team share the link with me.