AWS makes bringing an idea to life from proof-of-concept to full scale marketable solutions easy with their "Lego block"-like infrastructure components.
A project can go from a single EC2, to a few load balanced web servers, adding S3 storage as needed, then putting it all behind a CDN (CloudFront) and so on, and so on. As needed.
Inevitably though, we get the question:
"What IP should we whitelist?"
A quick Google search will return numerous stack exchange questions and about adding or mimicking a static IP in front of AWS services.
- https://security.stackexchange.com/questions/33616/how-to-whitelist-an-amazon-elb-in-any-firewall
- Where does Amazon publish the range of IP addresses used for EC2 instances by region?
- Using an IP Address to Point to Amazon S3 Storage (the sarcastic tone in rmalayter's answer prompted me to ask this)
I understand I'm not the only one questioning this IP obsession.
Whatever the scenario, there is usually a way, but what gets me, is that this often gets requested to please one client's IT department. It has nothing to do with improving the product.
(I can accept it when they require an IP address for our servers to interact with their API for example, but sometimes it seems they just want something to whitelist before they start using a service that is accessible to the general public)
Often this even hinders future architecture improvements since we can no longer freely add our AWS Lego blocks as needed.
Is it right to give in and find a way to get your static IP address?
Or should one dive into a rant about how 'the cloud' works in the hopes to confuse the "IT department" enough that they accept not having an IP address to whitelist? (this has worked before).
Is there an answer that can get the most "security aware" operations teams to back away while still being friendly.