> The only rationale we can reasonably come up with for AWS’s egress pricing: locking customers into their cloud, and making it prohibitively expensive to get c

> The only rationale we can reasonably come up with for AWS’s egress pricing: locking customers into their cloud, and making it prohibitively expensive to get customer data back out. So much for being customer-first. Presuming the average AWS customer is running an API SaaS, said customer’s API response load (= egress bandwidth) will tend to be a good measure of how much “useful work” said SaaS is doing for its own clients — either in the form of API responses, requests to third parties, emails sent, notifications pushed, uploads to other providers, etc. As such, egress bandwidth tends to be a good measure of having product-market fit, and thereby of willingness-to-pay. Things that trigger mostly ingress load — dropping attacks at a firewall, accepting uploads, receiving emails/notifications, etc. aren’t as emblematic of product-market fit (they aren’t things you would traditionally charge customers for “by volume”), but rather are usually activities done as loss-leaders; and so do not correlate to willingness-to-pay.