AWS free tier data transfer expansion

Nobody is actively encouraging the entire industry to consider optimization as an afterthought. This makes zero sense. Why would an organization pay a business to reduce their AWS costs if it wasn't worth the cost to realize the savings? Cost optimization is an easy task, as you've stated - so the cost/value proposition of a business like The Duckbill Group must not be worth it according to your statement. Yet they exist and do, seemingly, well. Maybe... Just maybe, cost optimization in AWS is not easy, not straightforward, and designed to be painful enough to where smart engineers are incented to leverage Amazon's dark patterns of hiding costs at time of deployment.

You even state...

> 80% of what he does is easily discovered a few clicks away from the AWS home page.

So why is cost optimization a constant point of conversation with AWS if it's so easy? Why do outfits like Digital Ocean advertise on the notion of clear billing as a positive differentiator compared to AWS?

> Cloudflare have a solid sales pipeline, but they're a sitting duck if any of the big clouds ever decide to replicate the business model like-for-like.

Let's take a stroll back in time. Do you think that Amazon and AWS have always posted a profit? Go look, they've posted many quarterly losses to get where they're at. That's how it works as you build a business like that. To your point - AWS does compete directly with Cloudflare in certain products, yet here we are, Cloudflare and AWS both continue to grow (negative operating income / net income are not a direct correlation of company growth BTW). A mistake you've made is around brand and reputation. Nobody thinks of AWS as a security company. Customers continue to buy Palo Alto Networks, Fortinet, Zscaler and, yes, Cloudflare - even though AWS offers some overlapping portfolio. Why? AWS isn't viewed as a security portfolio. Cloudflare has brand reputation in security and content distribution. And it's a pivot that easily works with both their brand and reputation.

> Outside some sweet Javascript cold start hacks they basically have no moat, and there are far more situations that could send the company into desperate measures than otherwise.

This just screams of the competitive argument low-road. I don't have ties to Amazon or AWS. I'm not an employee. When I read statements like this it's affirmation that there's some agenda. I laughed out loud reading that as the closing argument, thanks for that.

This doesn't feel to me like a generous enough statement about all of the finer points.

Would i be okay with my servers no longer responding to web requests (or doing so with reduced port speeds with some of the above hosts, e.g. Time4VPS) in case of spending limits being unexpectedly hit? Yes, definitely, since my projects going dark for a bit or slowing down would be preferable to me ending up not being able to pay my rent and having to rely on public outcry about the large bill and the vendor's mood that day towards letting that one slide.

Would i be okay with my servers running out of space and no longer writing new data to any database, merely responding with the appropriate errors instead? Yes, definitely, because that probably signals the fact that for some reason lots of space has suddenly started getting used for no reason, something that happened so fast that i didn't even get to plan the appropriate scaling, once Zabbix would warn me about 80% of the disk being full (or any monitoring tool would have an alert set up for this). This would probably also be yet another sanity check.

Would i be okay with my servers suddenly ceasing to exist and being wiped from existence, for any reason short of excessive abuse complaints or repeated ToS violations? I most certainly wouldn't want this to happen, yet in my experience it has never actually happened - some of the vendors listed allow you to pre pay for the resources you're about to use (e.g. Time4VPS and Contabo), even offering discounts for longer term reservations much like AWS does, without unexpected charges to you like AWS would do. In contrast, some of the other vendors listed allow you to pay based on hourly usage (Hetzner, Scaleway, Vultr, DigitalOcean, at least IIRC), but also won't have unpredictable pricing spikes because of set limits for the most part (ingress/egress charges may still apply, however, a worrying trend in the industry that muddies the waters), if i pay for a 5$ VPS every month, that's what i can expect to pay most months regardless of usage (assuming 100% uptime).

With these factors in mind, my servers being wiped would probably have to happen due to either me misusing those services, or alternatively me failing to pay the bills even with the more predictable pricing structures, which is actually very much like a person's electricity being turned off due to them not paying for it - as unpleasant as that'd be, no surprises there. Personally, i don't subscribe to the belief that managed services with unpredictable billing are the only way to do software nowadays, something that Richard Stallman coined as SaaSS (Services as a Software Substitute) and about which you can read more here: https://www.gnu.org/philosophy/who-does-that-server-really-s...

Alas, as long VPS uptime with predefined resources is the unit of computation that you're paying for, everything else should be fairly simple from there onwards.