Having a custom domain would get your far enough in this case, owning your own IP address isn't really needed.
Just as I move my domain around between email (and other) providers, I could then move my IPv6 block around. Instead of asking Linode for an IP address out of their block, I could tell them "serve this block." (Or can we already do that? For a reasonable price?)
If you control the DNS there's no real need for owning your own IP addresses as you can just change what they point to. That's usually also much faster to do in a pinch than find a new party that will announce your block and convince the old one to stop doing so. That's usually hours if not days in lead time.
Though if you truly care enough about it you'll also want your own DNS servers and own gTLD in order to own the whole chain.
However what you describe -- building out your own infrastructure -- still remains a lofty but worthy endeavor for anyone with the skills and determination.
Businesses of all sizes rent stable IP addresses. Some ISPs in some countries also provide them to non-business subscribers as part of the standard subscription.
It's really not too difficult to have a persistent IP address for an individual who only needs reachability and is not trying to run a high traffic, global enterprise. Avoid the popular large, high-volume providers. Obviously do not use something like AWS.
1. Misconfigured servers: This is the hardest part. Rule number 1: keep it as small as possible. Disable everything, and then only enable what you need. Rule 2: READ, I mean read every document, blog post etc you can find about e-mail, email attacks, etc.
2. IP address: The normal address you get from Digital Occean or Linode for example, are normally horrible as a lot of spam has been sent from them. If your lucky it might work, but check the lists for those IPs. Try to change if you can, until you find one with low scores. Also, I dont know if this helped, but i had my domain on both Gmail and Outlook servers for a while before migrating to my own servers. I dont know if this helped me to not get stuck in their filters etc? Or I was just lucky i guess :)
I finally moved because I did not have the energy to keep up with the server etc, and when i noticed that i had not even logged on to the server for 3 months, i just moved it to gmail. I am however looking at migrating away from google, as i don't like the web interface (use it from the iPad now). So who knows where I'll end up next.
Email dates back to timesharing, when people shared computers, each user having an "email account" on a single computer.
For people today, each with their own computer, it's still possible to do email without store and forward, with each user running SMTP servers that talk directly to each other instead of connecting to "email providers".
This is grassroots. Not meant to leverage existing "email provider" ecosystem.
What's America's largest ISP? I think Comcast? Good luck passing SMTP in either direction over Comcast's residential network: https://customer.xfinity.com/help-and-support/internet/list-...
I agree with your general point, but it's hard to keep up with this sort of thing, even for geeks. For the Muggles it's impossible.
You are sending email to people you know, on a port you have agreed to, who also have reachable IP addresses and will accept connections from your IP address.
ISP's are not going to do DPI on open ports like 22, 80, or 443 or any high port, in order to filter out SMTP traffic. Why would they bother?
If you must use the existing "email provider" system -- e.g., you need to send commercial email -- then you must use the standardized ports they use, and yes, you must jump through their hoops. And you will experience spam and other nuisances. It's supposed to be open and decnetralized, but the reality is it's quite centralized because everyone relies on "major email providers".
But there's nothing about SMTP protocol that requires anyone to use an "email provider" or any specific port, nor to accept email from all IP addresses.
Two people can communicate directly via SMTP on a non-standard port and can choose to filter by IP address.
Maybe no one does this anymore, but I've done it; it's possible.
It's not that difficult to set up. The showstopper is not the set up, nor the ISP filtering port 25. It's that most people do not have a reachable IP address.
When you have your own, then you can run your own "services", not to mention many other possibilities, such as true decentralized communications -- no third party servers needed.
1. "World-reachable" means you control the firewall, not a third party.
Alternatively, be satisfied and keep depending on third parties.
Google: "Let us make many consecutive requests to your webserver, download and store copies of your content." You: "Agreed." You: "Let me make many consecutive reuqests to your webserver, download and store copies of my own or other people's content." Google: "Sorry".
Then Google makes your content available to others, but only for a few searches at time[2]. To add insult to injury, it sells advertising and, if the regulators are correct, promotes its own websites -- all serving user-generated content -- as people are searching through publicly available information.
"Middleman" or "service"?
2. Any attempt to collect in bulk results in IP ban. Compare Internet Archive who encourages users to crawl their collection.
Organizing public information sounds like a lovely goal for a company, until you realize it will be access controlled (it's public information), laden with ads (because the company cannot support itself without advertisers), user tracking for the advertisers and that the changes to the algorithms that include or exclude results are secret. And there will be no index. You can only access the data through search.
Moving beyond the first page of results is OK only so long as you do not go "too fast" or go too deep -- you will never get to page 100 witout first being banned. WTF?