I know about Postfix MTA configuration concepts: there’s tree of them and the last one is most-appropriate only because it doesn’t require new Unix user to be created for new mailbox creation.
As a system hosts more and more domains and users, it becomes less
desirable to give every user their own UNIX system account.
How many Unix users is too much?
Ahh, this hearkens back to the good old days, when Unix systems were (nearly) the exclusive domain of universities and large enterprises, and Linux – with a version number starting with 0 – was almost unheard of. Back in these days the Web was just getting started; the Internet was email, Usenet, FTP, IRC, and not a whole lot more.
Everyone who needed to have access to the Internet would be given an account on one or more of the Unix systems, and often log on to it from a green screen dumb terminal. They would typically have “full” access to a shell, along with the email service they were expecting, and most simply used it as such.
Though, for security reasons, some places did not grant a full shell, and just dumped people into the program they were expected to use, such as an email client, factory application to run the conveyor belts, whatever. But the shell was still lurking in the background, and some of these programs had ways to get to it…
When Internet access became more common on regular PCs and Macs in the late 1990s, people began downloading email via the POP3 and IMAP protocols directly to their home computers. These people (usually) wouldn’t know what to do with a shell if they had one, and didn’t really need one to get their email and Web access anyway. So mail servers started implementing “virtual” users, for whom mail was handled but there was no corresponding Unix account.
This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.