We start by reviewing how a user views a normal Git server. Git repositories use URLs as locators, so when a user clones, fetches from, or pushes to a remote repository, it is done using a suitable URL. Git URLs are not very different from any other URL, and the man page for git-clone has a section on them, so you can see all the syntax variations that you can use.

However, for authenticated remote access, there are really only two protocols that are of interest: SSH and HTTP. Of these, ssh-based access is a lot more common, because it is quick and easy to set up; even on a freshly installed Unix, there's usually no extra configuration required to get it working.

As the main page mentioned previously says an ssh URL has the general form ssh://[user@]host.xz...