2018 2018 Ericsson AB. All Rights Reserved. Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. Terminology terminology.xml
General Information

In the following terms that may cause confusion are explained.

The term "user"

A "user" is a term that everyone understands intuitively. However, the understandings may differ which can cause confusion.

The term is used differently in OpenSSH and SSH in Erlang/OTP. The reason is the different environments and use cases that are not immediatly obvious.

This chapter aims at explaining the differences and giving a rationale for why Erlang/OTP handles "user" as it does.

In OpenSSH

Many have been in contact with the command 'ssh' on a Linux machine (or similar) to remotly log in on another machine. One types

ssh host

to log in on the machine named host. The command prompts for your password on the remote host and then you can read, write and execute as your user name has rights on the remote host. There are stronger variants with pre-distributed keys or certificates, but that are for now just details in the authentication process.

You could log in as the user anotheruser with

ssh anotheruser@host

and you will then be enabled to act as anotheruser on the host if authorized correctly.

So what does "your user name has rights" mean? In a UNIX/Linux/etc context it is exactly as that context: The user could read, write and execute programs according to the OS rules. In addition, the user has a home directory ($HOME) and there is a $HOME/.ssh/ directory with ssh-specific files.

SSH password authentication

When SSH tries to log in to a host, the ssh protocol communicates the user name (as a string) and a password. The remote ssh server checks that there is such a user defined and that the provided password is acceptable.

If so, the user is authorized.

SSH public key authentication

This is a stronger method where the ssh protocol brings the user name, the user's public key and some cryptographic information which we could ignore here.

The ssh server on the remote host checks:

That the user has a home directory, that home directory contains a .ssh/ directory and the .ssh/ directory contains the public key just received in the authorized_keys file

if so, the user is authorized.

The SSH server on UNIX/Linux/etc after a succesful authentication

After a succesful incoming authentication, a new process runs as the just authenticated user.

Next step is to start a service according to the ssh request. In case of a request of a shell, a new one is started which handles the OS-commands that arrives from the client (that's "you").

In case of a sftp request, an sftp server is started in with the user's rights. So it could read, write or delete files if allowed for that user.

In Erlang/OTP SSH

For the Erlang/OTP SSH server the situation is different. The server executes in an Erlang process in the Erlang emulator which in turn executes in an OS process. The emulator does not try to change its user when authenticated over the SSH protocol. So the remote user name is only for authentication purposes in the Erlang/OTP SSH application.

Password authentication in Erlang SSH

The Erlang/OTP SSH server checks the user name and password in the following order:

If a pwdfun is defined, that one is called and the returned boolean is the authentication result. Else, if the user_passwords option is defined and the username and the password matches, the authentication is a success. Else, if the option password is defined and matches the password the authentication is a success. Note that the use of this option is not recommended in non-test code.
Public key authentication in Erlang SSH

The user name, public key and cryptographic data (a signature) that is sent by the client, are used as follows (some steps left out for clearity):

A callback module is selected using the options key_cb. The callback module is used to check that the provided public key is one of the user's pre-stored. In case of the default callback module, the files authorized_keys and authorized_keys2 are searched in a directory found in the following order: If the option user_dir_fun is defined, that fun is called and the returned directory is used, Else, If the option user_dir is defined, that directory is used, Else the subdirectory .ssh in the home directory of the user executing the OS process of the Erlang emulator is used. If the provided public key is not found, the authentication fails. Finally, if the provided public key is found, the signature provided by the client is checked with the public key.
The Erlang/OTP SSH server after a succesful authentication

After a successful authentication an Erlang process is handling the service request from the remote ssh client. The rights of that process are those of the user of the OS process running the Erlang emulator.

If a shell service request arrives to the server, an Erlang shell is opened in the server's emulator. The rights in that shell is independent of the just authenticated user.

In case of an sftp request, an sftp server is started with the rights of the user of the Erlang emulator's OS process. So with sftp the authenticated user does not influence the rights.

So after an authentication, the user name is not used anymore and has no influence.