The ssh command has a number of options, and I don’t plan to cover all of them. Even the SSH documentation warns against the use of some of them, suggesting they are only for real experts. But I want to mention the ones that I think you will find important. These options take the form of switches in the command: -1 : Forces the connection to use SSH v.1 protocol only. The question here is why would you want to do that if you have SSH v.2 available. It is a real improvement, after all. -2 : Forces the connection to use SSH v.2 protocol only. -4 : Forces ssh to use IPv4 addresses only. -6 : Forces ssh to use IPv6 addresses only. -b : Bind address. Useful for machines that have two IP addresses, such as systems with two NICs. This tells SSH which IP address on the local machine to use for the connection. -L : Specifies that the given port on the local (client) host is to be forwarded to the given host and port on the remote side. This works by allocating a socket to listen to port on the local side, optionally bound to the specified bind_address. -p : Port to connect to on the remote host. This can be specified on a per-host basis in the configuration file. -R : Specifies that the given port on the remote (server) host is to be forwarded to the given host and port on the local side. This works by allocating a socket to listen to port on the remote side, and whenever a connection is made to this port, the connection is forwarded over the secure channel, and a connection is made to host port hostport from the local machine. -v : Verbose mode. This shows all commands and replies, and is useful for debugging. -W : Requests that standard input and output on the client be forwarded to host on port over the secure channel. Works with v.2 only. -X : Enables X11 forwarding. But note that this can open a vulnerability.

La commande ssh possède pas mal d'options, et je ne prévois pas de les lister toutes. Même la documentation SSH avertit contre l'utilisation de certaines, suggérant qu'elles sont réservées aux vrais experts. Mais je tiens à mentionner celles que vous trouverez importantes selon moi. Ces options prennent la forme de commutateurs à la commande :

-1 : Force la connexion à utiliser seulement le protocole SSH v.1 ; la question ici est pourquoi voudriez-vous faire cela si vous avez SSH v.2 disponible. C'est une réelle amélioration, après tout. -2 : Force la connexion à utiliser seulement le protocole SSH v.2. -4 : Force ssh à utiliser seulement des adresses IPv4. -6 : Force ssh à utiliser seulement des adresses IPv6. -b : adresse de liaison. Utile pour les machines qui ont deux adresses IP, deux cartes réseau. Indique à SSH quelle adresse IP utiliser sur la machine locale pour la connexion. -L : Indique que le port donné sur l'hôte local (client) doit être translaté sur l'hôte et le port spécifiés de l'autre côté. Cela fonctionne par l'attribution d'un « socket » pour écouter sur le port du côté local, éventuellement lié à l'adresse de liaison spécifiée. -p : Port sur lequel se connecter à l'hôte distant. Cela peut être spécifié selon l'hôte dans le fichier de configuration. -R : Indique que le port donné sur l'hôte distant (serveur) doit être translaté sur l'hôte et le port spécifiés du côté local. Cela fonctionne par l'attribution d'un « socket » pour écouter sur le port du côté opposé, et chaque fois qu'une connexion est faite sur ce port, la connexion est transmise sur le canal sécurisé et une connexion est établie sur l'hôte et le port depuis la machine locale. -v : Mode verbeux. Affiche toutes les commandes et les réponses, et est utile pour le débogage. -W : Demande que l'entrée et la sortie standard sur le client sont transmises à l'hôte sur le port via le canal sécurisé. Fonctionne en v.2 seulement. -X : Permet la translation X11. Mais notez que cela peut ouvrir une vulnérabilité.

Port Forwarding One of the handy things you can do, and something useful for tunneling. is port forwarding over SSH. The basic idea is to connect via ssh to a remote machine, and ask it to send something to a specific port other than the default port. The basic way you do this is to use the SSH command with the appropriate flags, -L and -R, which, not surprisingly, stand for Local and Remote. You need to specify the port you want to use, and what will be forwarded to it. • Local Port Forwarding – This takes a port on your local machine and forwards it to a specified port on the server. So you can make a request on a local port like 7280 on address 127.0.0.1, and your SSH client would intercept that call and send it to port 119 on the server. Then you would have a secure connection to get whatever port 119 is configured to serve (typically Usenet traffic, but this is just an example). So you use this to configure your newsgroup client to securely grab messages from a public server, assuming it allows SSH connections. • Remote Port Forwarding – This is the reverse of Local Port Forwarding. Here, the idea is to specify a port on the remote server and have it forwarded to your local server. This is not very common, and you may never need to do this. Essentially, all traffic coming in to the server on the specified port would then be forwarded to your local machine. • Dynamic Port Forwarding – This creates a SOCKS proxy and is not restricted to one port or one type of traffic.

Translation de port

Une des choses pratiques que vous pouvez faire, et d'utile pour le « tunneling ». est la translation de port sur SSH. L'idée de base est de se connecter via ssh sur une machine distante, et lui demander d'envoyer quelque chose sur un port spécifique autre que le port par défaut. La méthode de base pour faire cela est d'utiliser la commande SSH avec les indicateurs appropriés, -L et -R, qui, sans surprise, signifient Local et Remote ([NdT : distant]). Vous devez spécifier le port que vous souhaitez utiliser, et ce qui lui sera translaté. • Translation de port local - Ceci prend un port sur votre machine locale et le translate vers un port spécifié sur le serveur. Ainsi, vous pouvez faire une requête sur un port local comme 7280 sur l'adresse 127.0.0.1, et votre client SSH interceptera cet appel et l'enverra au port 119 sur le serveur. Ensuite, vous aurez une connexion sécurisée pour obtenir ce que le port 119 est configuré pour envoyer (généralement le trafic Usenet, mais cela est juste un exemple). Donc, vous pouvez utiliser ceci pour configurer votre client de « newsgroup » pour qu'il récupère en toute sécurité des messages à partir d'un serveur public, en supposant qu'il permet les connexions SSH. • Translation de port distant - Ceci est l'inverse de la section précédente. Ici, l'idée est de spécifier un port sur le serveur distant et qu'il soit translaté vers votre serveur local. Ce n'est pas très fréquent, et vous n'en aurez peut-être jamais besoin. Essentiellement, tout le trafic entrant sur le port spécifié du serveur sera ensuite transmis à votre machine locale. • Translation de port dynamique - Ceci crée un proxy SOCKS et ne se limite pas à un port ou un type de trafic.

Local Port Forwarding Suppose you are at work (or school), and you just cannot bear to miss out on your Facebook stream. But there’s a filter stopping you from accessing the site. However, for the sake of argument, you could create an SSH connection to a server outside the network (which could be your computer at home). You could then do something clever using Local Port Forwarding. Create a connection as follows: ssh -L 7280:facebook.com:80 address of home machine Now, your home machine does need to have a public IP address, or you would need to set up your router to forward the traffic, for this to get through. Once you have done this, you would open your browser and set it to go to http://localhost:7280, and traffic would then flow to your home machine, and from there to Facebook. You can now browse to your heart’s content on your Facebook stream. Of course, this also illustrates why network admins might want to shut down SSH traffic, which they might do by blocking any outbound traffic going to port 22 (the default SSH port). And you could then try changing the default port on your home SSH server to something other than port 22, and then the admins could do deep packet inspection, and so on. But SSH Port Forwarding is not just a matter of a security breach in the making, it can be used very legitimately in a number of situations. For example, you have a company with a number of geographically dispersed locations. In that case, SSH Port Forwarding would be a very useful way to connect sites to exchange data. You might have a database server that employees might need to connect to, and don’t want that traffic flowing through the Internet unsecured. Or perhaps you have set up a server for yourself, such as OwnCloud, and it is in a remote hosting center. Creating an SSH connection and using Port Forwarding might make your data a lot more secure.

Limitations There are a few things you need to watch out for. One is that not all ports may be available to you. If you are in a Unix-like environment, for instance, port 1024 and all ports below that can only be used by root. But any port above 1024 should be usable by a user with normal privileges as long as no one else is already using it. The other thing you need to remember is that if the connection is dropped the port forwarding is gone. And, in general, TCP connections are configured to close after a period of inactivity, and on some firewalls that can be as little as 300 seconds (5 minutes). This can be controlled by a rule (or perhaps more than one) in your iptables, or directly by /proc/sys/net/ipv4/tcp_keepalive_time. But, if you want a persistent connection, you need to use a Keep Alive.

Keep Alives There are two, basically. One is the TCP Keep Alive, which is simple but spoofable, and the other is the SSH keepalive, also called serveralive. Serveralive messages travel through the encrypted connection between you and the server, and thus cannot be spoofed. Assuming security is your reason for creating SSH connections, then it would be more secure to use serveralive messages, though I expect using TCP keepalives is far from the worst thing that could happen. In Linux, you can set this up either for everyone (if you have root privileges), or just for yourself, by editing the appropriate config file. For everyone – edit /etc/ssh/ssh_config and insert: Host * ServerAliveInterval 300 ServerAliveCountMax 2

For just you, edit ~/.ssh/config, and add the above code to it. ServerAliveInterval, specifies how often a null packet should be sent to the server to keep the connection alive. However, sometimes the server may go off or drop the connection, so the second line specifies how many times you should send a packet without getting a response. The setting I have shown will send a packet, and if no response is received, it will send a second packet 300 seconds later. If no response is received to the second consecutive packet the connection will be dropped by your client. On Windows, using PuTTY, there is a good explanation at http://blog.hazaveh.net/2013/10/keep-ssh-session-alive-in-putty/, but essentially you go to Connection, and then on the right under Sending of null packets to keep session active, you can set Seconds between keepalives (0 to turn off) to 300 seconds to get a similar result. Once you understand Port Forwarding and Keepalives, you are most of the way to tunneling.