.NET Remoting over TCP

Remoting in .NET is quite similar to RMI in Java. To begin with, we need an object that has the methods we would be executing on the server-side. This object inherits from MarshalByRefObject. This indicates that the client does not get a copy of the object but rather a reference to the remote object. We then have a hosting process and a client process.

In our example, the hosting process creates a new instance of the TcpChannel class (pass the TCP port number to the constructor of TcpChannel for the server hosting process) and registers the object using ChannelServices.RegisterChannel. Then, the class we are hosting has to be registered using the RemotingConfiguration.RegisterWellKnownServiceType method - it takes the type of our class as the first parameter, the registration name as the second parameter and the object mode (singleton & single call modes) as the third parameter. When creating a console application, we would usually add a Console.ReadLine statement at the end of the application to prevent it from exiting.

The client application would create and register an instance of TcpChannel. Then, obtain an instance of the class hosted on the server using Activator.GetObject - with the type of the class as the first parameter and the URL (Example: tcp://localhost:8080/my_service) as the second parameter. You can now call methods on the object as you would for a normal object.