Tutorials Client ServerJava builds client server systems with sockets. Sockets are the endpoints of two way communication between programs running in a network. They are software objects that connect applications to network protocols, so they become intelligible. For example, in UNIX a program opens a socket that enables it to send and receive messages from the socket. This simplifies software development because programmers only have to change or specify the socket, while the operating system is left in tact.
In client server models the server contains sockets bound to specific port numbers. Port numbers identify specific processes that are to be forwarded over a network, like the Internet, in the form of messages to the server. The server only has to monitor the socket and respond when a client requests a connection. Once connections have been made and bound to port numbers, the two endpoints, client and server, can communicate through the socket. Java sockets are client side sockets, known simply as sockets and server side sockets known as server sockets. Each belong to their own class within a Java package. The client initiates the socket, which contains a host name and port number, by requesting connections.
The server socket class in Java allows for servers to monitor requests for connections. As the socket communicates over the network, Java’s server socket class assigns a port number to each client in the server and creates a socket object or instance. This server socket instance creates a connection and produces a thread or string of processes through which the server can communicate with the client through the socket. Java web servers are built according to this model.
TCP (Transmission Control Protocol) works in conjunction with IP (Internet Protocol) to send messages between computers over the Internet. When communicating over the Internet, the client program and the server program each attach a socket to their end of the connection. Then the client and server can read from and write to the socket. Java provides operating system sockets that allow two or more processes to send and receive data, regardless of computer location.
The other method for using Java to build client server systems is RMI. RMI stands for Remote Method Invocation. By using Java language and functions, programmers write object oriented programming, so that objects that are distributed over a network can interact with each other. RMI allows for client objects in JVMs to request services through a network from another computer (host or server). Client objects included with the request may call upon methods in the remote computer that change the results. Methods are programmed procedures attributed to a class that are contained in each of its objects (or instances). It is a characteristic of object oriented programming.
Classes and, therefore, objects can have more than one method and methods can be used for more than one object. Responses then run as if they were local (on the same computer.) This passing back and forth of objects and methods attached to objects is called ‘object serialization’. Simply put, RMI requests call upon the method of a remote object. As previously stated, it uses the same syntax it would locally. To make this intelligible to the servers or sites being called upon requires three layers: a client side stub program, a remote reference layer, and a transport connection layer. Each request travels down the layers of the client computer and up the layers of the server.
The client side stub program initiates the request. Stub programs are small sections of programs containing routines from larger programs. It is a substitute for programs that may take too long to load or are located remotely on a network. They are provided by the server at the client’s request. Stubs accept client requests and communicate requested procedures (through another program) to remote applications. They also return the results to the client or requesting program. These stubs mimic the program being called for service.
In Java, stub programs are also referred to as ‘proxy’. Remote reference layers manage reference variables for remote objects, using the transport layer’s TCP (Transmission Control Protocol) connection to the server. Reference variables contain class data and therefore include methods. The transport layer protects and maintains end to end communication through a network. The most used transport layer is TCP. After the client requests pass through the transport layer, they pass through another remote reference layer before requesting implementation of the request by the server from a skeleton. These skeletons are written in high level IDL (Interface Definition Language). The server receives the request and sends the response along the same channels, but in the other direction.
First Page: Java and Client Server Models
