In building client/server applications with CA Gen, we generally adhere to a design structure known as Distributed Processing. With Distributed Processing, you have a Client procedure and a Server procedure, which are typically deployed on different machines. In actuality, you will likely have many client procedures and many server procedures, with the clients being deployed as a group on one or more client machines, and the servers being deployed separately or as a group across one or more server machines. CA Gen generates all the code for both programs, as well as any code necessary to support the communications between the two programs.
For Java (J2EE) web applications, Client procedures are generated as a combination of Java Server Pages (JSPs), Hyper Text Markup Language (HTML), JavaScript, and Cascading Style Sheets (CSS), and are deployed to a Web server. Server procedures can be generated as C, COBOL, or Java (EJBs) depending on the environment.
The following example illustrates these and other possible relationships:
Another characteristic of a distributed process system is that the DBMS typically resides on the same machine as the CA Gen Server. Because the elementary processes contain the business rules and most of the data access, performance is generally improved when the Server procedures implement the elementary processes. The Client procedure's logic can then concentrate on presentation, workflow, and some data validation.
|
Copyright © 2013 CA.
All rights reserved.
|
|