It all starts at the server [World Wide Web and FastCGI]

Recently, the words “browser war” have filled the Internet headlines. Each browser comes up with new schemes to be faster and to support newer, more powerful standards. Sometimes we seem to forget that behind the browser is a server, and, just like any other client-server architecture, the server´s quality, functionality, robustness and efficiency are important to the well-being of the entire system. There is a technology that promises to combine the benefits of both the Common Gateway Interface (CGI) and the application program interface (API): FastCGI, brought to the Web world by OpenMarket. Each FastCGI program runs in its own process and can be accessed by the server in one of two ways: through a direct pipe if the Web server and FastCGI processes are running on the same machine, or through a TCP/IP connection if the FastCGI is running on a different machine. The FastCGI model seems ideal in terms of server-side extensions. It combines the safety of a separate process with server software independence, yet without the overhead of starting that separate process on every request