Because embedded systems almost always serve Web pages that contain dynamic, or real-time, information, this networking tutorial begins with Rabbit and TINI examples that serve Web pages with dynamic content. Following the examples is an introduction to using HTTP and other protocols in serving Web pages.

Two Approaches

A Web browser such as Microsoft’s Internet Explorer is a client application that uses HTTP to request Web pages from servers on the Internet or in a local network. The servers don’t have to be PCs or other large computers. Even a small embedded system with limited memory can serve a page containing text and simple images, including pages that display real-time data and accept and act on user input. A browser provides a user interface for requesting and displaying pages. The computers that request Web pages typically have full-screen displays, but for some applications, an embedded system with limited display capabilities can function as an HTTP client. If the requested pages are very simple, even a text-only display of a few lines might suffice. Or an embedded system might receive and process the contents of a Web page without displaying the page in a browser at all. This networking tutorial focuses on Web servers. With an Internet connection, a Web server can serve pages to any browser on the Internet. Or a server may be programmed to respond to requests only from specific IP addresses. A Web server in a local network may serve pages to selected computers or to any computer in the local network.

An embedded system that functions as a Web server generally has all of the following:

• Non-volatile memory to hold pages to be served.

• Support for TCP and IP. Requests for Web pages and the pages sent in response travel in the data portion of TCP segments.

• Support for HTTP. The server must be able to understand and respond to received requests for Web pages. The HTTP standard specifies the format for the requests and replies.

• A local-network or Internet connection. To serve pages on the Internet, the Web server must have an Internet connection. Any firewalls must be configured so the system can receive HTTP requests, as described in Chapter 10.

• One or more pages to serve. The Web pages are files or blocks of text that use a form of encoding called hypertext markup language (HTML).

The HTML encoding specifies the formatting of text and images on the page, including text size and fonts and the positioning of text and other elements on the page. The HTML code may include links to images that appear on the page, as well as links to other pages or resources. In serving a Web page with dynamic content, the software must have a way of inserting the dynamic content as the page is being served. A variety of protocols and technologies can work along with HTTP and HTML to enable a server to provide Web pages that contain real-time data and respond to user input. This networking tutorial includes two approaches to serving real-time data, and Chapter 7 covers ways that Web servers can respond to user input.

Serving a Page with Dynamic Data

Many Web pages are static, where the information on the page doesn’t change unless someone edits the page’s HTML file and uploads the new file to the server. Static Web pages are useful for presenting product information, articles, or other information that remains constant. But most embedded systems have little use for static pages, other than possibly presenting a home page with links to other pages. An embedded system that functions as a Web server will almost certainly want to display real-time information such as sensor readings or other up-to-the-minute information about the processes or environments the system is controlling or monitoring.

This section shows how the Rabbit and TINI modules introduced in Chapter 3 can serve Web pages that display dynamic data. Dynamic, or real-time, data includes any data that can change over time and can be different each time the page is served. An obvious example is a counter that displays the number of times the page has been accessed. Dynamic data may also include sensor or switch readings and time and date information. The supporting code included with the Rabbit and TINI (and additional sources in the case of Java servlets on the TINI) greatly reduces the amount of the programming required to serve Web pages with dynamic content. The dynamic data served by the example applications in this networking tutorial consists of a message that displays the amount of time the system or application has been up and running. Network article 6-1 shows an example page. The embedded system stores the number of days, hours, minutes, and seconds in variables. When serving the page, the server application inserts the current values of the variables in the appropriate places in the page. You can use the same techniques to create Web pages that display the current values of any variables in a system. Although the result is the same, the Rabbit and TINI examples use different approaches to achieve the result. The Rabbit uses Server Side Include directives that instruct the server to insert the values of variables in the appropriate locations in the file being served. For the TINI, instead of storing the Web page in a separate file, the application creates the Web page as it’s being sent, using a series of writes to send the page’s contents to a TCP socket and inserting the values of variables in the designated locations in the page.