Side_1_360

page. The links at the pages are said to use

hypertext.

The client is located at the user. The browser

fetches the page requested, interprets the text

and formatting commands that it contains, and

displays the page to the user. Pages may contain

text, images, sound or video tracks, etc. The

browser may be assisted by a helper application

to be run with the page as input. Some helper

applications contain interpreters for particular

languages, enabling downloading and running

programs from WWW pages.

The way a browser fetches a page on a server is

to establish a TCP connection to the server and

then send a request on that connection.

An illustration of the protocols is given in Figure

23. Depending on the medium and traffic flows,
    other protocols could also be involved.

Every WWW server has a process listening to

TCP port 80 for incoming connections from

clients (browsers). After a TCP connection has

been established the server receives the request

and a response is sent. This is usually done by

the HyperText Transfer Protocol (HTTP). An

address of the page is commonly in a Uniform

Resource Locator (URL) format. One example is

http://www.telenor.com/Telektronikk/TraffEng.

html. This consists of three parts: httpthe name

of the protocol; http://www.telenor.comthe server

where the page is located; and Telektronikk/

TraffEng.htmlbeing the name of the file contain-

ing the page (in case the file name is omitted, a

default file is assumed, like index.html). This

shows that the user has to know where the page

is located; that is the server name. To support

referencing to pages without knowing their loca-

tion (as well as supporting replication) work on

Universal Resource Identifiers (URIs) has been

undertaken.

The DNS is used for translating the server name

to the corresponding IP address. This allows the

browser/client to establish a TCP connection on

port 80 to the proper server. Then an HTTP mes-

sage ‘Get’ is sent identifying the file (Telektron-

ikk/TraffEng.htmlin the example above). The

server returns this file and the TCP connection is

released. The browser displays the information

in the page. Several TCP connections may be

established to fetch different parts of the page,

like a text frame, an image frame, and so forth.

WWW pages are written in a language called

HyperText Markup Language (HTML), includ-

ing text, images, links to other pages, etc. By

HTML can be described a “static” layout of

WWW pages. Java can be applied in order to

allow for more rapid interactions. Java borrows

many ideas from C and C++ languages, although

it is not fully compatible with any of them. The

main idea is that a page can point to a Java pro-

gram called an applet. When the browser re-

ceives it, the applet is placed on a client machine

and executed. This allows for animation and

sound as well as simplifying forms, and so forth.

The applets developer writes the applet in Java

and then compiles it into byte code. Then the

browser needs to understand the applet and a

byte code interpreter is needed in the client sys-

tem. Such an interpreter may also be called a

Java virtual machine. Thus, if a new data format

is stored at a site, the only thing needed for

someone to download and view the data is to

also fetch the applet allowing the client to view

the data. This may also go for protocols, which

can be written in the Java language and loaded

dynamically when needed.

Plug-ins are dynamically loaded code modules

that become part of the browser’s code. This is a

flexible way to add functionality to a browser.

Some plug-ins are placed in a special directory

into which the browser looks. However, plug-ins

can also be downloaded from a server when

needed. Each plug-in commonly handles a few

document types. This means that the plug-in is

not using memory as long as no document of

that type has been loaded.

A complete screen as observed by a user may

consist of a number of pages so the browser may

decide the sequence of fetching such pages. Text

pages can normally be downloaded quite rapid-

ly, also keeping the user occupied, while the

more data voluminous pages are transferred, like

images. Thus the user may also decide to abort

the transmission if the information already avail-

able does not warrant the wait for the rest of the

information. To support this, most browsers

have a status line displaying which step they are

currently in.

Some browsers support the use of local disk to

cache pages that have been fetched. Then, before

Figure 23 Some protocols
used for Web surfing

Web

server

Web

client

HTTP HTTP

TCP TCP

IP IP

lower

layers

lower

layers