Site Overlay

Just how Does the web Work?

Just how Does the web Work?

Just how Does the web Work?




Where to start? Web Addresses

Protocol Stacks and Packets

Marketing Infrastructure

Web Infrastructure

The internet Routing Hierarchy

Domain names and also Address Resolution

Internet Protocols Revisited

Application Protocols: HTTP and also the World wide Web

Application Protocols: Electronic Mail and SMTP

Transmission Control Protocol

Web Protocol

Wrap Up




Just how does the web work? Great question! The internet ‘s development is now intense which appears to be not possible to get away from the bombardment of’s watched continuously on tv, seen on radio, and also observed in magazines. Simply because the web has become such a big section of the lives of ours, a great understanding is necessary to utilize this brand new tool most effectively.

This whitepaper explains the main technologies and infrastructure which make the Internet succeed. It doesn’t go into amazing depth, but protects enough of every area to give a basic understanding of the concepts involved. For virtually any unanswered questions, a summary of materials is offered at the conclusion of the paper. Any comments, questions, suggestions, etc. are prompted and could be directed to the author at


Where to start? Web Addresses

Simply because the web is a worldwide community of computer systems each computer attached to the Internet should have a distinctive address. Internet addresses are located in the form nnn.nnn.nnn.nnn in which nnn should be a selection from zero – 255. This address is generally known as an IP address. (IP means Internet Protocol; much more on this particular later.)

The picture below illustrates 2 computers attached to the Internet; the computer of yours with IP address and a different pc with IP deal with The web is represented being an abstract item in between. (As this paper advances, online part of Diagram one is explained and redrawn several times since the details of the Internet are exposed.)


Diagram one

Diagram one


In case you hook up to the web through an Internet Service Provider (ISP), you’re generally given a short lived IP address for the length of your dial in session. In case you hook up to the web from a neighborhood area system (LAN) the computer of yours may have a lasting IP address or maybe it may get a temporary one from the DHCP (Dynamic Host Configuration Protocol) server. At any rate, in case you’re connected towards the Internet, your computer has a distinctive IP address.

Check it out – The Ping Program If you are making use of A flavor or microsoft Windows of Unix and also have a link with the Internet, there’s a neat system to determine if a laptop on the web is in existence. It is called ping, possibly after the audio made by older submarine sonar systems.1 If you’re using Windows, begin a command prompt windowpane. When you are making use of a taste of Unix, reach a command prompt. Type ping The ping application is going to send a’ ping’ (actually an ICMP (Internet Control Message Protocol) echo demand note) to the named system. The pinged computer system is going to respond with a reply. The ping application is going to count time expired until the reply arrives back again (in case it does). Furthermore, in case you enter a website (i.e. rather than an IP standard address, ping is going to resolve the domain name and display the computer ‘s IP address. More on domain names and also address resolution later.


 Packets and protocol Stacks So your personal computer is attached to the Internet and has a distinctive address. Just how does it’ talk’ to various other computer systems attached to the Internet? A good example should work here: We need to say the IP address of yours is and you are looking to send out a message to the pc The idea you wish to send is “Hello computer system!”. Certainly, the message should be transmitted over whatever wire type connects the computer of yours to the Internet. Suppose you have dialed into your ISP coming from home and the message should be transmitted over the telephone line. Therefore the note should be interpreted from alphabetic text into electric signals, transmitted throughout the Internet, then translated back into alphabetic textual content. How’s this done? Through the usage of any protocol stack. Every computer system needs one to speak on online as well as it’s normally included in the computer ‘s os (i.e. Windows, Unix, etc.). The protocol stack utilized on the web is refered to as the TCP/IP process stack due to the 2 leading communication protocols used. The TCP/IP stack looks as this:


 Protocol Layer Comments

Application Protocols Layer Protocols particular to uses for example WWW, FTP, e-mail, etc.

Transmission Control Protocol Layer TCP directs packets to a specific application on a computer with a port number.

Internet Protocol Layer IP directs packets to a certain pc with an IP address.

Hardware Layer Converts binary package information to network signals and returned.

(E.g. ethernet community flash card, modem for telephone lines, etc.)


 If we were following the route that the information “Hello computer!” gathered from our pc to the pc with IP tackle, it will occur something as this:


Diagram two

Diagram two


The message would begin at the upper part of the process stack on the computer of yours and work it is way downward.

If the information to be routed is very long, each stack level which the information passes through might separate the idea up into little chunks of information. This’s because data delivered within the Internet (and most computer networks) are submitted manageable chunks. On the internet, these chunks of information are referred to packets.

The packets would proceed through the Application Layer and also continue towards the TCP layer. Each package is given a port number. Ports is describe afterwards, but suffice to point out that some plans might be going with the TCP/IP stack as well as sending emails. We have to know which system on the destination pc has to get the idea since it’ll be listening on a certain port.

After going via the TCP level, the packets proceed to the IP layer. This’s exactly where each packet gets it’s location address,

After our communication packets have a port quantity as well as an IP address, they’re prepared to be sent over the internet. The hardware level takes care of turning our packets containing the alphabetic copy of the message of ours into electric indicators as well as transmitting them through the telephone line.

On another end of the telephone line your ISP has an immediate link with the Internet. The ISPs medialink wireless router examines the spot address in each package and also determines where to post it. Frequently, the packet ‘s subsequent quit is another router. More on routers and also Internet infrastructure later.

Ultimately, the packets achieve computer Here, the packets begin at the bottom part of the spot computer ‘s TCP/IP stack as well as function upwards.

As the packets go upwards through the stack, everything routing information that the sending computer ‘s stack included (such as IP street address and port number) is removed from the packets.

When the data reaches the top of stack, the packets are re-assembled into the original form of theirs, “Hello computer!”

Marketing Infrastructure

Now you know exactly how packets journey from one computer to the next over the internet. But what is in between? What really compensates the Internet? Let us consider another diagram:


Diagram three

Diagram three


Right here we see Diagram one redrawn with much more detail. The physical relationship through the telephone system on the Internet Service Provider may have been simple to guess, but beyond that may bear some explanation.

The ISP keeps a pool of modems for their dial in customers. This’s handled by some kind of computer (usually a separate one) which controls information flow out of the modem pool to some backbone or maybe specific line router. This installation might be refered to being a port server, as it’ serves’ permission to access the system. Billing and usage info is generally collected here too.


After your packets traverse the telephone networking along with your ISP’s local tools, they’re routed upon the ISP’s backbone or maybe a backbone the ISP purchases bandwidth from. From here the packets often journey through many routers and over a few backbones, committed lines, along with other networks until they find the destination of theirs, the pc with address But would not it’d be good in case we knew the exact route our packets have been taking throughout the Internet? As it seems, there’s a way…


Check it out – The Traceroute Program If you are making use of A flavor or microsoft Windows of Unix and also have a link with the Internet, here’s one neat Internet program. This you’re named traceroute which shows the road your packets are shooting to a particular Internet destination. Like ping, you have to make use of traceroute originating from a command prompt. In Windows, use tracert From a Unix timely, type traceroute Like ping, you might in addition enter IP addresses rather than domain names. Traceroute is going to print out a summary of most of the routers, computer systems, and every other Internet entities your packets must go through getting to the destination of theirs.


If you are using traceroute, you will discover your packets must travel through many points to get to the destination of theirs. Most have extended names such as and These’re Internet routers that choose where to post the packets of yours. Several routers are revealed in Diagram three, but only a couple of. Diagram three is intended to exhibit a simple community structure. The web is a lot more complicated.


 Web Infrastructure

The internet backbone consists of numerous massive networks which interconnect with one another. These large networks are called Network Service Providers or maybe NSPs. Several of the massive NSPs are UUNet, PSINet, SprintNet, BBN Planet, IBM, CerfNet, and also others. These networks peer with one another to exchange packet visitors. Each NSP is necessary to hook up to 3 Network Access Points or maybe NAPs. At the NAPs, package website traffic could go through a single NSP’s backbone to the next NSP’s backbone. NSPs likewise interconnect at Metropolitan Area Exchanges or maybe MAEs. MAEs provide exactly the same objective because the NAPs but are privately owned or operated. NAPs had been the first Internet interconnect points. Both MAEs and NAPs are called Internet Exchange Points or maybe IXs. NSPs also market bandwidth to smaller networks, like Smaller bandwidth providers and isps. Below is a picture demonstrating this particular hierarchical infrastructure.


Diagram four

Diagram four


This’s not really a true representation of a real slice of the Internet. Diagram four is just intended to show just how the NSPs might interconnect with each smaller and other ISPs. None of the bodily system elements are revealed in Diagram four as they’re in Diagram three. This’s because an individual NSP’s backbone infrastructure is an intricate design by itself. Most NSPs use maps of the network infrastructure of theirs on the web sites of theirs and could be found very easily. To bring a real map of the Internet will be nearly impossible due to it is sizing, complexity, and actually changing structure.


The internet Routing Hierarchy

So how do packets discover their way throughout the Internet? Does every laptop attached to the Internet know where other computer systems are? Do packets just get’ broadcast’ to each pc on the web? The solution to both preceeding concerns is’ no’. No computer knows exactly where any of the other computer systems are, and packets don’t be delivered to every computer. The info used to get packets for their destinations are contained in routing tables maintained by each router linked to the internet.

Routers are packet changes. A router is generally attached between networks to path packets between them. Each wireless router is aware of it is sub networks and which IP addresses they utilize. The wireless router generally does not understand what IP addresses are’ above’ it. Examine Diagram five below. The black colored boxes linking the backbones are routers. The sizable NSP backbones at the top part are hooked up at a NAP. Under them are a number of sub networks, and also under them, a lot more sub networks. At the bottom part are 2 local area networks with computer systems attached.


Diagram five

Diagram five


When a packet comes at a router, the router examines the IP street address placed there by the IP process coating over the originating computer. The router determines it’s routing dinner table. If the system with the IP address is discovered, the package is delivered to that system. If the system with the IP address isn’t discovered, then the router transmits the package on a default path, typically up the backbone hierarchy to the subsequent router. Perhaps the following router is going to know where to send out the packet. If it doesn’t, once again the packet is routed upwards before it gets to a NSP backbone. The routers linked to the NSP backbones keep the biggest routing tables as well as below the package is routed to the appropriate backbone, exactly where it is going to begin its journey’ downward’ through more compact as well as smaller sized networks until it discovers it’s location.


Domain names and also Address Resolution But imagine if you do not understand the IP address of the computer system you wish to link to? What if the you have to get into a web server called Just how does the web browser of yours know exactly where on the Internet this particular computer lives? The answer to all these questions is the Domain Name Service or maybe DNS. The DNS is a sent out data source which keeps track of computer ‘s labels as well as their corresponding IP addresses within the Internet.

Many computer systems attached to the Internet host with the DNS database and the application which allows others to get into it. These computer systems are called DNS servers. No DNS server has the entire database; they only have a subset serotonin. If a DNS server doesn’t have the website requested by an alternate computer system, the DNS server re directs the requesting pc to another DNS server.


Diagram six

Diagram six


The Website Service is organized as a hierarchy much like the IP routing hierarchy. The pc requesting a name resolution is going to be re-directed’ up’ the hierarchy until a DNS server is discovered that could solve the website in the petition. Figure six illustrates a percentage of the hierarchy. At the upper part of the tree would be the domain roots. Several of the older, more widespread domains are seen close to the top. What’s not shown are the large number of DNS servers across the world which form the majority of the hierarchy.

When internet access is set up (e.g. to get a LAN or maybe Dial Up Networking in Windows), a single main along with 1 or even more secondary DNS servers tend to be specified together with the set up. This particular way, any Internet programs that require domain name resolution will have the ability to perform properly. For instance, when you get into a web address into the web browser of yours, the browser first links to your main DNS server. After acquiring the IP address with the website you entered, the internet browser then links to the target pc as well as requests the web site you wanted.


Check it out – Disable DNS within Windows If you are utilizing Windows 95/NT along with gain access to the internet, you might see your DNS server(s) and turn off them.

If you make use of Dial Up Networking: Open your Dial Up Networking window (which is usually discovered in Windows Explorer beneath your CD ROM drive as well as above Network Neighborhood). Right click the Internet connection of yours and click Properties. Near the bottom part of the connection qualities windowpane media the TCP/IP Settings… switch.


If you’ve a lasting link on the Internet: Right click Network Neighborhood and click Properties. Click TCP/IP Properties. Select the DNS Configuration tab in the pinnacle.


You ought to right now be checking out your DNS servers’ IP addresses. Below you might turn off DNS or even fixed your DNS servers to (Write bad your DNS servers’ IP addresses initially. You’ll most likely need to reactivate Windows as well.) Now go into addresses into the web browser of yours. The browser will not be able to solve the domain name and you’ll most likely get a nasty dialog package outlining that a DNS server could not be found. Nevertheless, in case you go into the corresponding IP address rather than the website, the web browser is going to be ready to access the sought-after net page. (Use ping to buy the IP address before disabling DNS.) Other Microsoft os’s are identical.



Internet Protocols Revisited

As hinted to earlier within the department about protocol stacks, one particular might surmise which you can get numerous protocols which are used on the internet. This’s true; there are lots of communication protocols required for online to run. These are the IP and TCP protocols, routing protocols, moderate access management protocols, application amount protocols, and so on. The following sections describe several of the more crucial and commonly used protocols within the Internet. Higher level protocols are discussed for starters, accompanied by reduced level protocols.

Application Protocols: HTTP and also the World wide Web Just about the most widely used products on the web is a World Wide Web (WWW). The software process which makes the internet tasks are Hypertext Transfer Protocol or maybe HTTP. Don’t mistake this together with the Hypertext Markup Language (HTML). HTML could be the dialect used-to create web pages. HTTP will be the process that web browsers and also web servers work with to speak with one another over the internet. It’s an application level protocol since it rests atop the TCP level in the protocol stack and it is utilized by certain programs to speak to each other. In this instance the apps are web browsers and also web servers.

HTTP is a connectionless textual content based protocol. Clients (net browsers) drive requests to net servers for web elements including images and web pages. After the petition is maintained by a server, the relationship between server and client across the web is disconnected. A new link have to be made for every request. Nearly all protocols are relationship oriented. It means that the 2 computers talking with each other keep the relationship open above the Internet. HTTP doesn’t however. Before an HTTP petition may be reached by a client, a brand new link should be put forth on the server.


When you kind a URL into a web browser, this’s what happens:


If the Url contains a website, the browser first links to a domain name server and also retrieves the corresponding IP standard address for the web server.

The web browser links to the internet server and directs an HTTP petition (via the protocol stack) for the preferred web page.

The net server receives the petition and checks for any preferred page. If the web page exists, the internet server directs it. If the server can’t get the requested page, it is going to send an HTTP 404 errors message. (404 means’ Page Not Found’ as anybody who has surfed the internet probably knows.)

The internet browser gets the page returned and the relationship is closed.

The internet browser then simply parses through the page and also looks for some other page elements it must finish the site. These typically include pictures, applets, and more.

For each element necessary, the browser makes extra connections and HTTP requests towards the server for each element.

When the web browser has finished loading each pictures, applets, etc. the web page is going to be totally packed in the internet browser window.

Check it out – Use your Telnet Client to Retrieve a page Using HTTP Telnet is a remote terminal system utilized on the internet. It’s pick has declined lately, though it’s a really helpful tool to learn the Internet. In Windows consider the default telnet program. It might be placed with the Windows directory called telnet.exe. When opened, pull on the Terminal menus and select Preferences. In the tastes windowpane, determine Local Echo. (This is really so you are able to determine your HTTP demand whenever you type it.) Now pull on the Connection menus and select Remote System. Enter for your Host Name plus eighty for the Port. (Web servers normally listen on port eighty by default.) Press Connect. However type

Get or HTTP/1.0


and press Enter two times. This’s a HTTP petition to a web server for it is root webpage. You need to visit a web page flash by then a dialog box must appear to let you know the hookup was lost. If you would want saving the retrieved page, switch on signing in the Telnet program. You might then browse through the page and watch the HTML that was utilized to create it.


 Most Internet protocols are specified by Internet papers widely known as being a Request For Comments or maybe RFCs. RFCs might be found at a few places on the internet. Check the Resources section beneath for suitable Url ‘s. HTTP model 1.0 is specified by RFC 1945.

Application Protocols: SMTP and Electronic Mail Another commonly used Internet assistance is electric mail. E-mail makes use of an application level protocol known as Simple Mail Transfer Protocol or maybe SMTP. SMTP is in addition a content based process, but in contrast to HTTP, SMTP is relationship oriented. SMTP is more complex compared to HTTP. You will find many more considerations and commands in SMTP than there’re in HTTP.

When you open your mail customer to look over the e-mail of yours, this’s what usually happens:


The mail customer (Netscape Mail, Microsoft Outlook, Lotus Notes, etc.) opens a link with it is default mail server. The mail server’s IP street address or maybe url is generally established whenever the mail customer is installed.

The mail server will transmit the first information to determine itself.

The customer is going to send an SMTP HELO command to that the server will answer with a 250 Ok email.

Based on whether the prospect is verifying mail, driving mail, etc. the proper SMTP commands will be delivered to the server, which will react appropriately.

This request/response transaction is going to continue until the prospect transmits an SMTP QUIT command. The server might say goodbye and also the connection is closed.

A simple’ conversation’ in between an SMTP customer and also SMTP server is found below. R: denotes messages delivered through the server (receiver) and also S: denotes messages delivered by the client (sender).

This SMTP instance shows mail delivered by Smith at host USC ISIF, Green, to Jones, and Brown at multitude BBN UNIX. Here we assume that

host USC ISIF contacts host BBN UNIX directly. The mail is

recognized for Jones as well as Brown. Green doesn’t possess a mailbox at

host BBN UNIX.




R: 220 BBN UNIX.ARPA Simple Mail Transfer Service Ready




S: MAIL FROM:<></>

R: 250 OK


S: RCPT TO:<></>

R: 250 OK


S: RCPT TO:<></>

R: 550 No this kind of user here


S: RCPT TO:<></>

R: 250 OK



R: 354 Start mail input; conclusion with<crlf>,<crlf></crlf></crlf>

S: Blah blah blah…

S:,..etc. etc. and so on.

S :.

R: 250 OK



R: 221 BBN UNIX.ARPA Service closing transmission channel

This SMTP transaction is had from RFC 821, which specifies SMTP.

Transmission Control Protocol

Under the application level within the protocol stack would be the TCP layer. When applications start a link with a different computer Online, the messages they send out (using a certain program layer protocol) get handed down the stack on the TCP layer. TCP is responsible for routing application program protocols to the appropriate program on the location computer. To do this, port volumes are used. Ports will be regarded as seperate channels on each pc. For instance, you are able to surf the internet while reading e-mail. This’s because these 2 applications (the internet browser as well as the mail client) used distinct port numbers. When a package arrives at a personal computer and also can make the way of its up the process stack, the TCP layer decides what application receives the package according to a port number.

TCP operates as this:


When the TCP layer gets the application level protocol information from above, it sections it within manageable’ chunks’ after which gives a TCP header with particular TCP info to each’ chunk’. The info found in the TCP header consists of the port quantity of the application the information has to be delivered to.

When the TCP layer gets a package from the IP level listed below it, the TCP layer strips the TCP header details in the packet, does several information reconstruction if necessary, after which sends the information on the proper program utilizing the port quantity taken from the TCP header.

This’s exactly how TCP routes the information going through the protocol stack on the proper program.

TCP isn’t a textual protocol. TCP is a connection oriented, dependable, byte stream program. Connection-oriented suggests that 2 programs using TCP have got to first build a relationship before exchanging data. TCP is solid because for each package gotten, an acknowledgement is delivered to the sender to verify the shipping. TCP also contains a checksum in it is header for error-checking the received data. The TCP header looks as this:


Diagram seven

Diagram seven


Notice that there is no spot for an IP address within the TCP header. This’s because TCP does not know something about IP addresses. TCP’s task is usually to get software level information from application to program reliably. The task of obtaining information from computer to computer could be the function of IP.

Check it out – Popular Internet Port Numbers Listed below will be the port figures for several of the more widely used Internet services.

FTP 20/21

Telnet 23



Quake III Arena 27960


 Web Protocol

Unlike TCP, IP is an unreliable, connectionless process. IP does not care whether a packet becomes to it is destination or not. Nor does IP be informed on connections plus port numbers. IP’s work is too transmit and path packets to other pcs. IP packets are impartial entities and could arrive out of order or perhaps not at all. It’s TCP’s duty to ensure packets arrive and are within the proper order. About the single thing IP has in frequent with TCP is the way in which it receives information and also gives it is very own IP header info to the TCP data. The IP header looks as this:


Diagram eight

Diagram eight


Above we come across the IP addresses of the receiving and sending pcs wearing the IP header. Below is what a packet is like as a result of passing thru the application layer, IP layer, and TCP layer. The application level data is segmented within the TCP level, the TCP header is put, the package will continue to the IP level, the IP header is put, after which the package is transmitted throughout the Internet.


Diagram nine

Diagram nine


Wrap Up

And now you know exactly how the internet works. But how much time will it remain this way? The version of IP now utilized on the Internet (edition four) only permits 232 addresses. At some point there will not be some totally free IP addresses left. Amazed? Do not care. IP version six has been tested at this time holding an analysis backbone by a consortium of investigation institutions along with companies. And after that? Exactly who understands. The web has come quite a distance since it’s inception as being a Defense Department research task. No one truly understands what the Internet will be. One particular thing is certain, however. The Internet will unite the world like absolutely no different mechanism ever has. The information Age is within full stride and I’m glad to become an aspect of it.

Rus Shuler, 1998

Updates made 2002



Below are a number of intriguing links regarding several of the subject areas discussed. (I wish they each still work. Most wide open in new window.) would be the website on the Internet Engineering Task Force. This body is considerably accountable for the development of The like and internet protocols. is the group accountable for administering domain names. is a good RFC search engine helpful for finding some RFC. exhibits animated maps of Internet latency. is Internet Weather from ClubNET. This web page displays package damage for several carriers. is Russ Haynal’s ISP Page. This’s a fantastic website with links to many NSPs plus their backbone infrastructure maps.



The next books are excellent resources and also helped considerably in the writing of the paper. I believe Stevens’ guide is the greatest TCP/IP reference ever and may be considered the bible on the Internet. Sheldon’s guide covers a much broader scope and possesses a great quantity of marketing info.

TCP/IP Illustrated, Volume one, The Protocols.

  1. Richard Stevens.

Addison-Wesley, Reading, Massachusetts. 1994.

Encyclopedia of Networking.

Tom Sheldon.

Osbourne McGraw Hill, New York. 1998.

Although not utilized for creating the newspaper, the following are a few additional excellent books on the subjects on the Internet as well as networking: Firewalls as well as Internet Security; Repelling the Wiley Hacker.

William R. Cheswick, Steven M. Bellovin.

Addison-Wesley, Reading, Massachusetts. 1994.

Data Communications, Open Systems and Computer Networks. Fourth Edition.

Fred Halsall.

Addison-Wesley, Harlow, England. 1996.

Telecommunications: Design and protocols.

John D. Spragins with Joseph L. Hammond and also Krzysztof Pawlikowski.

Addison-Wesley, Reading, Massachusetts. 1992.

Leave a Reply

Your email address will not be published. Required fields are marked *