1.1 Prior Knowledge and Revision

1.1. PRIOR KNOWLEDGE AND REVISION 3 1.1 Prior Knowledge and Revision This topic assumes you already have some background knowledge of the Internet. You may have studied The Internet unit of Intermediate 2 Information Systems, or the Internet section of Standard Grade Computing Studies, or you may simply know about the Internet through your own reading and experience. Before you study this topic, you should be able to describe what the Internet is explain the difference between the Internet and the World Wide Web describe some of the services provided by the Internet, including web pages electronic mail file transfer describe how a user can connect to the Internet Check your understanding by trying the revision questions: Revision Q1: The Internet a) is a local area network (LAN) b) is a wide area network (WAN) c) is a network of networks spread across the world d) is a world-wide network, controlled by the US military Q2: The World Wide Web (WWW) a) is the common name for the Internet b) is a vast collection of information stored and accessed through the Internet c) is the international network used by the Internet for transmitting data d) is the central network of the Internet, to which users must connect Q3: File transfer a) is only possible on local area networks b) is not allowed over the Internet c) is one of the services provided by the Internet d) is only possible using the World Wide Web Q4: An ISP a) is an organisation that provides access to the Internet b) is a device used to connect a computer to the Internet c) is an address used to identify a computer attached to the Internet d) is a user s e-mail address

4 TOPIC 1. HOW THE INTERNET WORKS Q5: www.thisisagoodsite.org.uk/index.html a) is an e-mail address b) is an IP address c) is a MAC address d) is a URL 1.2 Introduction In this first topic, you will review the basic structure of the Internet, and begin to learn about how it all works - all the hardware and software right across the world, made up of many different types of system, yet they can all work together. The Internet began in 1969, when 4 US military computers were connected together to form ARPANET, using 50kbps transmission lines. A few years later, some University computers were connected, and the growth of the Internet began. By 1984, there were over 1000 hosts (computers linked to the network), and 4 years later, over 60,000. In 1991, the World Wide Web (WWW) was "invented" by Tim Berners-Lee of CERN, making it possible to store and access information easily, with hyperlinks between resources. This new Internet service boosted the continuing expansion of the Internet, which passed the 1,000,000 hosts landmark in 1992. There is no sign yet of this growth slowing down. The latest figures (July 2004) indicate over 285 million Internet hosts, and growing. The Internet is not a single massive network. It is a network of networks. What binds them together is the use of the TCP/IP protocol. This means that many different types of computers on different types of networks, running different types of software, can all communicate. Internet Growth 15 min Take a few minutes to look in more detail at how the Internet has grown since it began in 1969. You will find useful information at: http://www.zakon.org/robert/internet/timeline/ (or one of the many other websites which describe the history and growth of the Internet). Make a list of key events - try to select between 6 and 12 of the most important ones, and summarise them in a table like this: Date Event 1969 4 US computers linked to form ARPANET 1971 Extended to 15 hosts

1.3. STRUCTURE AND TOPOLOGY 5 The Internet in diagrams Browse the following website, to get an impresiion of how extensive and complex the Internet has now become: http://www.cybergeography.org/atlas/atlas.html Look particularly at the diagrams in the following sections: 20 min Geographic Cables & satellites Census Topology ISP maps Historical Choose any 2 diagrams which you found particularly interesting or surprising, and write a very brief description of what they show. 1.3 Structure and topology The Internet is a network of networks. The backbone consists of high bandwidth connections between major hosts, which may belong to government agencies, military organisations, commercial companies or academic institutions. Some of these act as Internet Service Providers (ISPs), which means that they allow other users to access the Internet by connecting to their servers. An individual computer may be connected to an ISP s server, perhaps using a dial-up modem over a telephone line, or a broadband connection. Many smaller companies, schools, offices and other organisations with several computers, connect their LAN (local area network) to the Internet through a router to the ISP server. Remote Office Local Area Network 1 I.S.P. Local Area Network 3 Internet Local Area Network 2 Traveller Remote EDI & FTP Site Local & Wide Area Networks using Internet backbone

6 TOPIC 1. HOW THE INTERNET WORKS The high speed broadband connections between the main Internet nodes form a mesh topology, like Figure 1.1. Figure 1.1: This mean there are many routes between any two nodes on the network. This makes the Internet very reliable. Even if one node or one link fails, then the whole system will continue to work. 1.4 Internet hosting The nodes in the mesh network are called Internet hosts. A host is simply a computer system which provides a service. Hosts are also called servers. The Internet operates on a client/server basis. An Internet user accesses the Internet using his or her own computer system, running suitable software - this might be a web browser, an FTP client, or some other type of communication software. The user requests a resource (perhaps a file or a web page) using the client computer. A message is sent to the appropriate host or server, which responds by sending the requested resource back to the client. Any computer system connected to the Internet can act as a host, providing services to any other Internet users. As a result of this system, Internet resources are not localised in only one place. It is a truly distributed system. Many resources are hosted on multiple hosts. For example, a shareware application may be downloaded from any one of many

1.5. HOW DATA IS TRANSMITTED 7 host servers, known as mirror sites. Each host consists of A computer Server software An operating system which includes the TCP/IP protocols, such as Unix, Windows NT, Linux or Mac OS Because the host uses the TCP/IP protocol to handle all communication with clients, the user will be completely unaware of the type of computer or operating system being accessed. 1.5 How Data is Transmitted There are two main methods which can be used to transmit data across a mesh network - these are circuit switching and packet switching. 1.5.1 Circuit switching Circuit switching is where there is a direct connection established between two networks or two computers. All data follows the same physical path during the communication. Circuit switching is expensive because it means dedicating a particular connection for the time that the communication is occurring. Asynchronous Transfer Mode (ATM) is a popular circuit switching technology. ATM is often used to connect networks over a leased telephone line. This is expensive, particularly if large distances are involved, but provides a high bandwidth connection.

8 TOPIC 1. HOW THE INTERNET WORKS 1.5.2 Packet switching Packet switching is a system where the data to be exchanged is broken up into blocks of data called packets. These packets are given a destination address and a sequence number, and this information is used to re-assemble the communication when it reaches its destination. Breaking the data down into packets means that packets may follow different routes between the transmitter and the receiver, but is extremely efficient because packets from different users can be mixed, and the network hardware can decide on the most efficient transmission route. Packet switching is usually much cheaper than circuit switching because it does not need a dedicated connection. Most Internet traffic uses packet switching. The difference between circuit and packet switching On the Web is an animation that shows the difference between circuit and packet switching. You should now look at this animation. 1.5.3 Packets and TCP/IP For most data transmission, the Internet uses packet switching, with all files split up into chunks of data called packets during transmission. This process is controlled by the TCP/IP protocol. TCP/IP is really a combination of two protocols - TCP (Transmission Control Protocol) and IP (Internet Protocol) - which work together to ensure reliable transmission of data across the Internet maze. TCP is responsible for splitting the data into packets and adding a header to each one. This header includes a sequence number which allows the file to be recreated at the receiving end by re-assembling the packets in the correct order. TCP is also responsible for retransmitting any packet if it does not receive an acknowledgement that the packet has been received. The IP protocol is responsible for taking each packet and adding its own header to allow the packet to be routed correctly. As it is perfectly possible for packets to follow different routes to their destination, they might easily take different amounts of time to arrive and therefore be out of sequence when they do. As long as the packets can be re-assembled in the correct order by TCP using the sequence numbers which were added when the packets were created, then the transmission will have been successful and TCP will issue an acknowledgment. 1.5.4 Packet switching Consider a file (for example, a shareware games program) being downloaded from an Internet host. TCP divides the file up into packets of around 1000 to 1500 bytes. Each packet is allocated a unique packet number which is added to the packet in a header.

1.5. HOW DATA IS TRANSMITTED 9 1/5 2/5 3/5 4/5 5/5 IP also adds the source and destination IP addresses, and the length of the packet to the header of each of the packets. It also adds a trailer to the end of each packet. This indicates the end of he packet and also contains error checking data. The packets are now sent out over the Internet. Different packets may follow different routes as shown in Figure 1.2. Step 1 2/5 1/5 Destination Source Step 2 2/5 3/5 1/5 Destination Source Step 3 2/5 1/5 5/5 4/5 3/5 Destination Source Step 4 5/5 1/5 2/5 3/5 4/5 Destination Source Step 5 5/5 1/5 2/5 3/5 4/5 Destination Source Figure 1.2: Eventually all the packets arrive at the destination. However, because the packets took different routes through the Internet, they have arrived in the wrong order. The web server at the destination address uses TCP to sort the packets into the correct order using the information in the packet headers.

10 TOPIC 1. HOW THE INTERNET WORKS 1.6 Routers and routing tables The hardware responsible for routing the data packets through the Internet are called routers. A router is a device which receives data packets on its input ports, reads the header to find out where the packet is to be sent, then sends it onwards towards its destination via one of its output ports. If both source and destination computers are connected directly to the same router, the router can simply send the packet to the destination computer. However, this is not usually the case. Often a data packet will need to pass through many routers between the source and destination. How does a router know where to send a packet? Routers use routing tables to decide where to send a packet next. A routing table is simply a list of destination addresses; if the router cannot send the packet directly to its destination, it consults the routing table, which lists the address of another router which is nearer the destination. Routing tables are usually dynamic; that means that they are constantly updated depending on network traffic. This is why different packets may be sent by different routes. Tracing a route 5 min It is possible to investigate the route taken by a packet using the ØÖ ÖØ command in Windows. To do this, type ØÖ ÖØ ÓÐ Öº Ûº ºÙ at the C: command prompt. This will then display the route taken by a packet across the Internet from your computer to the Scholar host. Your results should look something like Figure 1.3. Figure 1.3: In this example, the packet passed through 12 routers. The first of these is a local router on the LAN. The next two are at Pipex (the ISP being used to access the Internet). Routers 4 to 9 appear to be a route through the Internet backbone. Router 10 is at Interactive University, leading through routers 11 and 12 to the SCHOLAR host.

1.7. TCP/IP AND IP ADDRESSING 11 Try this from your computer, and compare routes to various destinations of your choice. 1.7 TCP/IP and IP addressing 1.7.1 Introduction to IP Addresses and the Domain Name Service An Internet Protocol (IP) address is the way the Internet Protocol distinguishes computers (or more specifically Network Interfaces) that exist on the same network. On any one network, you cannot have two computers sharing an IP address, as this would make it impossible to distinguish between them when trying to send data from one computer to another. You can however have a single computer with two network Interface Cards (NIC)s fitted, each one with its own IP address. In this case the computer would be connected to two different networks. Routers are computers like this. All computers that are connected together on the Internet (or on a private LAN) need unique IP addresses. Because IP addresses are difficult to remember, we use Domain names to identify them instead. When we connect to a remote computer using a domain name such as google.co.uk, a system known as the Domain Name Service (DNS) translates the domain name into an IP address (216.239.41.100 in this case) so that the correct computer can be found. DNS On the web is an animation that shows how DNS operates. animation now. You should view this 1.7.2 The structure of an IP address An IP address is a 32-bit number separated in to four 8-bit parts called octets. The four parts are like a postal address, with the difference that the detail is ordered from right to left instead of left to right. The first number of the IP address is the most general and the last is the most specific. An IP address can be divided into 2 sections, - the network identifier and the host identifier. There are three ways to split the IP address. Class A Class B Class C nnn.hhh.hhh.hhh nnn.nnn.hhh.hhh nnn.nnn.nnn.hhh where n = network identifier, h = host identifier The network identifier is used to route packets to the correct destination network. Once a packet reaches the network which requested it, the network part of the address is ignored and the packet is routed to the correct host using the host identifier. 1.7.3 Classes of IP address Since each computer on the Internet needs a different IP address, there has to be some way of sharing out the IP addresses so that large companies and organizations have

12 TOPIC 1. HOW THE INTERNET WORKS one for each of their computers, while letting smaller organizations have some as well. Since there are a small number of large organizations and a large number of small organizations, ranges of IP addresses can be allocated by giving large organisations blocks of Class A addresses and small organisations blocks of Class B or Class C addresses. Class A addresses have their first octet in the range 1 to 126. Class B addresses have their first octet in the range 128 to 191 and Class C addresses have their first octet in the range 192 to 223. A very large company with very complex internal networks may be allocated a Class A address block such as 115.*.*.*. This would give a total of about 16 million possible IP addresses. There are only 127 Class A address blocks, and no more are to be allocated. Class B addresses are common for large companies, allowing a block of around 65000 IP addresses. When a Class B IP address is allocated, (say 135.113.*.*), the first two numbers identify that company network. The company can give any IP address in that block to any computer on their network. Class C addresses are the third type, giving 254 possible IP addresses for any one block (0 and 255 are reserved for particular functions). Here, the first three octets are specified, and the remaining field is allocated by the owner of the address. Networks that are directly connected to the Internet are usually connected to an Internet Service Provider (ISP) via a full time connection (such as a cable or leased line) The ISP informs the network administrator which IP addresses can be used on the network, and a router controls the delivery of packets to the appropriate IP address on the network. Small ISPs typically have 1 or 2 Class C licenses, providing 250 to 500 IP addresses. When you dial up an ISP with a modem, your computer is temporarily allocated an IP address. This will be in the range of the Class C licences that they own. If you want to know more about classes of IP address have a look at the following links: http://www.networkclue.com/routing/tcpip/index.php http://compnetworking.about.com/library/weekly/aa042400b.htm

1.7. TCP/IP AND IP ADDRESSING 13 1.7.4 Class Summary Class Number of Hosts per Network Range of First Networks Octet A 126 16,777,214 1 to 126 B 16,384 65,534 128 to 191 C 2,097,152 254 192 to 233

14 TOPIC 1. HOW THE INTERNET WORKS Some IP addresses are reserved and cannot be used. For example, network address 127 is reserved to allow testing of network cards. Octets cannot be set to 255, as this value is used to set up a network broadcasts. Octets cannot be set to 0 (all packets which have an IP address which contain a 0 octet are restricted to the local network and not passed to the Internet). 1.7.5 Gateways, Masks and Subnets Within a class B network, all stations IP addresses will have the same first and second octet (which defines the network uniquely) but different 3rd and 4th octets (to define the actual station). For example: station 1 145.217.105.113 station 2 145.217.99.136 station 3 145.217.73.108 The network router uses a mask to seperate the network from the host identifier. In this case, it would be 255.255.0.0. The network mask for a class A network is 255.0.0.0, and for a class C network it would be 255.255.255.0. The router performs a "bitwise AND" between the address of an incoming packet and the network mask to identify the destination hosts. Administrators can use subnet masks, which look like 255.255.XXX.0 where XXX is between 0 and 255 to split a class B network into several sub-nets. The same can be applied to a class A or class C network. A router which provides access from one network to another is called a gateway. The router s network card has its own address, call the gateway address. If a packet is sent to a destination which is not on the network, it is sent to the gateway address. The router then routes the packet out of the network towards its correct destination. 1.7.6 Limitations of the IP address system The Class A B and C system of allocating IP addresses is very inefficient, particularly where Class A and B addresses are concerned as many allocated IP addresses may remain unused. With more and more devices being connected to the Internet there is

1.8. UNIFORM RESOURCE LOCATORS 15 concern that the world will run out of IP addresses. There are two possible solutions to this problem. One is to extend the number of IP addresses using a 6 octet system known as IPv6. IPV6 is a system of IP addressing which among other improvements, increases the IP address size from 32 bits to 128 bits, making a possible maximum number of 3.4X10 38 addresses The other solution is to dispense with the IP Class system and use Classless Inter Domain Routing (CIDR) uses IP addressing space more efficiently by making it possible to allocate part of a block to a network instead of the whole block. 1.8 Uniform Resource Locators IP addressing is fine for routers and electronic devices, but it is not a very userfriendly system for humans. Instead of having to remember IP addresses like 62.189.14.202, the domain name service (DNS) allows resources to be identified by a uniform resource locator (URL). 1.8.1 URL Structure Here is a typical URL http://www.fredtheplumber.com/business/info/contact.html It consists of 4 elements: element example purpose protocol http defines the protocol to be used to interpret the resource domain name www.fredtheplumber.com this is the unique identifier of an Internet host computer path /business/info/ defines where the resource is to be found within the directory structure of the system file identifier contact.html the filename of the resource Every URL has this same structure. Interpreting URLs Identify the protocol 5 min the domain name the path the file identifier

16 TOPIC 1. HOW THE INTERNET WORKS for each of the following URLs: Q6: http://www.dbethune.com/computing/higher.html Q7: http://scholar.hw.ac.uk/heriotwatt/scholarlogin.asp Q8: ftp://ftp.freebsd.org/pub/freebsd/readme.txt Q9: https://www.jobs.ed.ac.uk/jobs/index.cfm 1.8.2 Extended URLs Some URLs also include additional parameters. The commonest of these is a port number. Here is an example: http://www.squigglegraphics.com:8080/manage/index.html This might be required if a computer with a single unique IP address is hosting more than one server. The port number identifies that the request is to a particular server hosted on the system, rather than the default server. An example would be accessing a mail server on an Intranet. 1.9 The Domain Name System A Domain Name is the unique name that identifies an Internet site. Domain Names always have 2 or more parts, separated by dots. The part on the left is the most specific, and the part on the right is the most general. For example in the domain name: my company.com my company refers to a specific company and.com refers to the commercial domain. Examples of three letter top level domains are: Top level Domain Meaning.com Commercial business, a company.net Network provider, Internet Service Provider.gov Governmental agency.edu Educational institution.org Non-profit institution.mil US. military.int International organisation Two letter top level domains are allocated to countries other than the United States. Examples are:

1.9. THE DOMAIN NAME SYSTEM 17 Top level Domain.uk.fr.di.au.jp Country United Kingdom France Germany Australia Japan Note that my company.com and my company.co.uk are treated as different domains, but may point to the same Internet site. Two or more domain names may point to the same machine, but a domain name can only ever refer to one machine. Many companies will register their names in a number of top level domains so that people can find them easily, and so that they do not get their company name hijacked by competitors. If you want to know more about top level domains have a look at he Internet Assigned Numbers Authority (IANA) website http://www.iana.org/domain-names.htm It is possible for a Domain Name to exist but not be connected to an actual machine. This is often done so that a group or business can have an Internet e-mail address without having to establish a real Internet site. In these cases, some real Internet machine must handle the mail on behalf of the listed Domain Name. Name resolution is the system of mapping a domain name to its IP address. This is done on the Internet by the Domain Name Service. The domain name system on the WWW is regulated by an organisation called InterNIC. If an company wishes to have a presence on the WWW then the first step is usually to pay an accredited registrar a to register one or more domain names for them. In the UK the organisation responsible for overseeing the registration is Nominet. Once that has been done, the next step is to pay to host a website. The company hosting the website will provide details of the name servers which will resolve the domain name to an IP address and this information needs to be given to the registrar. Once this information has propagated around the WWW, typing the domain name into a browser will result in the name being resolved to an IP address and the appropriate web page should be accessible. Many companies offer an all-in one service of registration and hosting, and many will build company websites as part of this service as well. You can find out more about Nominet from this link. http://www.nic.uk/ You can find out more about registering domain names and InterNIC from this link http://www.internic.net/faqs/authoritative-dns.html 1.9.1 Domain Name Service (DNS) The Domain Name Service (DNS) sometimes called the Domain Name System translates the names which we use to identify hosts into an IP address. e.g. scholar.hw.ac.uk has the IP address 62.189.14.202. The DNS database is stored on a hierarchy of dedicated servers. (The DNS database

18 TOPIC 1. HOW THE INTERNET WORKS is the foundation of all communication on the Internet, so it is very important that this service is available at all times) When you request a web page on the Internet, a piece of software called the DNS resolver (usually built into the network operating system) first contacts a DNS server by sending it a UDP packet in order to determine the server s IP address.(obviously the resolver needs to know the IP address of the DNS server in order to do this) If the DNS server does not contain the information needed, it will in turn forward the request to a DNS server at the next higher level in the hierarchy. This continues until the correct IP address is returned to the machine requesting the file. Name servers are arranged in a hierarchy, with a top level domain for each country together with the six domains: edu, com, gov, mil, org and net. Each of these domains is sub-divided, with each sub division maintaining a name server (usually at least two, so that if one fails there is still another available to supply the relevant information) On a very small scale within a LAN, each computer may have a hosts file which is a text file which maps IP addresses on the LAN to host names. A hosts file may read something like this: 127.0.0.1 localhost 192.168.0.1 gateway 192.168.0.2 intranet 192.168.0.3 mailserver Early in the history of the Internet a central authority called the Network Information Centre maintained a table like this, with a list of all address to name bindings. This table was emailed out to all hosts on the Internet every time it was modified. The system became far too cumbersome once the number of hosts on the Internet grew, and certainly would not be practical nowadays. Hosts files like this can still be useful on a small network if they use a slow connection to the Internet. If you place commonly used hosts such as Google or Yahoo in your hosts file, then this will speed up access times as your machine will not need to contact a Domain Name Server in order to get the IP address of these commonly used sites. Using the Domain Name Service Use the following address to look up the IP address of some popular websites. http://www.zoneedit.com/lookup.html Using nslookup You can use the nslookup command from a command prompt to get a domain name from an IP address, or an IP address from a domain name. In this example the users commands are in bold with the response following

1.10. SUMMARY 19 nslookup www.microsoft.com Server: resolver2.svr.pol.co.uk Address: 195.92.195.95 Name: www2.microsoft.akadns.net Addresses: 207.46.134.189, 207.46.249.29, 207.46.245.92, 207.46.249.189 207.46.245.156, 207.46.134.221, 207.46.134.157, 207.46.249.221 Aliases: www.microsoft.com, www.microsoft.akadns.net nslookup 207.46.134.189 Server: resolver2.svr.pol.co.uk Address: 195.92.195.95 Name: origin2.microsoft.com Address: 207.46.134.189 1.10 Summary the Internet is a world wide network of networks the Internet uses a mesh topology an Internet host is a computer system connected to the Internet that provides resources for users data is sent across the Internet using packet switching files are split up into packets, and assigned headers using the TCP protocol the packets are routed across the Internet by routers, using the IP protocol the packets are re-assembled at their destination routers make use of dynamic routing tables to decide where a packet should be sent next each computer on the Internet has a unique IP addresses of the format 62.189.14.202 there are 3 classes of IP addresses: class A, B and C class A addresses use 8 bits to identify the network and 24 bits for the device, allowing 126 networks each with up to 16 million devices attached class B addresses use 16 bits to identify the network and 16 bits for the device, allowing 16,384 networks each with up to 65,534 devices attached class C addresses use 24 bits to identify the network and 8 bits for the device, allowing 16 million networks each with up to 254 devices attached sub-net masks are used to mask out sections of the IP address which are not relevant to the situation

20 TOPIC 1. HOW THE INTERNET WORKS the structure of a URL is protocol://domain name/path/file identifier some URLs include an additional parameter (usually the port number) the URL is resolved into its corresponding IP address by the Domain Name System (DNS) 1.11 End of topic test An online assessment is provided to help you review this topic.