1. Prepared by: Eng. Ali H. Elaywe1 Arab Open University - AOU T209 Information and Communication Technologies: People and Interactions Twelfth Session – Part2

2. Prepared by: Eng. Ali H. Elaywe2  This part is based on the following reference: Book N: Networks Part2: Internet Protocols Reference Material

3. Prepared by: Eng. Ali H. Elaywe3  Ethernet is the protocol for LANs  TCP/IP (transmission control protocol and internet protocol) family of protocols is used for the Internet  For any part of the journey that crosses a LAN, TCP/IP packets are carried in the data field of Ethernet frames, a process called encapsulation  So why do we need TCP, IP and Ethernet? TCP, IP and Ethernet all occupy different layers in the OSI Reference Model TCP is at the transport layer IP at the network layer Ethernet covers both the data-link and physical layers Continue Part2 of Book N: Internet Protocols Topic 1: Introduction

4. Prepared by: Eng. Ali H. Elaywe4 Ethernet provides a best-effort delivery across local area networks using the CSMA/CD protocol, which works well within the confines of a LAN, which is after all essentially local, and uses local addressing The Internet Protocol (IP) uses network addresses that cover the entire Internet, and uses these to determine if a packet is local or should be sent on via a router IP still provides a best effort service, but one that extends over the entire global reach of the Internet For many of the applications that are run over the Internet it is not good enough to rely upon a best effort service, after all in a global Internet you are much more likely to encounter network problems!! TCP provides a reliable service on top of IP Continue

5. Prepared by: Eng. Ali H. Elaywe5  Activity 38 (exploratory) The header of an Ethernet frame contains various fields dedicated to specific tasks. If I told you that IP uses packets to send data across wide area networks, which two fields from the Ethernet frame would be needed to route an IP packet, and why? The two most important fields for the routing of packets are the destination and source addresses The destination address can be used by a network node to determine where a packet should be delivered The source address can be used by the destination node to address any reply. Both of these are contained in each IP packet, so it is always possible for a packet to find its way across a network Continue

6. Prepared by: Eng. Ali H. Elaywe6  Activity 39 (exploratory) Why do you think packets traveling across the Internet are more likely to encounter network problems, when compared to packets that remain within the boundary of a LAN?  1- The scale: The Internet is enormous, it carries millions of packets, all trying to navigate their way across and between countries. Just the difference in scale between this and a LAN is likely to lead to more problems  2- Management: A LAN can be planned and managed reasonably accurately. After all, the LAN manager should have a good idea about the traffic generated by all of the users. Also if a user has a particular need, say for extra bandwidth for a short period, they can ask the manager to ensure that this is not a problem. The demands made upon the Internet are much more variable, and different parts of the network are owned and operated by different organizations. Getting an overall picture of traffic demands is much more difficult than within the controlled domain of a LAN

7. Prepared by: Eng. Ali H. Elaywe7  Figure 1 shows three local area networks connecting into a WAN cloud. A cloud is used because the detail contained within it is not needed for our explanation  The principle function of a WAN is to provide communication over very large distances, for example, within a country, between countries or between continents A WAN can be provided for the exclusive use of a group of organizations  private service  A group of networks joined together so that they appear as a single network is called an internetwork; hence, the term Internet, which is a global internetwork Because the Internet is available to anyone, it is called a public service  A backbone is a very high bandwidth communication link/medium to connect together the major nodes of the Internet Continue Topic 2: Wide Area Networks

8. Prepared by: Eng. Ali H. Elaywe8 Continue Figure 1 Wide Area Network

9. Prepared by: Eng. Ali H. Elaywe9  Router: Each LAN connects to the WAN via a router A router is a switching device very similar to a bridge A router operates at the network layer of the OSI model, and makes routing decisions based upon the address information supplied by the packets it receives For example, a router is able to terminate the Ethernet protocol, extract the data and address information (if necessary) and re-package the data in a format suitable for transmission across the WAN

10. Prepared by: Eng. Ali H. Elaywe10  LANs, WANs and Internet etc... use Layered Protocols for data communication purposes  Each layer in a communication process is given a particular responsibility  How it discharges (solves) that responsibility is determined by rules of communication, called protocols For example, one layer may be responsible for simply sending any packets given to it, while another may be responsible for ensuring that all the packets needed to send a message do get to their destination Continue Topic 3: Protocols

11. Prepared by: Eng. Ali H. Elaywe11  Activity 42 (exploratory) I can think of three elements within a computer which this module has mentioned as necessary to send a word-processed file between two networked computers. Can you list them? 1- Application software 2- Network interface card 3- CSMA/CD software Continue

12. Prepared by: Eng. Ali H. Elaywe12  Activity 43 (self-assessment) You were introduced to the OSI Reference Model in Book M. Describe, in your own words, the function of the four lowest layers? 1- The transport layer provides a service concerned with communication between destination and source 2- The network layer controls the routing of data through the network 3- The data link layer looks after the transmission of data across a link. It checks the data integrity across this link 4- The physical layer deals with the transmission of the bits Continue

13. Prepared by: Eng. Ali H. Elaywe13  TCP, IP and Ethernet each occupy different layers, and when they operate together they illustrate nicely how easy it is for a packet at one layer to be carried by the one below  The technique of packaging the information used in a higher layer into the packets of a lower layer is called encapsulation  Each layer takes data from the layer above, applies protocols, and passes the new packet onto the layer below  The layers build up, one upon another, to form a layered architecture, often called a protocol stack  The way encapsulation works when TCP/IP is encapsulated into an Ethernet frame is shown in Figure 2 Sub-Topic 3.1: Encapsulation (Important) Continue

14. Prepared by: Eng. Ali H. Elaywe14 Figure 2 Encapsulation and layers Continue Protocol Stack (segment) (datagram)

15. Prepared by: Eng. Ali H. Elaywe15  At the top of the stack Application data is placed in the data field of a TCP packet, and a header is added  At each layer of the protocol stack a new header is added to the packet handed down from the higher layer For example: the TCP packet is encapsulated into an IP packet by adding the IP header to it. The IP packet is encapsulated into an Ethernet frame by adding the Ethernet header, which means that the Ethernet frame now contains a TCP packet encapsulated within an IP packet  Once the bottom of the stack is reached, then the Ethernet frame must be sent onto a physical medium, to begin its journey across the network  TCP packets are called segments  IP packets are called datagrams Continue

16. Prepared by: Eng. Ali H. Elaywe16  The IP address The Ethernet frame can be transmitted across a LAN using the CSMA/CD protocol and the address information contained in its header If the destination address is within the same LAN, then the received frame will pass up a protocol stack at the destination in the reverse order, allowing the IP and TCP packets to be unpacked on the way If the packet has to leave the LAN then the Ethernet address will direct it to a router. At this point the IP address, contained in the IP packet, is used to direct the packet toward the destination IP provides a universal addressing scheme We will look at IP addressing again later in this session Continue

17. Prepared by: Eng. Ali H. Elaywe17 Figure 3 Layered architecture Continue

18. Prepared by: Eng. Ali H. Elaywe18  Peer-to-peer communication: The arrangement shown in Figure 3 forms the basis of communication across networks It is always the lowest layer of the layered architecture that deals with the physical transmission, in this case part of the Ethernet protocol However, the function carried out at each layer can be understood as if communication takes place directly between layers at the same level It is usual to refer to this virtual communication between corresponding layers as peer-to-peer communication It must always be remembered that actually data is passed up and down protocol stacks at each end of a communication path, and at intermediate points if the network architecture dictates this Continue

19. Prepared by: Eng. Ali H. Elaywe19  Activity 44 (exploratory) A TCP connection is established from a computer on an Ethernet LAN, which then connects to the destination across a WAN. An intermediate node routes IP packets, and converts between Ethernet and a WAN protocol Sketch three protocol stacks, showing the layers at source, destination and intermediate locations using IP routing. Explain what happens at each location Figure 4 shows the situation with three stacks, a complete stack at the source and destination, and a shorter stack at the intermediate node Continue

20. Prepared by: Eng. Ali H. Elaywe20 Continue Figure 4 Intermediate node

21. Prepared by: Eng. Ali H. Elaywe21 At the source: A virtual connection needs to be established between the source and destination. TCP packets are passed down through IP to the Ethernet layer. Ethernet then forwards the packets to the edge of the LAN, which in terms of the connection can be called an intermediate node, or location At the intermediate node: Ethernet cannot forward frames any further because it has reached the boundary of the LAN, so the packet has to be passed up the stack to the IP layer. IP addresses are universal, so IP packets can be forwarded towards the destination. However, these IP packets need to be transported across the WAN, so they will be encapsulated by a WAN protocol At the destination: each of the layers is unpacked

22. Prepared by: Eng. Ali H. Elaywe22  Just as there is a limit on the size of Ethernet frames, both IP and TCP have limits on the maximum size of a packet  Many factors influence the choice of packet size, of which performance and efficiency are two examples. There is a trade-off between them  If packet size is too small, the communication channel will not be efficiently used because of the overheads incurred by each packet Overheads are parts of the packets needed for addressing, error checking and other purposes. The ratio between the actual data content of the packet (the data payload) and the total packet size is one way to express the efficiency of channel use Continue Sub-Topic 3.2: Packets: data format

23. Prepared by: Eng. Ali H. Elaywe23  If the packet size is very large then it is necessary to send a large packet even when a much smaller packet could contain the data. Under these circumstances, interactive 1- performance may suffer because long delays occur during the transmission of large packets, and the user must wait longer for a response to a very short message. Another disadvantage of large packets is the impact if an 2- error occurs during transmission. A lot of data has to be re-transmitted because the whole packet has to be sent again!!  Activity 45 (self-assessment) What is the minimum and maximum data capacity of an Ethernet frame? The minimum size of the data field is 46 bytes, and the maximum is 1500 bytes Continue

24. Prepared by: Eng. Ali H. Elaywe24  Activity 46 (exploratory) One way of expressing efficiency is to calculate the proportion of data carried by a packet as a percentage of the overall packet length. Calculate the percentage efficiency of the smallest and largest Ethernet frames. Each Ethernet frame has 6 bytes for source address, 6 for destination address, 2 for type and 4 bytes for frame checking, giving a total of 18 bytes. Knowing that the data field ranges from 46 to 1500 bytes, we can calculate the minimum and maximum frame sizes to be 64 and 1518 bytes respectively. This gives the following efficiencies: Minimum efficiency = 46/64 = 72% Maximum efficiency = 1500/1518 = 98.8% Continue

25. Prepared by: Eng. Ali H. Elaywe25  Example: Let us summarise by looking again at the size of TCP, IP and Ethernet packets as they cross the different layers. We have also used the ‘proper’ names for packets at each of the layers 1- When TCP is used over an Ethernet data link, the maximum TCP segment size is 1460 bytes 2- At the network layer we have IP datagrams, which have a maximum size of 65 535 bytes 3- At the data link layer the information is handled as frames. When the data link layer protocol is Ethernet, the maximum IP datagram size is 1500 bytes

26. Prepared by: Eng. Ali H. Elaywe26  TCP/IP is the main technology underpinning the development of the Internet and of the World Wide Web  This makes the TCP/IP protocol family of great importance in communication  The TCP and IP protocols are layered, with TCP occupying a higher layer than IP. Together, the TCP and IP layers form the core of the family (see Figure 5) Topic 4: The TCP/IP protocol family Continue

27. Prepared by: Eng. Ali H. Elaywe27 Figure 5 Position of the TCP and IP protocols in the TCP/IP protocol family Continue

28. Prepared by: Eng. Ali H. Elaywe28  A- Duties of the IP Layer: (Important) In OSI terms, IP is a network layer protocol, so it is concerned with routing (or find a route) and getting messages across the network It provides no guarantee that it will succeed, although it will do its best. For this reason the service is often described as a ‘best effort’ service The main aim in IP is to have a network system that is: (1) Flexible (in the sense of being usable with many different types of networks) (2) Robust (in the sense that it should work reasonably well even in the face of breakdowns and damage) Sub-Topic 4.1: TCP and IP Continue

29. Prepared by: Eng. Ali H. Elaywe29  Activity 48 (self-assessment) Discuss whether or not you would consider Ethernet to be a best effort service? I defined Ethernet as a best effort service in Part 1. Remember that if multiple collisions occur, then after 16 attempts the frame is discarded. No action is taken by Ethernet to recover the lost frame, having made its best effort to deliver it Continue

30. Prepared by: Eng. Ali H. Elaywe30  B- Duties of the TCP Layer: (Important) In OSI terms, TCP is a transport layer protocol, which sits on top of IP, and is therefore concerned with end-to-end issues rather than routing messages across the network In contrast to IP, TCP provides a reliable service in the sense that it guarantees error-free transportation of messages between source and destination application layers TCP uses two primary mechanisms to provide a reliable service: 1- The first is called sequencing, and just ensures that packets are sorted out into the same order as they were transmitted. This can be achieved by simply numbering the packets, and making sure each packet is passed up the protocol stack in the correct sequence. Sequencing also allows the receiver to spot if any packets are missing. It can then request the transmitter to send these again Continue

31. Prepared by: Eng. Ali H. Elaywe31 2- The second is less obvious, and is called flow control  Flow control regulates the transmission rate of packets, ensuring that they will not arrive at the receiver faster than they can be processed Note: The information necessary to regulate 1- sequence and 2- flow packets is carried in the header. TCP is designed to be flexible in the sense that it can offer such a service to a wide variety of applications Continue

32. Prepared by: Eng. Ali H. Elaywe32  Concept of Connections: (Important) 1- A connection-oriented service: A connection-oriented service requires the user to keep the connection established (before sending any data) and active throughout the transmission of information, and then to clear the connection afterwards This need of keeping the connection established throughout the transmission arises from the requirements where a Sender may require the Receiver to send back Acknowledgement messages (often called ack/nack communication) and the sender keeps the connection till the acknowledgment is received or no acknowledgment is received This is usually associated with Reliable services such as the TCP. Packets sent as part of this connection should arrive at the destination in the same order they are sent Continue

33. Prepared by: Eng. Ali H. Elaywe33 2- A connectionless service: Does not require the user to keep the connection established (before sending any data) and active throughout the transmission of information. Each packet of data is sent and routed independently. This is usually associated with best effort services such as the IP Whereas TCP provides a connection-oriented service, the IP provides a connectionless service  C- The UDP: (Important) The User Datagram Protocol (UDP) is an alternative transport layer protocol UDP is a connectionless protocol for use with applications that do not need 1- sequencing or 2- flow control Figures 5 and 6 show clearly that more than one protocol is available at a layer Continue

34. Prepared by: Eng. Ali H. Elaywe34 Figure 6 Position of the application protocols in the TCP/IP protocol family Continue

35. Prepared by: Eng. Ali H. Elaywe35  D- Application layer protocols: When you use networked computers in your work, or at home, to connect to Internet service providers you will have direct experience of using application layer protocols that run on top of TCP Examples: 1- Electronic mail uses the simple mail transfer protocol (SMTP) 2- Web pages are accessed using hypertext transfer protocol (HTTP) 3- When you look at network management in Book NM, you will look at an application layer protocol called simple network management protocol (SNMP)

36. Prepared by: Eng. Ali H. Elaywe36  Both the TCP and IP packets have headers and a data field  The IP header contains source and destination addresses, which have the same function as Ethernet addresses, but cover all IP hosts globally  Each address is unique, and identifies a particular host or computer on the network  The IPv4: Each IPv4 address is 32 bits long, and is conventionally divided into four bytes, which in turn are written as four denary numbers, each separated by a full stop For example, the address ‘10000000 00001010 00000110 00100000’ is 128.10.6.32 IPv4 is still widely used, in spite of a perceived demand for larger address fields, which are incorporated in its successor, IPv6 Sub-Topic 4.2: IP Address Continue

37. Prepared by: Eng. Ali H. Elaywe37  Activity 49 (exploratory) How many unique addresses can be represented by the 32 bit IP address? As each IP address is 32 bits long, there is a theoretical maximum of 2 32 (about 4 billion) different addresses. However, because IP addresses are divided into different classes, it may not be practical to use every possible address Continue

38. Prepared by: Eng. Ali H. Elaywe38  Concept of the Domain Name System (DNS): Consider an e-mail addresses of the form A.Student@oufcnt2.open.ac.uk A.Student@oufcnt2.open.ac.uk This e-mail address can be divided into two parts, local- part@domain-name, separated by the @ symbollocal- part@domain-name Resolver software: convert the domain name (part of an e-mail address) into an IP address The information needed to convert from domain name to IP is stored in a series of distributed databases, collectively called the Domain Name System (DNS), and allows the Internet to be divided up into a hierarchy of domains as illustrated in Figure 7 Continue

39. Prepared by: Eng. Ali H. Elaywe39 Figure 7 Domain structure Continue

40. Prepared by: Eng. Ali H. Elaywe40  Activity 50 (self-assessment) If my computer is connected to a LAN, and the person I am sending my message to is on the Internet, my message will have to be sent to a router using Ethernet. Give a general description of how my e-mail is encapsulated into a sequence of Ethernet frames. Assume TCP and IP are used. My e-mail is sent using a mail application on TCP on IP. If it is a long message it will be split up into many parts, with the whole message being compiled as a string of TCP packets. Each TCP packet is encapsulated in an IP packet, which is then encapsulated in an Ethernet frame Continue

41. Prepared by: Eng. Ali H. Elaywe41  Example (related to Activity 50): If we want to send an e-mail then we should enter the e-mail address, type in my message and finally click on the send icon Once the domain name has been resolved, my computer has the IP address of the destination server The destination address of the mail server is resolved from the DNS into a 32 bit IP address. This is then passed down through TCP and written in the IP packet header. This is shown in Figure 8 TCP uses a port number to uniquely identify the application using the TCP connection The source IP address is known by the application and can be passed down in a similar way. The IP layer now contains the destination and source addresses Continue

42. Prepared by: Eng. Ali H. Elaywe42 Figure 8 Address handling Continue

43. Prepared by: Eng. Ali H. Elaywe43 If the final delivery of the packet is over the Ethernet, then the IP addresses have to be converted to Ethernet addresses for final delivery. The problem is how to find the Ethernet address of a node when we only know the IP address? The solution is to ask them all, using a TCP protocol called Address Resolution Protocol (ARP). Figure 9 shows how ARP works  Address Resolution Protocol (ARP) role: IP addresses can be discovered using the Address Resolution Protocol (ARP). A packet containing an ARP request message is sent onto the LAN, which when read by a node, asks that node if its IP address matches the one it is looking for. If it does then the matching node sends back a packet containing its Ethernet address, using the source address of the ARP packet as its destination address Continue

44. Prepared by: Eng. Ali H. Elaywe44 Figure 9 Address resolution Continue

45. Prepared by: Eng. Ali H. Elaywe45 1- If the destination and source are on the same LAN, then the computer can return a packet directly to the originating node. This is shown as node A in Figure 9 2- When the destination is beyond the boundary of the LAN, then a router recognizes the IP address as being on the Internet, and returns its address as the Ethernet destination. The router will then forward packets containing the e-mail onto the Internet Continue

46. Prepared by: Eng. Ali H. Elaywe46  Activity 51 (self-assessment) What special type of destination address can the Ethernet frame use to ensure that an ARP message is read by all nodes? A broadcast address is recognized by all interfaces. This means that it is received and read by all computers, and forwarded by all bridges

47. Prepared by: Eng. Ali H. Elaywe47  1) Must read Books NM & E  2) Do associate activities  3) Complete the associate Journal items  4) Contribute to your tutor group forum  5) Read the TMA02 and prepare questions to be addressed in the next session or preferably in the forum Topic 5: Preparation for Next Session