T 305 DIGITAL COMMUNICATIONS T 305 Digital Communications

T 305: DIGITAL COMMUNICATIONS T 305: Digital Communications Block I – Part I - Communication Networks Arab Open University-Lebanon Tutorial 2 1

T 305: DIGITAL COMMUNICATIONS Topics of today’s session Topic n 1: Introduction 2: Communicating Across a Network 3: The Internet 4: Ethernet (CSMA/CD) 5: Fast and Gigabit Ethernet 6: Switched Ethernet 7: Preparation for Next Tutorial Arab Open University-Lebanon Tutorial 2 2

T 305: DIGITAL COMMUNICATIONS Topic I - Introduction Telephone networks (connection of physical path) is an example of circuit switching connection type. n Arab Open University-Lebanon Tutorial 2 3

T 305: DIGITAL COMMUNICATIONS Packet networks operate on a different principle, which breaks the message into many small packets and then sends each one across the network separately. Each packet has a header, which contains address and control information. n Arab Open University-Lebanon Tutorial 2 4

T 305: DIGITAL COMMUNICATIONS Packet switching Message data is carried in the body of the packet. A message will have been successfully sent only when all the packets have arrived at the destination. n Header Data Arab Open University-Lebanon Tutorial 2 Trailer 5

T 305: DIGITAL COMMUNICATIONS Topic 2 -Communicating Across a Network A modem is an interface device that is connected between your computer and your telephone line. Arab Open University-Lebanon Tutorial 2 6

T 305: DIGITAL COMMUNICATIONS n International and national networks Communication links across the oceans are very expensive to install and maintain. This is why advanced digital multiplexing techniques (like time division multiplexing TDM) by means of a device called a multiplexer (mux) are used to carry very large numbers of channels on single cable or satellite link. n Using digital multiplexing it is possible for many thousand telephone calls as well as mobile phone calls, email, facsimile (fax), web connections, home entertainment (such as video-on-demand), and many more services to be carried on a single cable. n Arab Open University-Lebanon Tutorial 2 7

T 305: DIGITAL COMMUNICATIONS earth ground station Ground Station Both satellites and cables are used to communicate with many different paths of the world. Some advantages of Satellites over cable are: n. Cables terminate at cable stations whereas satellite links terminate at Earth stations. n Cable stations are normally located near to the coast where the cables land. However, Earth stations have been in remote locations to isolate them from electrical interference. n Arab Open University-Lebanon Tutorial 2 8

T 305: DIGITAL COMMUNICATIONS Capacity of the network links n. One technique that provides the bandwidth necessary for the delivery of home entertainment services is called asymmetric digital subscriber line (ADSL). The first trans-Atlantic telephone cable (TAT-1) was installed in 1956 and could carry 36 simultaneous telephone calls. TAT-12 and TAT 13, brought into service in 1996, each could carry 60, 000 phone calls n Arab Open University-Lebanon Tutorial 2 9

T 305: DIGITAL COMMUNICATIONS Topic 3: The Internet Connecting networks n. Networks that connect together computer users within a limited geographical area are called local area networks (LANs). LANs can vary significantly in size, from very small networks to very large networks joining together many users situated in different buildings. n. A LAN is designed to provide communication between computer facilities within a limited range, and will usually connect computer users to other shared devices such as printers and servers. n Arab Open University-Lebanon Tutorial 2 10

T 305: DIGITAL COMMUNICATIONS Topic 3: The Internet A network that can connect widely dispersed computers or LANs together is called a wide area network (WAN). n. A combined network of WANs and LANs is called an Internet. n. Routers are the devices that are most often used to connect a LAN to a WAN, and are the most common devices that are used to route packets across wide area network. Because routers sit on the boundary between a LAN and WAN, they will often need to convert between LAN and WAN protocols. n Arab Open University-Lebanon Tutorial 2 11

T 305: DIGITAL COMMUNICATIONS The Internet IP networks n. It is common practice to refer to the sub-networks that make up the Internet as IP networks. This is because these networks all use the Internet protocol (IP). n. If you are accessing Internet from home, you need to connect to an Internet service provider (ISP). n Arab Open University-Lebanon Tutorial 2 12

T 305: DIGITAL COMMUNICATIONS IP network architecture n. The solution to the problem of scalability of the Internet is to break the network down into manageable parts, or domains. Each domain is called an autonomous system (AS). n. An autonomous system is made up from many nodes, each one containing one or more routers, all connected together using SDH or some other transport technology. n. SDH is a hierarchy of bit-rates that can be constructed using TDM techniques. Standard bit-rates include 155 Mbit/s and 2. 5 Gbit/s. n. The routing of packets among nodes within an autonomous system is carried out by the routers using a protocol called interior gateway protocol (IGP). n. Routing between autonomous systems uses a different protocol called a border gateway protocol (BGP). n. The central cores of the Internet are called backbones that join together the nodes of autonomous systems. n Arab Open University-Lebanon Tutorial 2 13

T 305: DIGITAL COMMUNICATIONS Topic 4: Ethernet (CSMA/CD) n Ethernet is an example of LAN technology, and the IEEE standardized it in the 802. x series of standards. n It is general practice to refer to the networks conforming to the IEEE 802. 3 set of specifications as Ethernet LANs. Arab Open University-Lebanon Tutorial 2 14

T 305: DIGITAL COMMUNICATIONS n Ethernet architecture The initial interpretation of what meant by an Ethernet LAN was a 10 Mbits/s carrier sense multiple access with collision detection (CSMA/CD) network with thick co-axial cable connected to a bus configuration. Today, most LANs use twisted pair cable, based on a star configuration. n. Each computer connects to the network through a network interface card (NIC). The simplest arrangement in common use is to connect all the computers to a central hub (repeater) in a star topology. n. The job of the repeater is to regenerate (bit-by-bit) each packet received and send data to every other computer. This allows every computer to send data to every other computer. n. Typical commercial hubs can accept a maximum of 24 connections but can be extended by connecting two or more hubs together. n Arab Open University-Lebanon Tutorial 2 15

T 305: DIGITAL COMMUNICATIONS Carrier sense Multiple access n. A network interface card (NIC) can be in one of three states: transmitting, receiving, or cycling its state between transmission and reception. Before a NIC can transmit, it monitors its receivers and waits a period of silence (9. 6 μs according to IEEE 802. 3 standard) before transmitting a frame. n. The reason for the 9. 6 μs inter-frame gap is to allow the station that last transmitted to cycle its circuitry from transmit mode to receive mode. Without the inter-frame gap, such a station might miss a fame that was destined for it because it had not yet cycled back into receive mode. n. In practice, some NIC manufacturers design their NICs with a smaller inter-frame spacing to give a higher data transfer rate. n Arab Open University-Lebanon Tutorial 2 16

T 305: DIGITAL COMMUNICATIONS Collision Detection n. Two NICs may simultaneously detect ‘silence’ and both start to transmit resulting in collision. Collision detection refers to the ability of a NIC to detect the resulting ‘collision’ of frames, and react to resolve the conflict. The combination of carrier sensing (to prevent transmission that otherwise would result in collisions) and collision detection are the key features of the Ethernet protocol. n Arab Open University-Lebanon Tutorial 2 17

T 305: DIGITAL COMMUNICATIONS When a collision is detected, the NIC waits a short while and then retransmit. n. The events leading up to a collision n. Station A detects that the physical medium has been quiet for 9. 6 s and begins a data frame transmission. n. At some time before the signal from A reaches B, station B also transmits a data frame. n. At some point the frames from stations A and B pass each other (collide). n. The station closest detects the collision first (station B). it stops transmission and transmits a 32 -bit jam message onto the channel. n Arab Open University-Lebanon Tutorial 2 18

T 305: DIGITAL COMMUNICATIONS n Ensuring detection of collision For CSMA/CD protocol, it is necessary to ensure that a NIC stays in the transmit mode long enough to detect a collision, i. e. , all frames are longer than minimum size. n. For a CSMA/CD network, twice the propagation time between the two most distant stations is called the slot time. This time is also referred to as the contention time or round-trip delay. n. The slot time is often described in terms of bits rather than time. For example, 802. 3 define the slot time as 512 bits. This can be converted into a round trip delay if you know the bit-rate that is being used. n. The time for signals to travel between stations is the sum of the time it takes for signals to propagate and the delay introduced by repeaters. n Arab Open University-Lebanon Tutorial 2 19

T 305: DIGITAL COMMUNICATIONS n Events following a collision n When a station detects collision, it immediately stops the transmission of the frame. Then it calculates a delay period using an algorithm known as the binary exponential backoff algorithm before retransmit again. As more stations place demands on the network, the share each gets of the total available bandwidth decreases. In addition, as the demand for network bandwidth increases, efficiency begins to drop because more frames have to be retransmitted. In an extreme case, an Ethernet LAN can enter a collapse state where all frames collide for a significant amount of time. Typically, collapse is triggered by an interface with a failure in its circuitry causing it to send a continuous stream of frames- a condition known as jabbering. Arab Open University-Lebanon Tutorial 2 20

T 305: DIGITAL COMMUNICATIONS Collision domains n. A bridge prevents the spread of collisions by detecting n collided frames and discarding them n. Two or more stations in the same collision domain cannot transmit simultaneously without causing a collision, whereas stations in different collision domains can. The bridge will handle the simultaneously reception and transmission of frames. Arab Open University-Lebanon Tutorial 2 21

T 305: DIGITAL COMMUNICATIONS All computers connected together by repeater hubs are said to share a collision domain. The method of sending packets to every connected computer is called broadcast. n. The advantage of broadcast is that it is an easy protocol to implement, but the problem is that it generates a lot of traffic as all the packets are copied and sent around the network. n. In order to avoid spread data too widely across a network, a bridge is used. This device directs traffic by acting intelligently upon address information contained within the header of every packet it receives. n. A bridge knows where to send packets because it has a forwarding table (implemented in software). It constructs its forwarding table by analyzing source address in each packet it receives, and recording which port the packet arrived on. From this data, it is able to list the addresses of all the computers on the LAN, and for each address, indicate which port will reach that computer. n Arab Open University-Lebanon Tutorial 2 22

T 305: DIGITAL COMMUNICATIONS Topic 5: Fast and Gigabit Ethernet The increase in the number of users, together with bandwidth-hungry applications (like full-color publishing, medical applications requiring complex images, network back-up of servers and storage systems) requires new methods of interconnecting computers on LANs. n. The performance of Ethernet can be improved through the use of fast (100 Mbit/s), gigabit (1000 Mbit/s) and 10 gigabit (10 Gbit/s) Ethernet Technologies. n Arab Open University-Lebanon Tutorial 2 23

T 305: DIGITAL COMMUNICATIONS The IEEE 802. 3 Sub-Layers n. The idea of layers helps explain how it has been possible to keep separate the changes in protocol and medium access as the speed of Ethernet has increased from 10 Mbit/s to 10 Gbit/s. n. The IEEE 802. 3 specification divides the data link layer into two sub-layers: the logical link control (LLC) sublayer and the medium access control (MAC) sub-layer. n. The CSMA/CD protocol is one implementation of MAC sub-layer. n Network Data Link Physical LLC Logical Link Control sublayer MAC Medium Access Control sublayer Arab Open University-Lebanon Tutorial 2 24

T 305: DIGITAL COMMUNICATIONS When frames are sent across a physical medium, it will be necessary to convert the data contained within the frame into a physical signal that can be sent over the medium. This process is called line coding. n. The electrical signals that represent an IEEE 802. 3 MAC frame are created and communicated to the LAN cable through the physical sub-layers n. The 10 Gbit/s LAN is primarily intended to operate over optical-fiber cable. n Arab Open University-Lebanon Tutorial 2 25

T 305: DIGITAL COMMUNICATIONS Fast Ethernet n. Fast Ethernet provides an extension of the IEEE 802. 3 MAC sub-layer protocol that increases the transmission rate from 10 Mbit/s to 100 Mbit/s. The fast Ethernet specification is named 100 base. T and contains a set of options that allow the fast Ethernet protocols to be used with different media like: n 100 base. T 4: designed to operate using four pairs in unshielded twisted-pair (UTP) cables and uses 8 B 6 T coding (8 binary digits represented as 6 ternary symbols). The T stands for twisted pair and the 4 for the number of pairs. n 100 base. FX: designed for use with optical fiber. n. Fast Ethernet retains CSMA/CD protocol enabling data to move between basic Ethernet and fats Ethernet nodes on the LAN without protocol translation. n. The IEEE 802. 3 specification allows for a maximum cable length of 2. 5 km with four repeaters. n Arab Open University-Lebanon Tutorial 2 26

T 305: DIGITAL COMMUNICATIONS Gigabit Ethernet n. The key objectives of the Gigabit Ethernet Task Force were to develop a gigabit Ethernet standard that: n. Allows half- and full-duplex operations at speeds of 1000 Mbit/s. n. Uses the 802. 3 frame format. n. Uses the CSMA/CD access method. n. Is compatible with earlier 10 base. T and 100 base. T technologies. n Arab Open University-Lebanon Tutorial 2 27

T 305: DIGITAL COMMUNICATIONS Gigabit Ethernet The gigabit Ethernet protocol provides an unacknowledged, best-effort delivery service (delivery of the message across a network is attempted but not guaranteed) and requires no changes in higher layers. n. In 10 and 100 Mbit/s , the minimum MAC frame length is set to 512 bits. This determines the slot time and the maximum distance between the two most distant stations on the LAN. n. The physical sub-layers in gigabit Ethernet are architecturally identical to fast Ethernet but the details of conversion between the MAC sub-layer and the physical medium will be different (for example, line coding). n Arab Open University-Lebanon Tutorial 2 28

T 305: DIGITAL COMMUNICATIONS Topic 6: Switched Ethernet One approach applicable to increase the data transmission rates to 10, 100 or 1000 Mbit/s systems is to re-organize the architecture of a network to provide higher bit-rates only where they are needed. n. The introduction of switches, together with the change towards UTP cabling boost a move away from the shared half-duplex medium to full-duplex operation in Ethernet. n Arab Open University-Lebanon Tutorial 2 29

T 305: DIGITAL COMMUNICATIONS Switched Ethernet Full-Duplex operation n. The normal half-duplex operation of an Ethernet NIC means that it can either transmit or receive, but not both together. The advantage of moving to full duplex operation is doubling the throughput of the LAN. For example, a 100 Mbit/s link is capable of carrying 200 Mbit/s. n. Full-duplex Ethernet operation is becoming possible through moves to a general use of UTP cable as an alternative to the original coaxial cable. n Arab Open University-Lebanon Tutorial 2 30

T 305: DIGITAL COMMUNICATIONS Bridges and switching hubs n. One solution to increasing the effective bandwidth is to segment a LAN into sub-LANs, each of which has a coherent user population, for example a department or a workgroup, and its own server. n. The rationale for creating a sub-LAN is that communication between most individual users has a higher probability of being with other members of the sub-LAN than with members of other sub. LANs. n. A bridge then connects a pair of sub-LANs. n Arab Open University-Lebanon Tutorial 2 31

T 305: DIGITAL COMMUNICATIONS Bridges and switching hubs Partitioning the workgroups with a bridge means that local workgroup traffic will no longer affect the operation of the other workgroup. n. A bridge will contain a look-up table that maps individual node addresses on all its sub-LAN segments to its ports. In this way, the bridge ‘knows’ to which of its ports any frame has to be sent. n. The bridge inspects every frame and looks up the destination address in the table and sends it out on the appropriate port. If the destination address is unknown, the bridge ‘floods’ all ports except the receiving port with the received frame. n Arab Open University-Lebanon Tutorial 2 32

T 305: DIGITAL COMMUNICATIONS Arab Open University-Lebanon Tutorial 2 33

T 305: DIGITAL COMMUNICATIONS n Switches are bridges with large numbers of ports, which were capable of forwarding frames at ‘wire speed’ on all ports. n. A switched LAN is an alternative to a sharedbandwidth LAN. n In a switched LAN, each port on a switch constitutes a single collision domain. So, if more than one station is attached to a port, collisions can occur between them. If only a single station is attached to a port, collisions cannot occur between stations. n Arab Open University-Lebanon Tutorial 2 34

T 305: DIGITAL COMMUNICATIONS The benefits of full-duplex operations are not apparent for computers or workstations used in standard business operations such as word processor. Full-duplex operation is really directed to server operation, in which several applications and users are supported simultaneously. The advantages are: n. Full-duplex NICs: simultaneous transmission and reception. n. Full-duplex transmission medium: independent transmit and receive paths. n. Full-duplex switches: provision of contention-free paths between stations. n. One significant benefit of full-duplex operation comes from the elimination of collision domains. Hence, full-duplex operation with no collision domain, allows LANs to span greater distances. n. When the receiver's input buffer approaches its capacity, the receiver's transmit section inserts a pause frame in order to slow down the sending station. n Arab Open University-Lebanon Tutorial 2 35