The Ethernet Technology Report - Essay Example

 Ethernet Introduction Introduced in 1980, Ethernet is presently the most widely used LAN technology that defines the cabling and signaling specification for the physical and data link layer of the OSI model. It refers to the protocols defined by the underlying data rates (10 Mbps, 100 Mbps, 1 Gbps or 10 Gbps), the physical medium (thick/thin coaxial cable, twisted pair or fiber-optic cables), running different signaling methods (Manchester, 4B/5B, 8B/10B) providing an access method for the shared medium Carrier Sense Multiple Access with Collision Detection (CSMA/CD). Ethernet provides an inexpensive, high-performance networking environment where all types of computer devices can be connected to one another. 1. Ethernet Packets Every bit of information that is transmitted on an Ethernet network is in form of packets. A packet simply refers to a portion of data (payload) packed with identification headers for routing it through various networks to the destination. A data packet on Ethernet is called frame. A typical Ethernet frame for a 10/10 Mbps Ethernet is as shown in Figure 1. Figure 1. A Typical Ethernet Frame Each field holds significance as explained below: Preamble (PRE) (7 bytes): an alternating pattern of 1s and 0s which helps an Ethernet card in synchronizing with the start of an incoming frame. Start-of-frame delimiter (SFD) (1 byte): an alternating pattern of 1s and 0s, ending in two 1s as 10101011, to indicate the next bit to be start of frame i.e. the destination address. Destination address (DA) (6 bytes): identifies the Media Access Control (MAC) address of the recipients Ethernet card. Source addresses (SA) (6 bytes): identifies the sender’s MAC address. Length/Type (2 bytes): indicates either the number of data bytes contained in frame or the frame ID in case an optional format of frame is used. Data: a sequence of bytes ranging between 46 to 1500 containing any value. Frame check sequence (FCS) (4 bytes): is a cyclic redundancy check (CRC) value, which the sender MAC creates. The check is recalculated by receiver’s MAC to check for any damaged frames. 2. Encoding Schemes For transporting digital bits of data through the carrier waves, Ethernet has some encoding techniques. These encoding techniques aid in synchronizing the data sending and data receiving stations and providing a signal with strength sufficient enough to deliver a required amount of data rate. So, the difference in encoding techniques relate to the difference in the data transfer speed provided by the Ethernet cable. For instance, the three encoding techniques Manchester, 4B/5B and 8B/10B are used for the providing Ethernet LAN of speeds 10Mbps, 100 Mbps and 1Gbps respectively (TechFest.com, 1999). Manchester encoding delivers a signal of frequency 10MHz which is sufficient to provide a data rate of 10Mbps for the 10Base5 and 10Base2 coaxial cables. The installation is easy and new systems can be added easily. However, owing to the bus topology (in 10Base5) and the fact that the cabling is half duplex (i.e. same cable used for sending and receiving), chances of collisions would be more and the throughput of the system would be less. In case of daisy chains (10Base2), routing difficulties would be present. Secondly, the higher data rates cannot be achieved due to coaxial cables. In case of using 10BaseT twisted pair cables, 20Mbps can be achieved owing to full duplex mode. New hosts can be added easily in the supported star topology network. However, repeaters are required after every 100m unlike 10Base5 and 10Base2. When using 10BaseFL, repeaters are not needed till 1000m. The 4B/5B encoding delivers a 62.5MHz frequency signal on the 100Base-Tx Cat 5 twisted pair cabling. A signal of maximally 100MHz is enough to be transmitted over a medium of 100Mbps. In full duplex mode, any new host can be added without affecting the collision rates. 4B/5B encoding is also used in 100Base-Fx cables. In case of 8B/10B encoding, as 10 bits are required in encoding 8 data bits, the data rate goes lower than the line speed. So for delivering a data rate of 1Gbps, the line speed of 1.25Gbps is required when using 8B/10B encoding. 8B/10B is used in 1000Base-LX, 1000Base-SX and 1000-Base-CX cables where a transmission rate of 2000Mbps can also be achieved in full duplex mode. Since the full duplex mode is supported, by adding hosts the collision chances do not increase. 3. CSMA/CD When Ethernet is set to operate in half duplex mode (as in the case of coaxial cables using Manchester coding), the transmit (Tx) and receive (Rx) signals are sent along the same wire. So a protocol to access the shared, contentious medium called CSMA/CD is used (Shinder, 2001, pp. 122). The medium access and transmission works as follows: 1. Before transmitting a frame, the station senses the network to detect any pre-existing activity. 2. In case there is no activity for a time period called InterFrame Gap (IFG), the station transmits its frame. 3. In case, there is activity, the station keeps listening. 4. In case two stations happen to detect a free medium for IFG and choose to transmit at the same time, a collision of frames occurs. 5. Any collision is detected by the network and a jam signal is sent to all stations connected to the medium to discard the corrupted frames. All transmissions stop when a collision is detected. 6. After collision, the transmitting stations wait a time interval determined by a random exponential back-off algorithm before re-transmitting the frame. 7. The whole listening process repeats from Step 1. In the Fast Ethernet setup where a data rate of 10Gbps is required, the full-duplex mode is used, where separate Tx and Rx wires are used. This removes the requirement of CSMA/CD. 5. References Shinder, D.L. (2001). Computer Networking Essentials, Cisco Press. TechFest.com. (1999). TechFest Ethernet Technical Summary. Retrieved from http://www.techfest.com/networking/lan/ethernet4.htm Read More