1 / 16

Module Ethernet Technology

1973 Development begins1976 First paper on Ethernet published1977 Ethernet is patented1979 DEC, INTEL, and XEROX collaborate on Ethernet Version 1.1980 Ethernet V.1 is published1980 Xerox ships first Ethernet1982 Ethernet version 2 is published1983 IEEE approves 802.3 stand

yannis
Download Presentation

Module Ethernet Technology

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


    2. 1973 Development begins 1976 First paper on Ethernet published 1977 Ethernet is patented 1979 DEC, INTEL, and XEROX collaborate on Ethernet Version 1. 1980 Ethernet V.1 is published 1980 Xerox ships first Ethernet 1982 Ethernet version 2 is published 1983 IEEE approves 802.3 standard 1985 Ethernet is produces by more than 200 vendors 1985 Installed base exceeds 20.000 Units 1992-1994 Fast Ethernet / 100 Mbps/ Switching Ethernet Experimental Ethernet was developed by XEROX back in 1975. The experimental Ethernet was running 2.94 Mbps. Ethernet Version I was a joint development by XEROX, DEC and INTEL, and was the first step towards a standardized way to communicate across a Local Area Network. Ethernet Version II was defined as a result of the XEROX, DEC and Intel development and the works in the IEEE standards committee. Primary changes from version I to version II: - Physical signal levels at the transceiver - Drop cable specifications ( from 3 pair to 4 pair ) IEEE 802.3 was published in 1985. Major Ethernet Vendors adopt the new standard. Primary changes from Ethernet version II to IEEE 802.3: - Physical signaling at transceiver - Cabling ( shield moved from pin 1 to pin 4 ) - Frame format (type field to length field) 100 Mbps Ethernet. In 1992 the first fast Ethernet product was announced. Larger LANs and the increasing demand for bandwidth paved the way for a ten-fold increase in the basic LAN signalling speed. The 100BASE-T standard describes the most popular type of fast Ethernet. 100VG-AnyLAN. Another approach to overcome the 10Mbps limitation of Ethernet towards high-speed networking. Originally, 100VG-AnyLAN was proposed to the IEEE committee by HP and AT&T. Today, 100-VG is backed up by IBM as well. The problem with 100-VG AnyLAN though, seams to lie in the different access method to the media. 100 VG-AnyLAN is NOT Ethernet (CSMA/CD), since the access method is referred to as Demand Priority. Demand Priority needs to be implemented in hubs, PC adapters and applications. Therefore 100 VG-AnyLAN does NOT provide a smooth migration, i.e. is NOT backward compatible with existing 10 Mbps Ethernet. Switched Ethernet. In order to further increase the bandwidth available to each LAN user, switched Ethernet products are increasingly installed instead of traditional shared media hubs. Ethernet switches will typically offer dedicated 10Mbps to each pair of communicating stations and 100Mbps fast Ethernet ports for server and backbone connections. Ethernet Experimental Ethernet was developed by XEROX back in 1975. The experimental Ethernet was running 2.94 Mbps. Ethernet Version I was a joint development by XEROX, DEC and INTEL, and was the first step towards a standardized way to communicate across a Local Area Network. Ethernet Version II was defined as a result of the XEROX, DEC and Intel development and the works in the IEEE standards committee. Primary changes from version I to version II: - Physical signal levels at the transceiver - Drop cable specifications ( from 3 pair to 4 pair ) IEEE 802.3 was published in 1985. Major Ethernet Vendors adopt the new standard. Primary changes from Ethernet version II to IEEE 802.3: - Physical signaling at transceiver - Cabling ( shield moved from pin 1 to pin 4 ) - Frame format (type field to length field) 100 Mbps Ethernet. In 1992 the first fast Ethernet product was announced. Larger LANs and the increasing demand for bandwidth paved the way for a ten-fold increase in the basic LAN signalling speed. The 100BASE-T standard describes the most popular type of fast Ethernet. 100VG-AnyLAN. Another approach to overcome the 10Mbps limitation of Ethernet towards high-speed networking. Originally, 100VG-AnyLAN was proposed to the IEEE committee by HP and AT&T. Today, 100-VG is backed up by IBM as well. The problem with 100-VG AnyLAN though, seams to lie in the different access method to the media. 100 VG-AnyLAN is NOT Ethernet (CSMA/CD), since the access method is referred to as Demand Priority. Demand Priority needs to be implemented in hubs, PC adapters and applications. Therefore 100 VG-AnyLAN does NOT provide a smooth migration, i.e. is NOT backward compatible with existing 10 Mbps Ethernet. Switched Ethernet. In order to further increase the bandwidth available to each LAN user, switched Ethernet products are increasingly installed instead of traditional shared media hubs. Ethernet switches will typically offer dedicated 10Mbps to each pair of communicating stations and 100Mbps fast Ethernet ports for server and backbone connections.

    3. The 802.3 standards for 10Mbps Ethernet cover several physical media: 10BASE-2 ‘Cheapernet’. (RG-58 Coax). Segment length up to 185m. 10BASE-5 ‘Yellow Coax’. Segment lengt up to 500m. 10BASE-T Twisted Pair (UTP) Cabling. Requires cabling hub. 10BASE-F Fiber Optic Cable. Requires cabling hub. The fast Ethernet standards include: 100BASE-TX 100 Mbps Ethernet shared media 100BASE-T4 100 Mbps full duplex. Requires cabling hub. 100BAS-FX 100 Mbps on Fiber Optic cables.The 802.3 standards for 10Mbps Ethernet cover several physical media: 10BASE-2 ‘Cheapernet’. (RG-58 Coax). Segment length up to 185m. 10BASE-5 ‘Yellow Coax’. Segment lengt up to 500m. 10BASE-T Twisted Pair (UTP) Cabling. Requires cabling hub. 10BASE-F Fiber Optic Cable. Requires cabling hub. The fast Ethernet standards include: 100BASE-TX 100 Mbps Ethernet shared media 100BASE-T4 100 Mbps full duplex. Requires cabling hub. 100BAS-FX 100 Mbps on Fiber Optic cables.

    4. CSMA/CD Carrier Sense “Listen before trying to transmit - is the cable free? “ Multiple Access “Many Stations are allowed to use the same media” Collision Detect “Be aware of signal collisions - back off when detected” The Carrier Sense Multiple Access / Collision Detect CSMA/CD) media access method is the traffic regulation system which allows stations in an Ethernet LAN to share a common transmission medium (cable). All stations will continously ‘listen’ to activity on the cable (Carrier Sense). In silent periods, when no other station is transmitting on the medium, every station is allowed to initiate a transmission (Multiple Access). Two stations may sense the same ‘no carrier’ situation and initiate a transmission; but then shortly after - how long time depends on the signal propagation delay in the medium and the distance between the two transmitting stations -the two signals will collide and both messages will be destroyed. The collision will cause a significant rise of the average signal level at the cable, and this ‘new’ signal (level) will continue to propagate through the cable. When it reaches the two stations which started the transmissions, they will measure the signal level change (Collision Detect) and they will both know that their message has been destroyed. After a collision detect, both of the transmitting stations will discontinue their transmission (back off) and wait a random time before attempting to transmit the message again. If a new collision occurs on the next try, the new back off period will be two times that of the previous period etc. (binary exponential back off). Upon 15 unsuccessful retries the station considers the medium to be faulty and upper layers will be notified.The Carrier Sense Multiple Access / Collision Detect CSMA/CD) media access method is the traffic regulation system which allows stations in an Ethernet LAN to share a common transmission medium (cable). All stations will continously ‘listen’ to activity on the cable (Carrier Sense). In silent periods, when no other station is transmitting on the medium, every station is allowed to initiate a transmission (Multiple Access). Two stations may sense the same ‘no carrier’ situation and initiate a transmission; but then shortly after - how long time depends on the signal propagation delay in the medium and the distance between the two transmitting stations -the two signals will collide and both messages will be destroyed. The collision will cause a significant rise of the average signal level at the cable, and this ‘new’ signal (level) will continue to propagate through the cable. When it reaches the two stations which started the transmissions, they will measure the signal level change (Collision Detect) and they will both know that their message has been destroyed. After a collision detect, both of the transmitting stations will discontinue their transmission (back off) and wait a random time before attempting to transmit the message again. If a new collision occurs on the next try, the new back off period will be two times that of the previous period etc. (binary exponential back off). Upon 15 unsuccessful retries the station considers the medium to be faulty and upper layers will be notified.

    5. Minimum frame size and maximum span of the LAN. To be able to detect collisions, a transmitting station should monitor the medium for a period of time called a slot time. Slot time is the time during which a collision may occur and is the maximum delay for a transmission to reach the far end of the LAN and for a collision to propagate back. Slot time is defined to be 51.2 microseconds (512 bits time in 10Mbps LAN). This time imposes a maximum length (span) of the size of the LAN. It also imposes a minimum (64 bytes, excluding preample and FCS) on the size of the frames transmitted by each station. Frame formats The frame formats for Ethernet V.II and IEEE 802.3 are not the same. However, both protocols use the same medium and access method. This means that while LAN stations running these protocols could share a common bus, they could not communicate with each other. The layout of the frames are as follows: Preample Syncronization bits Sync/SFD (Eth. II: Synchronize / IEEE: Start Frame Delimiter). Indicates that the data portion of the frame will follow. DA (Destination Address). MAC address SA (Source Address). MAC address Type/Length Type (Eth. II): Identifies the higher layer protocol which is used Length (IEEE): Indicates the number of data bytes (excluding the PAD) Data Min. 46 data bytes, max. 1500 data bytes. If less than 46 data bytes are available, PAD bytes must be inserted. For Eth. V II frames the protocol(s) above the MAC layer inserts PAD. For IEEE frames the MAC layer inserts the PAD). FCS Frame Check Sequence. CRC check on DA through to DATA/PAD. Interframe spacing An interframe gap of minimum of 9.6 microseconds must be inserted after each transmission. This ensures sufficient recovery time for the stations and all physical components.Minimum frame size and maximum span of the LAN. To be able to detect collisions, a transmitting station should monitor the medium for a period of time called a slot time. Slot time is the time during which a collision may occur and is the maximum delay for a transmission to reach the far end of the LAN and for a collision to propagate back. Slot time is defined to be 51.2 microseconds (512 bits time in 10Mbps LAN). This time imposes a maximum length (span) of the size of the LAN. It also imposes a minimum (64 bytes, excluding preample and FCS) on the size of the frames transmitted by each station. Frame formats The frame formats for Ethernet V.II and IEEE 802.3 are not the same. However, both protocols use the same medium and access method. This means that while LAN stations running these protocols could share a common bus, they could not communicate with each other. The layout of the frames are as follows: Preample Syncronization bits Sync/SFD (Eth. II: Synchronize / IEEE: Start Frame Delimiter). Indicates that the data portion of the frame will follow. DA (Destination Address). MAC address SA (Source Address). MAC address Type/Length Type (Eth. II): Identifies the higher layer protocol which is used Length (IEEE): Indicates the number of data bytes (excluding the PAD) Data Min. 46 data bytes, max. 1500 data bytes. If less than 46 data bytes are available, PAD bytes must be inserted. For Eth. V II frames the protocol(s) above the MAC layer inserts PAD. For IEEE frames the MAC layer inserts the PAD). FCS Frame Check Sequence. CRC check on DA through to DATA/PAD. Interframe spacing An interframe gap of minimum of 9.6 microseconds must be inserted after each transmission. This ensures sufficient recovery time for the stations and all physical components.

    6. Physical Signalling Collision Detect SQE Test Jabber Function The Medium Attachment Unit (MAU), also known as the Transceiver, provide the mechanical, electrical and functional interface between the LAN station and the particular media used on the Ethernet (IEEE) bus. A transceiver connects to the LAN station via a separate Attachment Unit Interface (AUI) cable, (also known as a ‘drop’ cable). The Transceiver will provide electrical isolation between the station and the physical media. The transceiver detects collisions. On coax media the collisions are detected by monitoring the voltage level on the center conductor. If the voltage is more than the allowed threshold (-1.6V nominally), it is reported as a collision. On twisted pair media (UTP), the transceiver just monitors activity on the receive wire-pair while transmitting on the transmit wire-pair. The Transceiver provides SQE function. The Signal Quality Error Signal (also known as heartbeat) is used as collision test function. After each frame is transmitted, a 10Mhz burst is sent to the station via the collision wires to inform that the transceiver is working properly. Note: Some repeaters must be attached to the LAN by transceivers that have the SQE disabled because they are not able to discriminate between an SQE and real collisions. This may cause SQEs to be repeated to other segments as collisions, or it may cause the (repeater) hub to partition (disable) the port. The Transceiver provides Jabber protection. Jabber occurs when a station sends data for more than 150 ms (more than maximum frame size ). This could be caused by hardware failure or a running process attempting to send too large frames. The Medium Attachment Unit (MAU), also known as the Transceiver, provide the mechanical, electrical and functional interface between the LAN station and the particular media used on the Ethernet (IEEE) bus. A transceiver connects to the LAN station via a separate Attachment Unit Interface (AUI) cable, (also known as a ‘drop’ cable). The Transceiver will provide electrical isolation between the station and the physical media. The transceiver detects collisions. On coax media the collisions are detected by monitoring the voltage level on the center conductor. If the voltage is more than the allowed threshold (-1.6V nominally), it is reported as a collision. On twisted pair media (UTP), the transceiver just monitors activity on the receive wire-pair while transmitting on the transmit wire-pair. The Transceiver provides SQE function. The Signal Quality Error Signal (also known as heartbeat) is used as collision test function. After each frame is transmitted, a 10Mhz burst is sent to the station via the collision wires to inform that the transceiver is working properly. Note: Some repeaters must be attached to the LAN by transceivers that have the SQE disabled because they are not able to discriminate between an SQE and real collisions. This may cause SQEs to be repeated to other segments as collisions, or it may cause the (repeater) hub to partition (disable) the port. The Transceiver provides Jabber protection. Jabber occurs when a station sends data for more than 150 ms (more than maximum frame size ). This could be caused by hardware failure or a running process attempting to send too large frames.

    7. Max. 500 m. Cable Segments Max. 100 Stations per Cable Segment 2.5 m. between Transceivers 50 Ohm terminator at each end 10BASE-5: Specification for the IEEE 802.3 Coax media, the ‘yellow’ Ethernet cable, which is a very high quality coaxial cable with a center conductor and two outer shields, comprise: Maximum Stations 100 Cable Impedance 50 Ohm +- 2 Ohm Propagation delay 77 % of light speed Bending radius 114 mm Distance to power cable < 2 KV A 6 cm 2-5 KV A 15 cm > 5 KV A 30 cm Attenuation Max. 8.5 dB on 500 Meter cable Connectors AUI The cable is marked at each 2.5 m. and transceivers may only be mounted at these marks. This will ensure that reflections will be out of phase. The transceiver is attached by mounting an AMP clamp. A hole is drilled into the cable through the shield, directly into the center conductor; the tap is mounted and the transceiver is placed on the clamp. A maximum of 100 stations can be attached to a single segment. This limitation is set to prevent signal distortion caused by impedance mismatch between transceivers and the attenuation when several stations connect to the same segment. In both ends of a cable segment, a 50 Ohm terminator is mounted. It matches the characteristic impedance of the cable and prevents reflections which would othervise be monitored as collisions. Grounding must be performed in one end of the cable only in order to prevent ground loop currents. In modern environments the 10BASE-5 topology is not very practical. The difficulties of manipulating the bus cable, rerouting AUI (drop) cables, attaching transceivers etc. means that installations of this nature are inherently inflexible and unable to accomodate the rate of change that is expected on most LANs today. 10BASE-5: Specification for the IEEE 802.3 Coax media, the ‘yellow’ Ethernet cable, which is a very high quality coaxial cable with a center conductor and two outer shields, comprise: Maximum Stations 100 Cable Impedance 50 Ohm +- 2 Ohm Propagation delay 77 % of light speed Bending radius 114 mm Distance to power cable < 2 KV A 6 cm 2-5 KV A 15 cm > 5 KV A 30 cm Attenuation Max. 8.5 dB on 500 Meter cable Connectors AUI The cable is marked at each 2.5 m. and transceivers may only be mounted at these marks. This will ensure that reflections will be out of phase. The transceiver is attached by mounting an AMP clamp. A hole is drilled into the cable through the shield, directly into the center conductor; the tap is mounted and the transceiver is placed on the clamp. A maximum of 100 stations can be attached to a single segment. This limitation is set to prevent signal distortion caused by impedance mismatch between transceivers and the attenuation when several stations connect to the same segment. In both ends of a cable segment, a 50 Ohm terminator is mounted. It matches the characteristic impedance of the cable and prevents reflections which would othervise be monitored as collisions. Grounding must be performed in one end of the cable only in order to prevent ground loop currents. In modern environments the 10BASE-5 topology is not very practical. The difficulties of manipulating the bus cable, rerouting AUI (drop) cables, attaching transceivers etc. means that installations of this nature are inherently inflexible and unable to accomodate the rate of change that is expected on most LANs today.

    8. Max. 185 m. Cable Segments Max. 30 Stations per Cable Segment Min. 0.5 m. Distance between Stations BNC Connectors used to attach stations 50 Ohm terminator at each end 10BASE-2: Specification for the RG-58 coax ‘Cheapernet’ cable, which is a high quality coaxial cable with a center conductor and one outer shield, comprise: Maximum Stations 30 Cable Impedance 50 Ohm +- 2 Ohm Propagation delay 65 % of light speed Bending radius 2 Inches Distance to power cable < 2 KV A 6 cm 2-5 KV A 15 cm > 5 KV A 30 cm Attenuation Max. 8.5 dB on 185 Meter cable Connectors BNC The distance between the station must be at least 0.5m. to ensure that reflections will be out of phase. The transceiver is attached by cutting the cable and mounting BNC connectors at the cables. The use of BNC type connectors makes adding and removing transceivers in a 10BASE-2 LAN a straightforward task. Stations using Cheapernet has often implemented the transceiver on the adapter. This avoids the use of the AUI drop cable and reduces the cost per station attachment. The stations are then attached simply by connecting the bus cables to the adapter by means of a small BNC T-connector block. A maximum of 30 stations can be attached to a single segment. This limitation is set to prevent signal distortion caused by impedance mismatch between transceivers and the attenuation when several stations connect to the same segment. In both ends of a cable segment, a 50 Ohm terminator is mounted. It matches the characteristic impedance of the cable and prevents reflections which would othervise be monitored as collisions. Grounding must be performed in one end of the cable only in order to prevent ground loop currents. 10BASE-2: Specification for the RG-58 coax ‘Cheapernet’ cable, which is a high quality coaxial cable with a center conductor and one outer shield, comprise: Maximum Stations 30 Cable Impedance 50 Ohm +- 2 Ohm Propagation delay 65 % of light speed Bending radius 2 Inches Distance to power cable < 2 KV A 6 cm 2-5 KV A 15 cm > 5 KV A 30 cm Attenuation Max. 8.5 dB on 185 Meter cable Connectors BNC The distance between the station must be at least 0.5m. to ensure that reflections will be out of phase. The transceiver is attached by cutting the cable and mounting BNC connectors at the cables. The use of BNC type connectors makes adding and removing transceivers in a 10BASE-2 LAN a straightforward task. Stations using Cheapernet has often implemented the transceiver on the adapter. This avoids the use of the AUI drop cable and reduces the cost per station attachment. The stations are then attached simply by connecting the bus cables to the adapter by means of a small BNC T-connector block. A maximum of 30 stations can be attached to a single segment. This limitation is set to prevent signal distortion caused by impedance mismatch between transceivers and the attenuation when several stations connect to the same segment. In both ends of a cable segment, a 50 Ohm terminator is mounted. It matches the characteristic impedance of the cable and prevents reflections which would othervise be monitored as collisions. Grounding must be performed in one end of the cable only in order to prevent ground loop currents.

    9. Requires a Cabling Hub Max. 100 m. from Station to Hub 10BASE-T: Specifications for the Unshielded Twisted Pair (UTP) cable comprise: Maximum Stations 2 Cable Impedance 100 Ohm +- 15 Ohm Cable Type 0.5 mm. (24 AWG ) Four twisted wire pairs Propagation delay 59 % of light speed Attenuation 8.5 dB for 100 m. cable Connectors RJ-45 Stations using 10BASE-T has often implemented the transceiver on the adapter. This avoids the use of the AUI drop cable and reduces the cost per station attachment. 10BASE-T is a star topology in which the stations are attached to a central hub. The hub acts as a multiport repeater between a number of segments in which each segment is a point-to-point connection between the station and a port in the hub. A segment can also be a point-to-point connection between two hubs. It is possible to interconnect two stations directly without the use of any hub. It is done by cross-connecting the the two pair of wires (TX to RX) and attach a station at each end of the cable. 10BASE-T cabling require no further termination, and grounding is avoided in most installations. If grounding is needed, Shielded Twisted Pair (STP) cables are used and the hub must be prepared with a common ground connection in every RJ-45 connector. 10BASE-T: Specifications for the Unshielded Twisted Pair (UTP) cable comprise: Maximum Stations 2 Cable Impedance 100 Ohm +- 15 Ohm Cable Type 0.5 mm. (24 AWG ) Four twisted wire pairs Propagation delay 59 % of light speed Attenuation 8.5 dB for 100 m. cable Connectors RJ-45 Stations using 10BASE-T has often implemented the transceiver on the adapter. This avoids the use of the AUI drop cable and reduces the cost per station attachment. 10BASE-T is a star topology in which the stations are attached to a central hub. The hub acts as a multiport repeater between a number of segments in which each segment is a point-to-point connection between the station and a port in the hub. A segment can also be a point-to-point connection between two hubs. It is possible to interconnect two stations directly without the use of any hub. It is done by cross-connecting the the two pair of wires (TX to RX) and attach a station at each end of the cable. 10BASE-T cabling require no further termination, and grounding is avoided in most installations. If grounding is needed, Shielded Twisted Pair (STP) cables are used and the hub must be prepared with a common ground connection in every RJ-45 connector.

    10. Noise Immune No Ground Loops Long Distance (up to 2Km.) Stations connect to a Central Hub Max. Two Stations per Cable Segment Typically used for Inter-Hub Connectivity Fiber Optgic cables presents an attractive solution for high-speed transmission rates used in backbone LANs. The cable is relatively immune to the types of electrical noise and grounding trouble that can plague metallic conductors in some environments. Thus, it is also an ideal medium for outdoor connections or for factories or locations in which cabling has to run near higher voltage wiring. Fiber Optic cables are also characterized by a very little signal attenuation (signal loss due to the medium) and thus, it is able to carry a signal without regeneration over much longer distances than that of metallic cables. The use of Fiber Optic cables in Ethernet is defined in the 10BASE-FL standard (formerly FOIRL). It requires the use of a separate transmit and receive path (2 cables). It also requires the use of repeaters in a central hub acting as the concentration point for a group of stations. 10BASE-F extends the allowable distance between two hubs to 2000 m. Similar to 10BASE-T, link integrity is performed by the station and the hub in each end of the cable segment. Each 10BASE-F transceiver transmits a 1Mz signal when no data transmission is taking place.Fiber Optgic cables presents an attractive solution for high-speed transmission rates used in backbone LANs. The cable is relatively immune to the types of electrical noise and grounding trouble that can plague metallic conductors in some environments. Thus, it is also an ideal medium for outdoor connections or for factories or locations in which cabling has to run near higher voltage wiring. Fiber Optic cables are also characterized by a very little signal attenuation (signal loss due to the medium) and thus, it is able to carry a signal without regeneration over much longer distances than that of metallic cables. The use of Fiber Optic cables in Ethernet is defined in the 10BASE-FL standard (formerly FOIRL). It requires the use of a separate transmit and receive path (2 cables). It also requires the use of repeaters in a central hub acting as the concentration point for a group of stations. 10BASE-F extends the allowable distance between two hubs to 2000 m. Similar to 10BASE-T, link integrity is performed by the station and the hub in each end of the cable segment. Each 10BASE-F transceiver transmits a 1Mz signal when no data transmission is taking place.

    11. Increased Flexibility Improved Availability Inherent Management Capabilities The introduction of 10BASE-T also introduced the star-wiring concept to Ethernet LANs. In star-wired systems the cabling is radiated from a wiring concentrator (the hub) to service a defined area within the building. Changes and moves of stations become easy tasks because all wiring is led to the same point of service. Typically, re-arrangements just require small moves of patch cables within a wiring closet. The introduction of hubs also increased the overall network availability because hubs inherently provide error checking and recovery functions. If a station generates more than 32 successive collisions, the hub will automatically disable the port to the station (partitioning). The introduction of intelligent hubs also allows for the the network administrator to perform management functions at the physical level. Ports can be disabled / enabled, the traffic can be monitored, and statistical information can be retrieved on individual segments as well as on the complete hub. Typically, statistical values can be obtained on: Alignment errors Packet contains an odd number of bytes Pygmies Packet contains 74-82 Bits Runts Packet contains between 552 and 565 Bits Collisions Number of collisions Late Collisions Number of collisions occurred after SQE Etc..... Link segments carry link pulses to indicate that active stations are attached to it. Each station sends link pulses periodically and most hubs will contain a port status indicator (LED) that lights up when a station is attached to the link segment. The introduction of 10BASE-T also introduced the star-wiring concept to Ethernet LANs. In star-wired systems the cabling is radiated from a wiring concentrator (the hub) to service a defined area within the building. Changes and moves of stations become easy tasks because all wiring is led to the same point of service. Typically, re-arrangements just require small moves of patch cables within a wiring closet. The introduction of hubs also increased the overall network availability because hubs inherently provide error checking and recovery functions. If a station generates more than 32 successive collisions, the hub will automatically disable the port to the station (partitioning). The introduction of intelligent hubs also allows for the the network administrator to perform management functions at the physical level. Ports can be disabled / enabled, the traffic can be monitored, and statistical information can be retrieved on individual segments as well as on the complete hub. Typically, statistical values can be obtained on: Alignment errors Packet contains an odd number of bytes Pygmies Packet contains 74-82 Bits Runts Packet contains between 552 and 565 Bits Collisions Number of collisions Late Collisions Number of collisions occurred after SQE Etc..... Link segments carry link pulses to indicate that active stations are attached to it. Each station sends link pulses periodically and most hubs will contain a port status indicator (LED) that lights up when a station is attached to the link segment.

    12. The maximum number of stations in an Ethernet collision domain is 1024. A collision domain is bordered by Bridges or Routers. Many segments and repeaters can be installed within a single collision domain as long as no two stations in the same collision domain are separated by more than four repeaters. No two stations in the same collision domain can be separated by more than three coax segments. The other two segments in a max. configuration must be link segments. Link segments can be 10BASE-T and 10BASE-F - not coax. When five link segments are connected in series, the max. length of a 10BASE-Fsegment is 500 m. When four link segments are connected in series, the max. length of a 10BASE-F segment is 1000 m. Each stand-alone 10BASE-T hub counts as one repeater Each stackable hub counts as one repeater The maximum number of stations in an Ethernet collision domain is 1024. A collision domain is bordered by Bridges or Routers. Many segments and repeaters can be installed within a single collision domain as long as no two stations in the same collision domain are separated by more than four repeaters. No two stations in the same collision domain can be separated by more than three coax segments. The other two segments in a max. configuration must be link segments. Link segments can be 10BASE-T and 10BASE-F - not coax. When five link segments are connected in series, the max. length of a 10BASE-Fsegment is 500 m. When four link segments are connected in series, the max. length of a 10BASE-F segment is 1000 m. Each stand-alone 10BASE-T hub counts as one repeater Each stackable hub counts as one repeater

    13. Bridges Receive and Buffer the Frames Each Side of a Bridge is a Separate Collision Domain Bridges Decide to Discard or Forward Packets Bridges allow for a more Effective Utilization of the Overall Bandwidth Transparent Bridging Ethernet Bridges receive the frame from one segment, buffer the frame and verify in a MAC address table to forward the frame before it is transmitted at the other segment. Due to the buffering of the frames, the Bridge will hold each frame until the transmitting CSMA/CD circuitry has successfully delivered the frame in a collision free time slot. In this way, Bridges effectively extend the LAN into several independent collision domains allowing for longer distances and more stations to be added. The Bridges are inherently transparent to the traffic on the LAN. However, during normal operation, they will automatically ‘learn’ to pass only frames that are addressed to stations placed at different segments. Additionally, the LAN administrator can configure special traffic filters according to MAC addresses, LSAPs and higher layer protocols. In this way, Bridges are very useful to prepare for a more effective utilization of the overall LAN bandwidth. Transparent Bridging Ethernet Bridges receive the frame from one segment, buffer the frame and verify in a MAC address table to forward the frame before it is transmitted at the other segment. Due to the buffering of the frames, the Bridge will hold each frame until the transmitting CSMA/CD circuitry has successfully delivered the frame in a collision free time slot. In this way, Bridges effectively extend the LAN into several independent collision domains allowing for longer distances and more stations to be added. The Bridges are inherently transparent to the traffic on the LAN. However, during normal operation, they will automatically ‘learn’ to pass only frames that are addressed to stations placed at different segments. Additionally, the LAN administrator can configure special traffic filters according to MAC addresses, LSAPs and higher layer protocols. In this way, Bridges are very useful to prepare for a more effective utilization of the overall LAN bandwidth.

    14. Two 100 Mbps technologies are competing to be the preferred Fast Ethernet standard: 100BASE-T 100BASE-T is an extension to the proven 10BASE-T standard. It retains the CSMA/CD access method and the IEEE802.3 frame format. This allows for data to easily flow between 10BASE-T and 100BASE-T stations. No changes to protocols, drivers or frames are required. The cabling for 100BASE-T is similar to that of 10BASE-T; Fiber Optic and UTP Meanwhile, 100BASE-T includes three media specifications: - 100BASE-TX is equivalent to cat. 5 UTP cabling in 10BASE-T installations (two pairs unshielded twisted pair and same pin layout in the RJ-45 connectors). - 100BASE- T4 supports 100Mbps over four pairs of cat. 3, 4 or 5 UTP cabling. The RJ-45 connectors are wired as in most of the existing 10BASE-T installations, but the signalling is different (100BASE-T4 uses three pairs for transmissions and one pair for collision detection). - 100BASE-FX defines 100Mbps over two strands of 62.5/125 micron fiber and uses the same connectors as defined for FDDI and 10BASE-T (MIC, ST and SC). 100VG-AnyLAN 100VG-AnyLAN is a new technology that uses a centrally controlled access method called demand priority. It supports both Ethernet and Token Ring frame formatting and thus, provides for a migration path to high speed Token Ring LANs. The demand priority method requires special drivers in all LAN stations. 100VG-AnyLAN operates over UTP cat. 3,4 and 5 , STP and Fiber Optic cables. Few products (adapters, drivers, Hubs and switches) that support 100VG-AnyLAN are currently available. Two 100 Mbps technologies are competing to be the preferred Fast Ethernet standard: 100BASE-T 100BASE-T is an extension to the proven 10BASE-T standard. It retains the CSMA/CD access method and the IEEE802.3 frame format. This allows for data to easily flow between 10BASE-T and 100BASE-T stations. No changes to protocols, drivers or frames are required. The cabling for 100BASE-T is similar to that of 10BASE-T; Fiber Optic and UTP Meanwhile, 100BASE-T includes three media specifications: - 100BASE-TX is equivalent to cat. 5 UTP cabling in 10BASE-T installations (two pairs unshielded twisted pair and same pin layout in the RJ-45 connectors). - 100BASE- T4 supports 100Mbps over four pairs of cat. 3, 4 or 5 UTP cabling. The RJ-45 connectors are wired as in most of the existing 10BASE-T installations, but the signalling is different (100BASE-T4 uses three pairs for transmissions and one pair for collision detection). - 100BASE-FX defines 100Mbps over two strands of 62.5/125 micron fiber and uses the same connectors as defined for FDDI and 10BASE-T (MIC, ST and SC). 100VG-AnyLAN 100VG-AnyLAN is a new technology that uses a centrally controlled access method called demand priority. It supports both Ethernet and Token Ring frame formatting and thus, provides for a migration path to high speed Token Ring LANs. The demand priority method requires special drivers in all LAN stations. 100VG-AnyLAN operates over UTP cat. 3,4 and 5 , STP and Fiber Optic cables. Few products (adapters, drivers, Hubs and switches) that support 100VG-AnyLAN are currently available.

    15. When traditional shared media LANs are becoming more and more congested by new traffic, it is time to search for new solutions that can provide dedicated bandwidth where and when it is needed. An Ethernet Switch is a LAN device that goes into existing installations and minimizes congestion problems in an inexpensive way, and also, it paves the way for a smooth migration to a high-speed LAN technology like Fast Ethernet. The Ethernet switch can be thought of as a very fast (wire speed), low latency multiport Bridge. Separate LAN segments with one or more stations on each segment can be created off the Switch. The Switch simply replaces the traditional Hub, and the new physical layout of the segments is made by re-arranging patches in the wiring closet. No replacement of protocols, drivers, adapters or cabling components is required. The Switch will forward packets according to the preconfigured port modes: - Fast forwarding does not provide any buffering in the Switch. Each frame starts being directed to the output port as soon as the Switch has received the MAC address section of the incoming frame. This mode should not be used on ports with shared segments since it will allow for collisions to be transferred. - Store and forward will secure that collisions and corrupted frames are not transferred to other segments. CRC check is performed on arrival at the port and, if errorfree, the packet is sent through filters and buffered before transmission. Due to the buffering, the store and forward mode is used where speed conversion from 100Mbps to 10Mbps is required. - Fragment free combines the best of fast-forward and store-and-forward. Forwarding occur nearly instantly, but it will only start after the first 64 bytes have been received. This will secure that collisions are not transferred, and that the frames have been checked for a valid format. In a switched LAN, more stations may simultaneously send data to the same server, and in order to prevent packets from being discarded due to occupied ports or buffer overflow, some kind of flow control mechanism has to be implemented at the physical level. Hence, the Switch will generate fake collisions, i.e. activate the TX wires to the sending station, when it is not possible to deliver more frames to the addressed port.When traditional shared media LANs are becoming more and more congested by new traffic, it is time to search for new solutions that can provide dedicated bandwidth where and when it is needed. An Ethernet Switch is a LAN device that goes into existing installations and minimizes congestion problems in an inexpensive way, and also, it paves the way for a smooth migration to a high-speed LAN technology like Fast Ethernet. The Ethernet switch can be thought of as a very fast (wire speed), low latency multiport Bridge. Separate LAN segments with one or more stations on each segment can be created off the Switch. The Switch simply replaces the traditional Hub, and the new physical layout of the segments is made by re-arranging patches in the wiring closet. No replacement of protocols, drivers, adapters or cabling components is required. The Switch will forward packets according to the preconfigured port modes: - Fast forwarding does not provide any buffering in the Switch. Each frame starts being directed to the output port as soon as the Switch has received the MAC address section of the incoming frame. This mode should not be used on ports with shared segments since it will allow for collisions to be transferred. - Store and forward will secure that collisions and corrupted frames are not transferred to other segments. CRC check is performed on arrival at the port and, if errorfree, the packet is sent through filters and buffered before transmission. Due to the buffering, the store and forward mode is used where speed conversion from 100Mbps to 10Mbps is required. - Fragment free combines the best of fast-forward and store-and-forward. Forwarding occur nearly instantly, but it will only start after the first 64 bytes have been received. This will secure that collisions are not transferred, and that the frames have been checked for a valid format. In a switched LAN, more stations may simultaneously send data to the same server, and in order to prevent packets from being discarded due to occupied ports or buffer overflow, some kind of flow control mechanism has to be implemented at the physical level. Hence, the Switch will generate fake collisions, i.e. activate the TX wires to the sending station, when it is not possible to deliver more frames to the addressed port.

More Related