Networking Test Friday Online

Bell0216

chapter_9_powerpoint.pdf

Home >Computer Science homework help >Networking Test Friday Online

Ethernet

Network Fundamentals – Chapter 9

Objectives  Identify the basic characteristics of network media used

in Ethernet.

 Describe the physical and data link features of Ethernet.

 Describe the function and characteristics of the media access control method used by Ethernet protocol.

 Explain the importance of Layer 2 addressing used for data transmission and determine how the different types of addressing impacts network operation and performance.

 Compare and contrast the application and benefits of using Ethernet switches in a LAN as apposed to using hubs.

 Explain the ARP process.

Ethernet

 OSI model layers 1 (physical) and 2 (data link)

 TCP/IP model Network Access layer

Application

Presentation

Session

Transport

Network

Data link

Physical

Application

Transport

Internet

Network Access Ethernet

Ethernet

 The most common LAN technology

 Different media (copper cable, optical fibre)

 Different bandwidths (10, 100Mbps, Gbps, +)

 Same addressing scheme

 Same basic frame format

Ethernet history

 First LAN was Ethernet, designed at Xerox

 1980 Ethernet standard published by DIX (Digital, Intel, Xerox)

 1985 IEEE modified Ethernet standard and published as 802.3

Ethernet 802.3

802.2 MAC LLC

Sublayers

 Logical Link control sublayer links to upper layers, is independent of equipment.

 Media Access Control sublayer provides addressing, frame format, error detection, CSMA/CD.

 Physical layer handles bits, puts signals on the medium, detects signals.

MAC LLC

Advantages of Ethernet

 Simplicity and ease of maintenance

 Ability to incorporate new technologies (e.g. fiber optic, higher bandwidths)

 Reliability

 Low cost of installation and upgrade

Shared medium

 Physical bus topology 10Base5 (thick coaxial cable up to 500m) 10Base2 (thin coaxial cable up to 185m)

 Physical star topology 10BaseT (UTP cable up to 100m)

 Collisions happen – managed with CSMA/CD

Hubs and switches

 “Legacy Ethernet”, 10Base5, 10Base2 or 10BaseT with hubs is designed to work with collisions, when devices transmit at the same time. Collisions are managed by CSMA/CD.

 Performance is poor if there is a lot of traffic and therefore a lot of collisions.

 Collisions can be avoided by using switches and full duplex operation.

Hubs and switches

Switch forwards

frames only to the

destination once the

address is known.

Hub forwards frames

through all ports

except incoming port.

Half duplex

 One-way traffic. Necessary on a shared medium.

 If PC1 is transmitting but also detects incoming signals then there is a collision.

Full duplex

 Two way traffic

 PC can transmit and receive at the same time

 Not on shared medium – must have dedicated link from switch

 No collisions

Fast Ethernet, Gigabit Ethernet

 Along with the move to switches came higher bandwidth: 100 Mbps or Fast Ethernet.

 Later came 1000 Mbps, Gigabit Ethernet.

 Gigabit Ethernet requires fully switched and full duplex operation. Collisions are no longer defined and cannot be managed.

LAN, MAN, WAN

 Ethernet was developed for local area networks confined to a single building or group of buildings on one site.

 Using fiber optics and Gigabit speeds, Ethernet can be used for Metropolitan Area Networks – throughout a town or city.

 Ethernet can even be used over larger areas so the distinction between LAN and WAN is no longer clear.

Ethernet Frame

Packet

Packet Trailer Frame header

Packet from

Network layer is

encapsulated

Preamble

Destination

address

Start of

frame

delimiter

Source

address

Length

/type

7 1 6 6 2

Packet

Data

Frame

Check

Seq.

46-1500 4

Field size in bytes. Preamble and SFD are not counted in

frame size. Frame is 64-1518 (later 1522) bytes.

Frame fields

 Preamble and start of frame delimiter: act as a wake-up call, help synchronisation, show where frame starts.

 Destination Address: MAC address of destination, 6 bytes hold 12 hex digits.

 Source Address: MAC address of sender, 6 bytes hold 12 hex digits.

Frame fields

 Length/type field: DIX used this for type, the original IEEE 802.3 standard used it for length. The later IEEE standard allows it to be used for either.

 A value less than 0x0600 hex (1536 decimal) is length. A greater value is the type, a code showing which higher layer protocol is in use.

Frame fields

 Data field: This contains the layer 3 protocol data unit, usually an IP packet.

 If the packet is less than 46 bytes then the field length is made up to 46 bytes with a “pad”.

 The frame trailer contains the Frame Check Sequence field, used for the cyclic redundancy check to detect corrupt frames.

Ethernet MAC address

 A unique identification for a device (or NIC).

 Burned into the ROM but copied to RAM.

 First 3 bytes identify the manufacturer (Organizationally Unique Identifier)

 A device reads the destination MAC address to see if it should process the frame.

 A switch reads the destination MAC address to see where it should forward the frame.

Writing a MAC address

 The 12 hex digits are written in different ways

 00-05-9A-3C-78-00

 00:05:9A:3C:78:00

 0005.9A3C.7800

 This is the same address

 00-05-9A is the manufacturer’s ID assigned by IEEE

 3C-78-00 is assigned by the manufacturer

Different addresses

 MAC addresses are used to identify devices within a network. They are layer 2 addresses in the frame header.

 IP addresses are used to pass data between networks. They are layer 3 addresses in the packet header. They identify the network as well as the device.

On a long journey…

 The packet header with IP addresses is created by the source host and stays the same throughout the journey.

 The frame header is stripped off and replaced by each router, so the MAC addresses are different for every step of the journey. If parts of the journey are not over Ethernet then there will be a different addressing system – not MAC.

Unicast, multicast, broadcast

 Unicast: a message sent to one particular host. It must contain the destination host’s IP address and MAC address.

 Broadcast: message for all hosts on a network. “Host” part of IP address is all binary 1s. E.g. 192.168.1.255 MAC address is all binary 1s, FF:FF:FF:FF:FF:FF in hex.

 Multicast: message for a group of devices. IP address 224.0.0.0 to 239.255.255.255

Collisions

 Ethernet originally used shared coaxial cable.

 If hosts transmit at the same time, there is a collision.

 Later networks used hubs and UTP cable but the medium is still shared and collisions occur.

Hubs and Collision Domains

 Collision domain – area where collisions occur.

 Add more hubs and PCs – collision domain gets bigger, more traffic, more collisions.

 Hosts connected by hubs share bandwidth.

 Only one PC can send

CSMA/CD

 Carrier Sense: ‘Listen’ to see if there are signals on the cable

 Multiple Access: Hosts share the same cable and all have access to it

 Collision Detection: Detect and manage any collisions of signals when they occur

 This is the ‘first come, first served’ method of letting hosts put signals on the medium

Listen for signals

Are there signals on

the cable?

Yes.

Wait if there are signals

Wait until there are

no more signals

Listen for signals

Are there signals on

the cable now?

No.

Put signals on cable

Put my signals on the

cable.

Listen for collisions: no

No collision.

All is well.

My message was

sent.

Listen for collisions: yes

There is a collision.

Stop sending signals.

Send jamming signal.

My message is lost.

Listen again

No signals now.

Wait for a random

length of time.

Send message again.

CSMA/CD

 Collisions happen if a host transmits when there is a signal on the cable but the host does not yet know about it.

 Latency is the time a signal takes to travel to the far end of a cable. The longer the cable and the more intermediate devices, the more latency.

All clear

CSMA/CD

 If a host detects a collision while it is sending the first 64 bits of a frame then CSMA/CD works and the frame will get resent later.

 If the host has sent 64 bits and then detects a collision, it is too late. It will not resend.

 Latency must be small enough so that all collisions are detected in time.

 This limits cable length and the number of intermediate devices.

Definitions

 Latency or propagation delay: the time it takes for a signal to pass from source to destination.

 Bit time: the time it takes for a device to put one bit on the cable. (Or for the receiving device to read it.)

 Slot time: the time for a signal to travel to the far end of the largest allowed network and return.

Interframe spacing

 The time between the end of one frame and the start of the next frame.

 Gives the medium a chance to stabilise.

 Gives devices time to process the frame.

 Devices wait a minimum of 96 bit times after a frame has arrived before they can send.

 9.6 microseconds for 10 Mbps Ethernet

 0.96 microseconds for 100 Mbps Ethernet

Different bandwidths

 Change from 10 Mbps to 100 Mbps

 The sender puts the bits on the cable 10 times as fast, but they still travel at the same speed along the cable.

 Collision detected at the same time as before.

Frame gone – too late

Still sending frame

So… for CSMA/CD to work

 The greater the bandwidth, the closer a collision must be in order to detect it in time.

 The greater the bandwidth, the shorter the possible cable length from one end of the collision domain to the other.

 10 Mbps can have reasonable lengths.

 100 Mbps can just manage 100 metres.

 1 Gbps needs special arrangements

 10 Gbps – not a chance. Can’t do collisions.

Get rid of collisions

 Replace all hubs with switches.

 Each device has a private cable and gets the full bandwidth.

 Use full duplex on each link.

 No collisions.

 Can use higher bandwidths.

Legacy Ethernet

 10 Base-T – 10 Mbps, uses UTP cables Transmits on wires 1/2, Receives on 3/6 Uses Manchester encoding.

 10 Base-2 and 10 Base-5 used coaxial cable. They are obsolete and are no longer recognised by the standards.

Fast Ethernet

 100 Base-TX – 100 Mbps, uses UTP cables Transmits on wires 1/2, Receives on 3/6 Uses 4B/5B encoding

 100 Base-FX – 100 Mbps, uses multimode fibre optic cables.

Gigabit Ethernet

 1000 base-T – 1Gbps uses UTP cables. Uses all 4 wire pairs, transmitting and receiving at the same time on the same wire. Complex encoding and detection system.

 1000 Base-SX – uses multimode fibre, shorter wavelength.

 1000 Base-LX – uses single or multimode fibre, longer wavelength.

10 Gbps Ethernet

 Still evolving

 Potential for operating over longer distances – MANs and WANs

 Still uses same basic frame format as other Ethernet versions.

 Higher bandwidths are planned.

Hub and Switch

 Shared medium

 Shared bandwidth

 Collisions

 Point to point links

 Dedicated bandwidth

 Use full duplex – no collisions

Hub

Switch

Switching table

 Switch builds a switching table matching its port numbers to the MAC addresses of devices connected to them.

 When a frame arrives, it reads the destination MAC address, looks it up in the table, finds the right port and forwards the frame.

Flooding

 If the switch does not find the destination address in its table then it floods the frame through all ports except the incoming port.

 Broadcast messages are flooded.

Learning addresses

 The switch learns addresses by looking at the source MAC address of an incoming frame.

 It then matches the address to the port where the frame came in and puts the information in its table.

 Entries are time stamped and removed from the table when the time runs out.

 They can be refreshed when another frame comes in from the same host.

ARP table

 A host wants to send a message.

 It knows the destination IP address and puts it in the packet header.

 It looks in its ARP table and finds the corresponding MAC address.

 It puts the MAC address in the frame header.

Address resolution protocol

 A host wants to send a message.

 It knows the destination IP address.

 The destination MAC address is not in its ARP table.

 Host broadcasts “Calling 192.168.1.7, what is your MAC address?”

 192.168.1.7 replies “My MAC address is…”

 Host sends message and updates ARP table.

Remote addresses

 Host can see that destination IP address is on another network

 It finds the IP address of the default gateway

 It sends an ARP request for the matching MAC address of the default gateway

 Default gateway router replies and gives its own MAC address

 Host sends message via router and updates ARP table.

Proxy ARP

 If a host cannot tell that the destination IP address is on another network, it will send an ARP request asking for the matching MAC address

 The router will reply, giving its own MAC address

 The host will send the message via the router

Characteristics of Network Media used in Ethernet  Identify several characteristics of Ethernet in its early

years.

Characteristics of Network Media used in Ethernet  Describe the emergence of the LAN switch as a key

innovation for managing collisions on Ethernet-based networks

Characteristics of Network Media used in Ethernet  Identify the characteristics of state-of-the-art Ethernet

and describe its utilization of cabling and point-to-point topography

Physical and Data Link Features of Ethernet  Standards and Implementation

Physical and Data Link Features of Ethernet  Describe how the Ethernet operates across two layers

of the OSI model

Physical and Data Link Features of Ethernet  Logic Link Control – Connecting the Upper Layers

Physical and Data Link Features of Ethernet  Media Access Control (MAC)

Physical and Data Link Features of Ethernet  Physical Implementations of the Ethernet

Function and Characteristics of the Media Access Control Method  MAC in Ethernet

Function and Characteristics of the Media Access Control Method  Carrier Sense Multiple Access with Collision Detection

Function and Characteristics of the Media Access Control Method

 Ethernet Timing

Layer 2 addressing and its Impact on Network Operation and Performance  The Frame – Encapsulating the Packet

Layer 2 addressing and its Impact on Network Operation and Performance  The Ethernet MAC Address

Layer 2 addressing and its Impact on Network Operation and Performance  Hexadecimal Numbering and Addressing

Layer 2 addressing and its Impact on Network Operation and Performance  Another Layer of Addressing

Layer 2 addressing and its Impact on Network Operation and Performance  Ethernet Unicast, Multicast and Broadcast

Compare and Contrast the Use of Ethernet Switches versus Hubs in a LAN.  Legacy Ethernet – Using Hubs

Compare and Contrast the Use of Ethernet Switches versus Hubs in a LAN.  Ethernet – Using Switches

Compare and Contrast the Use of Ethernet Switches versus Hubs in a LAN.  Describe how a switch can eliminate collisions,

backoffs and re- transmissions, the leading factors in reduced throughput on a hub-based Ethernet network

Explain the Address Resolution Protocol (ARP) process.

 Mapping IP to MAC Addresses

Explain the Address Resolution Protocol (ARP) process.  ARP – Destinations Outside the Local Network

Explain the Address Resolution Protocol (ARP) process.  ARP – Removing Address Mappings

Explain the Address Resolution Protocol (ARP) process.  ARP Broadcasts - Issues

Summary