Title: Ethernet
1Ethernet
- In 1976 a 2.94 Mbps a system named Ethernet using
CSMA/CD was implemented at Xerox. It had over 100
personal workstations on a 1-km long cable
2IEEE802.3 Medium Access Control
- Random Access
- Stations access medium randomly
- Contention
- Stations content for time on medium
- It all began with ALOHA
3ALOHA
- Packet Radio - The first Multiple Access network
was the ALOHA packet radio system developed in
University of Hawaii, back in 1971 - ALHOA true free for all - Transmit whenever you
like without regard to any one else. When
station has frame, it sends - Station listens (for max round trip time) plus
small increment - If ACK, fine. If not, retransmit
4ALOHA (continued)
- If no ACK after repeated transmissions, give up
- Frame check sequence (as in HDLC)
- If frame OK and address matches receiver, send
ACK - Frame may be damaged by noise or by another
station transmitting at the same time (collision) - Any overlap of frames causes collision
- Max utilization 18
5Slotted ALOHA
- Time in uniform slots equal to frame transmission
time - Need central clock (or other sync mechanism)
- Stations can only start transmitting at the
beginning of a time slot. Not at any other time - Transmission begins at slot boundary
- Frames either miss or overlap totally
- Slotted ALOHA - Discrete Time ALOHA
- Less opportunity to collide
- Max utilization 37
6(No Transcript)
7Aloha vs. Slotted Aloha
- Throughput efficiency increases dramatically for
Slotted Aloha.
8CSMA
- Propagation time is much less than transmission
time - All stations know that a transmission has started
almost immediately - First listen for clear medium (carrier sense)
- If medium idle, transmit
- If two stations start at the same instant,
collision - Wait reasonable time (round trip plus ACK
contention) - No ACK then retransmit
- Max utilization depends on propagation time
(medium length) and frame length - Longer frame and shorter propagation gives better
utilization
9Carrier Sensed Multiple Access (CSMA)
- CSMA/CD is an acronym for Carrier Sensed Multiple
Access with Collision Detection. - This is a way for many devices to access the same
network (Multiple Access). - The way it works is that every device needing to
send data to the network first "listens" to the
wire to determine if anyone else is sending
anything (Carrier Sense). - If there is any activity at the time, then the
device needing to transmit will wait until the
transmission ends, otherwise it sends its data.
10CSMA Protocol
- Listen before you speak
- If someone else is speaking then wait and let
them finish - backoff - Persistency
- More efficient than Slotted ALOHA
11If Busy?
- If medium is idle, transmit
- If busy, listen for idle then transmit
immediately - If two stations are waiting, collision
12CSMA/CD improves CSMA
- With CSMA, collision occupies medium for duration
of transmission - Stations listen whilst transmitting
- If medium idle, transmit
- If busy, listen for idle, then transmit
- If collision detected, jam then cease
transmission - After jam, wait random time then start again
- Binary exponential back off
13Collision Detection
- On baseband bus, collision produces much higher
signal voltage than signal - Collision detected if cable signal greater than
single station signal - For twisted pair (star-topology) activity on more
than one port is collision - Special collision presence signal
14Collision Detection (CD)
- Once a device determines that the network is
clear, it begins sending data. - While sending, it monitors what data is actually
on the wire, and compares it to what is being
sent. It is possible for two devices to see that
the network is open at the same time and for both
to start sending. When this happens, the data on
the wire will be garbled and neither device will
see that what is actually on the wire matches
what it sent. This is called a collision. - The devices detect this collision, and both stop
their current transmission, wait a random amount
of time, and try again.
15TEBB
- A and B after the first collision
- both select random number from 0 and 1.
- if both select 0 and 0 or 1 and 1, there will be
another collision. - Thus collision probability is 1/2.
16TEBB
- A and B after the second collision
- both select random number from 0 to 3.
- if both select the same number, there will be
another collision. - thus collision probability is 1/4.
17TEBB (cont.)
- A and B after the 10to 16th collision
- both select random number from 0 to 1023.
- if both select the same number, there will be
another collision. - thus collision probability is 1/1024.
- After 16th collision, transmission is aborted.
18Truncated Exponential Backoff Algorithm
- Slot time 512bits.
- r slots delay before nth retransmission.
- r is a uniformly distributed random integer.
- Kmin(n,10).
- 0lt r lt2K
19CSMA/CDOperation
20CSMA/CD Flow Chart
21Collision Detection Can Take As Long as 2?
22The Collision Domain
- Collision detection can take as long as 2t, worst
case. - This round-trip delay defines the max Ethernet
network diameter, or collision domain. - Round-trip delay 512 bit times for all
ethernets.
23Ethernet MAC (802.3)
- 802.3 MAC does the following
- Minimize access delay for low traffic
- Minimize collision for low traffic
- Maximize bandwidth utilization when few large
data chunks are transmitted. - Ensure fair access and bandwidth allocation by
random access and minimizing the maximum packet
size. - Does not give a bound on the access delay.
- Efficiency degrades under heavy loads.
24Ethernet
- Developed at Xerox PARC - Robert Metcalfe
- In 1976 a 2.94 Mbps a system named Ethernet using
CSMA/CD was implemented at Xerox. It had over 100
personal workstations on a 1-km long cable - Was meant to be a part of the office of the future
25DIX Ethernet
- DIX - Digital, Intel, Xerox
- First commercial Ethernet system
- Original Ethernet system was so successful that
DEC and Intel collaborated with Xerox to build a
10 Mbps Ethernet. - This project was the basis for IEEE standard
802.3, which includes specifications for 1 to 10
Mbps networks using different physical media and
topologies
26IEEE 802 Series of Standards
- Why LAN Standards?
- 802.1 - Overview
- 802.2 - Logical Link Control
- 802.3 - IEEE Ethernet
27IEEE 802 Reference Model
28LAN/MAN Standards
29IEEE 802.3 at 10Mb/s
30IEEE 802.3
- Uses Carrier sense multiple access with collision
detection - Four step procedure
- If medium is idle, transmit
- If medium is busy, listen until idle and then
transmit - If collision is detected, cease transmitting
- After a collision, wait a random amount of time
before retransmitting
31802.3 Medium Notation
- Notation formatltdata rate in Mbpsgtltsignaling
methodgtltmaximum segment length in hundreds of
metersgt - e.g 10Base5 provides 10Mbps baseband, up to 500
meters - T and F are used in place of segment length for
twisted pair and fiber
32IEEE Ethernet Nomenclature
- 10Base5 - Original Ethernet - Thicknet
- 10Base2 - Cheapernet - Thinnet
- 10BaseT - Twisted Pair, Star Topology Ethernet
- 10BROAD36
- 10BaseFL - Optical Fiber Ethernet
33Ethernet Topology
- Originally BUS
- Star and Tree
34Ethernet Media Types
- Coaxial Cable - Thick and Thin
- Unshielded Twisted Pair
- Optical Fiber
35Different Media and Wiring For Ethernet Types
36Ethernet Signaling
- Differential Manchester
- -0.85 V and 0.85 V
37Ethernet Frame Format
- Preamble - 7 bytes of 10101010
- Start of Frame Delimiter - SOFD - 101010111
- Destination Address - 2 or 6 bytes - MAC
Addresses, Physical Addresses - Broadcast Address - Source Address - 2 or 6 bytes
- Length of Data Field - 2 bytes, valid frames must
be at least 64 bytes long, 64-1500 bytes. Why? - Data - 0 to 1500 bytes
- Padding - 0 to 46 bytes
- Checksum - 4 bytes or CRC-32 EDC
38Ethernet Frame - IEEE 802.3 Frame Format
3910Base5
- Thick Ethernet, officially known as 10Base5
- Is the oldest form of Ethernet. It was originally
developed in the late 1970's by Digital Equipment
Corporation, Intel and Xerox - Became an international standard in 1983
4010BASE5 (Thick Ethernet)
- Original 802.3 medium specification
- 50-O coax and Manchester signaling
- Segment length can be extended past 500m with
repeaters - transparent at the MAC level
- maximum of 4 allowed
- No looping allowed--one path between any two
stations
41Topology Cabling
- 10 Base 5 is laid out in a bus topology, with a
single coaxial cable connecting all nodes
together - At each end of the coaxial cable is a terminator
- Each node on the network physically connects to
the coaxial cable through a device called a
transceiver and an AUI cable is connected between
the node and the transceiver
42Thick Ethernet Physical Topology
4310Base5 PROS and CONS
- PROS
- Long Distances Possible
- Noise Immunity
- Conceptually Simple
- CONS
- Inflexible
- Fault Intolerant
- Susceptible To Ground Loops
- Very Difficult Troubleshooting
4410Base5 as a Backbone
4510BASE2 (Thin Ethernet)
- Intended to provide lower-cost system for PC LANs
- Uses thinner cable and supports fewer taps than
10BASE5 - Can combine 10BASE2 and 10BASE5 segments in the
same network (but backbone must then be 10BASE5)
4610Base2 Basics
- Thin Ethernet, officially called 10Base2, is a
less expensive version of 10Base5 (Thick
Ethernet) technology - It uses a lighter and thinner coaxial cable and
dispenses with the external transceivers used
with 10Base5
4710Base2 Topology Distance Limits
- 10 Base-2 uses an RG-58A/U coaxial cable and is
wired in a bus topology. Each device on the
network is connected to the bus through a BNC "T"
adapter, and each end of the bus must have a 50
Ohm terminator attached. Each node on the bus
must be a minimum of 0.5 meters (1.5 feet) apart,
and the overall length of the bus must be less
than 185 meters (606 feet). Figure One
graphically illustrates a sample 10 Base-2
network. - NOTE It is not possible to locate the node away
from the BNC "T" adapter by connecting a coaxial
cable between the "T" and the node's BNC
connector. Doing so will detune the network, and
will almost certainly result in major problems.
4810Base2 Network
4910BASE-T
- Uses UTP, often prewired in buildings
- Star-shaped topology is well-suited to existing
wires terminating in a closet - Stations attach to central multi-port repeater
(hub) - Hubs can be cascaded
- Physical star, but logical bus (all transmissions
are repeated)
5010BaseT
- Ethernet was originally designed to operate over
a heavy coaxial cable, and was later updated to
also support a thinner, lighter, coaxial cable
type. Both systems provided a network with
excellent performance, but they utilised a bus
topology which made changing a network a
difficult proposition, and also left much to be
desired in regard to reliability. - Also, many buildings were already wired with
twisted-pair wire which could support high speed
networks. Installing a coaxial-based Ethernet
into these buildings would mean they would have
to be rewired. Therefore, a new network type
known as 10 Base-T was introduced to increase
reliability and allow the use of existing
twisted-pair cable.
5110BaseT Topology Cabling
- 10 Base-T utilizes Category 3 (or higher)
Unshielded Twisted Pair (UTP) cable in a star
topology. Each node on the network has its own
cable run back to a common hub, and each of these
cable runs may be up to 100 meters (330 feet) in
length. Figure One shows a simple 10 Base-T
network.
5210 Base-T Network
5310BaseT Tree Topology
- 10 Base-T can also be wired in a tree topology,
where one "main" hub is connected to other hubs,
which are in turn connected to workstations.
Please note that the depth of a 10 Base-T tree
network is limited to one layer below the main
hub. It is also possible to combine 10 Base-T
with any combination of the other 10 Mbps
Ethernet technologies in an infinite number of
ways to meet nearly any requirement. Figure Two
shows a combination of 10 Base-T and 10 Base-2.
5410 Base-T and Thin Ethernet Combination
55Advantages Disadvantages
- 10 Base-T has various advantages and
disadvantages which make it suitable for some
applications and less suitable for others. Some
of them are listed below - Fault Tolerant
- Uses UTP Cable
- Easy Moves Changes
- Easy Troubleshooting
56Advantages
- Fault Tolerant
- Since each node on a 10 Base-T network has its
own cable connecting it to a central hub, it is
far less likely that any node can cause the
entire network to fail. The hub also has a
"partitioning" function built into it which
allows it to detect a problem on any of its
ports. If a problem is found, the node is
disconnected from the rest of the network. This
isolates the problem until the node can be
troubleshot and repaired. - Easy Troubleshooting
- Because of the partitioning function built in to
the hubs and the star-wired topology, it is
generally easy to troubleshoot a 10 Base-T
network. In a worst-case scenario, one can be
troubleshot by simply disconnecting nodes from
the hub one at a time until the network recovers.
Usually, the hub will give an indication as to
which node is causing a problem, allowing the
technician to troubleshoot that node as opposed
to spending many hours finding where the problem
is. - Easy Moves Changes
- Disconnecting a node from the network has no
effect whatsoever on the rest of the network.
Therefore, moving an attached device is simply a
matter of unplugging it from the hub and
reconnecting it somewhere else. - Uses UTP Cable
- Many buildings are already wired with UTP cable
which can support a 10 Base-T network. Even in
the event a building is not wired with UTP
already, it is still preferable to install UTP
than any other type of cable, as UTP will support
other applications later, whereas other cable
types will generally be specific to one network
type. This allows leveraging the UTP cable
investment for other applications many years
later.
57Disadvantages
- Distance - 10 Base-T only allows distances from
the hub to the node of 100 meters (330 feet). In
some installations, this can be a major problem
if nodes need to be located farther away. - Sensitive To Noise - The nature of UTP cable
makes it considerably more sensitive to
electrical noise than coaxial cable. Generally,
this rules 10 Base-T out as an option for
installations on factory floor environments or
other locations with a high ambient noise level
5810BaseT vs. 10Base2/5
5910BaseT Cabling Considerations
- 10 Base-T uses two pairs of wires one pair for
transmission and the second pair for receive. - The physical connector used is an 8 position
modular plug, commonly referred to as an RJ-45. - All cables must be rated at a minimum of Category
3, and must be wired such that pins 1 2 are on
one twisted pair and pins 3 6 are on a second
pair. Common wiring standards which meet this
requirement are EIA/TIA T568A and T568B.
6010BaseT Cabling Considerations - II
- There are two pinouts used MDI for DTE devices
(such as computers, printers, etc.) and MDI-X
(hubs). Connecting an MDI port to an MDI-X port
requires a straight through cable, and connecting
either MDI to MDI or MDI-X to MDI-X requires a
crossover cable. Pinouts of the MDI and MDI-X
interfaces are shown in Table One.
61Crossover Cable Applications
- There are several applications for crossover
cables in 10 Base-T networks. The most common
reason is to cascade hubs together in a tree
topology. If both hubs have only MDI-X ports then
a crossover cable is needed. Another application
for a crossover cable is to connect two DTE
devices together without a hub.
62Crossover Cable Pinout
- A standard 10 Base-T crossover cable wiring
diagram is shown in Table
6310BaseT Hub Evolution
- Hub
- Active Hub
- Manageable Hub
- Switched Hub
- Switch
6410BROAD36
- Only 802.3 broadband spec
- Uses 75 -O CATV coax
- Maximum length of individual segment is 1800m
- Broadband is by nature analog, so analog encoding
must be used (DPSK)
6510BASE-F
- Standard includes 3 specifications
- 10-BASE-FP Passive star topology, up to 1km per
segment - 10-BASE-FL Point-to-point link connecting
stations or repeaters up to 2km - 10-BASE-FB Point-to-point backbone link
connecting repeaters at up to 2km - All specs use two fibers, one for transmission in
each direction - Manchester encoding, converted to optical signal
elements
6610BaseFL
- 10 Base-FL is basically a version of Ethernet
which runs over fiber optic cable. In physical
topology, it is very similar to 10 Base-T. - 10 Base-FL runs over 62.5/125 micron multimode
fibre optic cable. It supports distances of 2000
meters (6600 feet). - 10 Base-FL is wired in a star topology with all
of the fibre optic runs originating from a
central hub. It is also acceptable to connect a
pair of 10 Base-FL devices directly together with
a point to point link. - One point which must be made is that although 10
Base-FL links can be run up to 2000 meters, an
Ethernet network has an overall maximum network
diameter (distance between the two most widely
separated nodes) of 2500 meters. Therefore, it is
highly recommended that anyone planning to
install a long 10 Base-FL cable run checks the
overall network to ensure that adding the new run
does not exceed the overall network size. If the
overall size will exceed 2500 meters, then a
local Ethernet bridge or a switch will need to be
installed to split the network into multiple
subnetworks, each of which can be built to the
full 2500 meter limit.
6710BaseFL Applications
- Since 10 Base-FL uses fibre optic cable, it is
completely immune to any type of external
electrical interference. This makes it a very
attractive technology for use in areas such as
factory floors where electromagnetic interference
(EMI) levels are very high. - Another property of fibre optics is that fibre is
very difficult to tap into, thus providing a high
level of security. - Finally, 10BaseFL is very useful for use in
interconnecting buildings in a campus environment
where distances could be very long.
68Comparison
69Problems with Ethernet
- CSMA/CD Protocol - Contention, Collisions
- Distance Limits
- Capacity Limits
70Pros and Cons of CSMA/CD
- Every Ethernet network, regardless of type,
behaves in this manner. Naturally, this access
method is most efficient under periods of light
load, as there is a greater chance the network
will be available when the device wants to
transmit, and there is also less chance another
device will try to transmit at the same time. As
the load on the network increases, collisions
increase with it. Eventually, it is possible to
get to the point where there are too many
collisions and devices spend more time
retransmitting and waiting than they do actually
sending data. - The CSMA/CD protocol is usually robust enough to
keep the network running very well, however there
are situations which can cause it to fail. Proper
design of the network is important to prevent a
breakdown of the basic CSMA/CD protocol. Design
mistakes which cause problems are most often
either exceeding the maximum repeater count or
exceeding the distance limit of a network
segment. - One thing which should be kept in mind is that
CSMA/CD happens automatically in each machine's
network interface card, and there is nothing the
user needs to program or set up to implement it,
nor is the user usually notified of collisions.