127
COMP210 Physical Layer Dr Ahmad Al-Zubi

COMP210 Physical Layer Dr Ahmad Al-Zubi. The Physical Layer The Physical Layer performs bit by bit transmission of the frames given to it by the Data

Embed Size (px)

Citation preview

COMP210

Physical Layer

Dr Ahmad Al-Zubi

The Physical Layer

• The Physical Layer performs bit by bit The Physical Layer performs bit by bit transmission of the frames given to it transmission of the frames given to it by the Data Link Layer.by the Data Link Layer.

• The specifications of the Physical The specifications of the Physical Layer include:Layer include:

• Mechanical and electrical interfacesMechanical and electrical interfaces

• Sockets and wires used to connect the Sockets and wires used to connect the host to the networkhost to the network

• Voltage levels uses (e.g. -5V and +5V)Voltage levels uses (e.g. -5V and +5V)

• Encoding techniques (e.g. Manchester Encoding techniques (e.g. Manchester encoding)encoding)

• Modulation techniques used (e.g. square Modulation techniques used (e.g. square wave)wave)

• The bit rate and the baud rate.The bit rate and the baud rate.

Signal Transmission

Electronic energy to send signals that communicate from one node to another

Two methods of transmitting data Digital signalingAnalog signaling

Comparison of Digital and Analog

Bandwidth-Limited Signals

Bandwidth-Limited Signals

Digital Signaling

Digital signal represents discrete state (on or off)

Practically instantaneous change

Digital SignalingCurrent State

Encoding Data is encoding by the

presence or absence of a signal

A positive voltage might represent a binary zero or binary one or visa versa

The current state indicates the value of the data

Digital Signaling Current State

Digital Signaling State Transition

Analog Signaling

Signals represented by an electromagnetic wave

Signal is continuos and represents values in a range

Uses one or more of the characteristics of an analog wave to represent ones and zeros

Characteristics of an Analog Signal

Parts of a Wave The maximum intensity of a wave is

called the amplitude.

The distance between two crests is the wavelength.

The number of complete wave cycles every second is the frequency.

The phase difference measures (as an angle) how far ahead one wave is when compared to another wave.

Phase:

Relative state of one wave to another in regards to timing

Amplitude (a)

Wavelength ()

or period

Phase differenc

e ()

ELECTROMAGNETIC SPECTRUM

Lambda = c / f = (3 * 10**8)[m / sec] / f

f = 10 MHz -> lambda = 30 mf = 10 GHz -> lambda = 3 cmf = 10**15 Hz -> lambda = 3 * 10**-4 mm

Three equations describing data transmissions

1. lambda = c / frequency lambda is the wave length c is the speed of light (3 * 10**8 m/sec)

2. Max_data_rate = 2 * freq * log(2) #of_levels

3. delta(freq) = c * delta(lambda) / lambda ** 2

Wire Propagation Effects

Propagation Effects Signal changes as it travels Receiver may not be able to recognise it

Distance

OriginalSignal

FinalSignal

Propagation Effects: Attenuation

Attenuation: signal gets weaker as it propagatesAttenuation becomes greater with

distanceMay become too weak to recognise

SignalStrength

Distance

Propagation Effects: Distortion

Distortion: signal changes shape as it propagatesAdjacent bits may overlapMay make recognition impossible

for receiver

Distance

Propagation Effects: Noise

Noise: thermal energy in wire adds to signalNoise floor is average noise energyRandom signal, so spikes

sometimes occur

SignalStrength

Time

Signal

Noise

Spike

Noise Floor

Propagation Effects: SNR

Want a high Signal-to-Noise Ratio (SNR)Signal strength divided by average

noise strengthAs SNR falls, errors increase

SignalStrength

Distance

Signal

Noise FloorSNR

Propagation Effects: Interference

Interference: energy from outside the wireAdds to signal, like noiseOften intermittent, so hard to

diagnose

SignalStrength

Time

Signal

Interference

Propagation Effects: Termination

Interference can occur at cable terminator (connector, plug)Often, multiple wires in a bundleEach radiates some of its signalCauses interference in nearby wiresEspecially bad at termination, where

wires are unwound and parallel

Termination

Bandwidth

Capacity of a media to carry information

Total capacity may be divided into channels

A channel is a portion of the total bandwidth used for a specific purpose

Bandwidth

BasebandThe total capacity of the media

is used for one channelMost LANs use baseband

BroadbandDivides the total bandwidth

into many channelsEach channel can carry a

different signalBroadband carries many

simultaneous transmissions

Analog versus Digital

Digital Is less error proneDistortion of the signal between

the source and destination is eliminated

AnalogLittle control over the signal

distortionOld technology

Analog versus Digital

voltage

digital signal

time

analogue signal

time

voltage

time

voltage

analogue signal + noise

time

voltage

digital signal + noise

When noise When noise effects an effects an analogue analogue signal, it is hard signal, it is hard to deduce the to deduce the original signal.original signal.

The original The original digital signal digital signal can be can be deduced deduced despite the despite the noise.noise.

In digital communication, it is often possible to reconstruct the original signal even after it has been effected by noise

Benefits of Digital Transmission

Reliability

Can regenerate slightly damaged signals

There are only two states. Change to closest

E.g., if two states are voltages +10v (1) and -10v (0) and the signal is +8v, the signal is a 1

Original ReceivedRegenerated

Benefits of Digital Transmission

Error detection and correctionCan correct errors in

transmission

- Add a few bytes of error-checking information

- Can ask for retransmission if an error is detected

Benefits of Digital Transmission

EncryptionEncrypt (scramble) messages so

that someone intercepting them cannot read them

Benefits of Digital Transmission

CompressionCompress message before

transmissionDecompress at other endCompressed message places

lighter load on transmission line, so less expensive to send

Not always used

10101001 1010

OriginalSignal

CompressedSignal

• Because attenuation is frequency dependent, modems use a sine wave carrier of a particular frequency, and then modulate that frequency. Various modulations include:

Amplitude modulation: Two different amplitudes of sine wave are used to represent 1's and 0's.

Frequency modulation: Two (or more) different frequencies, close to the carrier frequency, are used.

Phase modulation: The phase of the sine wave is changed by some fixed amount.

Binary Signal

Modulation

Nyquist’s Limit

Suppose we know the bandwidth (H) of a channel and the number of signal levels (V) being used. What is the maximum number of bit we could transmit?

Nyquist’s Limit says:

Max_bps = 2*H*log(base 2)(V)

For example, if the bandwidth is 3100Hz and we are using 16 level modulation then the maximum number of bits per second is:

max_bps = 2 3100 log2(16) = 24800 bps

bits/sec

Analog Signaling Amplitude Modulation

(ASK)

Analog Signaling Frequency Modulation

(FSK)

Analog Signaling Phase Shift Keying

(FSK)

Constellation Diagrams

We can represent the different combinations of phase shifts and amplitudes using a constellation diagram.

Each dot represents a combination of phase shift (the angle from the positive x axis) and amplitude (the distance from the origin).

15

45

000

111

110

101

100010

001

011

0110

1000

0000

Baud Rate / Bit Rate

The maximum number of times a signal can change in a second is called the baud rate.

The number of bits (1s and 0s) transmitted in a second is called the bit rate.

In the examples we have seen so far, the bit rate and the baud rate are the same. This is not always the case.

The bit rate can be higher than the baud rate if we use more than two signal levels (more later on…).

Modems

Modulation demodulation Used to connect a digital

computer to an analog phone system

Can be installed internally a card inserted into the motherboard

Can be connected to the serial port (external modem)

Modems

Transfer speedsBit rate BPSBaudBandwidth

Compressionappears to increases speed by

decreasing the number of bits sent (usually some data does not compress well)

sending and receiving modem must use same compression standard

Modems

Error detection and correctionasynchronous modems use

parity checkchecksum counting the number

of data words sent

Digital Modem

Miss named Used to connect to a

digital telephone ISDN (integrated Services

Digital Network) is an example

Again do not connect digital modems to an analog phone

Higher quality lower errors

Transmission Modes

Parallel Mode

Serial Mode

Channel Types

A channel is any conduit for sending information between devices.

There are three basic types of channel: simplex, half-duplex and full-duplex.

A simplex channel is unidirectional, which means data can only be sent in one direction. For example, a TV channel only carries data from the transmitter to your TV set. Your TV set cannot send information back.

Channel Types

A half-duplex channel allows information to flow in either direction (but not simultaneously).

Devices at either end of the channel must take it in turns to transmit information whilst the other listens. For example, a walky-talky either transmits or receives but not both at the same time.

Channel Types

A full-duplex channel allows data to be sent in both directions simultaneously.

A full-duplex channel can be formed from two simplex channels carrying data in opposite directions. This may make it more expensive than a half-duplex channel.

There is no waiting for turns or for the devices swap roles, as is the case with a half-duplex channel. This means full-duplex can be faster and more efficient.

Network Media Types

Types of MediaCable (conducted media)

- Coaxial- Twisted pair (UTP)- Shielded twisted pair (STP)- Fiber optic

Radiated- Infrared- Microwave- Radio- Satellite

Media Selection Criteria

CostFor actual media and

connecting devices such as NICs hubs etc

InstallationDifficulty to work with mediaSpecial tools, training

Media Selection Criteria

CapacityThe amount of information that

can be transmitted in a giving period of time

Measured as- Bits per second bps (preferred)- Baud (discrete signals per

second)- Bandwidth (range of

frequencies)

Media Selection Criteria

Node CapacityNumber of network devices that

can be connected to the media

AttenuationWeakening of the signal over

distance

Media Selection Criteria

Electromagnetic Interference (EMI)Distortion of signal caused by

outside electromagnetic fieldsCaused by large motors,

proximity to power sources

Other noise sourcesWhite (Gaussian) noiseImpulse noiseCrosstalkEcho

Cable Media

Unshielded Twisted Pair UTP

Shielded Twisted Pair STP Coaxial Fiber optic

Unshielded Twisted Pair (UTP)

One or more pairs of twisted copper wires insulated and contained in a plastic sheath

Twisted to reduce crosstalk

Unshielded Twisted Pair (UTP)

Categoriescategories 1 and 2

- voice grade - low data rates up to 4 Mbps

category 3- suitable for most LANs- up to 16 Mbps

category 4- up to 20 Mbps

Unshielded Twisted Pair (UTP)

Categoriescategory 5

- supports fast ethernet- more twists per foot- more stringent standards on

connectors

Data grade UTP cable usually consists of either 4 or 8 wires, two or four pair

Uses RJ-45 telephone connector

Unshielded Twisted Pair (UTP)

Shielded Twisted Pair (STP)

Same as UTP but with a aluminum/polyester shield

Connectors are more awkward to work with

Usually comes in pre made lengths

Different standards for IBM and Apple

Shielded Twisted Pair (STP)

Coaxial Cable Coax

Two conductors sharing the same axis

A solid center wire surrounded insulation and a second conductor

Coaxial Cable

Coaxial Cable Coax

Size of CoaxRG-8, RG-11

- 50 ohm Thick Ethernet

RG-58- 50 ohm Thin Ethernet

RG-59- 75 ohm Cable T.V.

RG-62- 93 OHM ARCnet

Fiber Optic Cable

Thin strand(s) of glass or plastic protected by a plastic sheath and strength wires or gel

Transmits laser (single mode) or LED (multi mode)

Single mode more expensive but can handle longer distances

Fiber Optics

Three examples of a light ray from inside a silica fiber impinging on the air/silica boundary at different angles.

Light trapped by total internal reflection.

The main advantage of optical fiber is the great bandwidth it can carry.

There are three main bands of wavelength used.

Wavelen (microns)

0 0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.55 band

1.30 band

0.80 band

Attenuat

dB/km

0.5

1.0

1.5

2.0

Optical Fibers

One problem with optical fibers is that the light pulses become distorted over distance due to dispersion.

Dispersion occurs when photons from the same light pulse take slight different paths along the optical fiber. Because some paths will be longer or shorter than other paths the photons will arrive at different times thus smearing the shape of the pulse.

Over long distances, one pulse may merge with another pulse. When this happens, the receiving device will not be able to distinguish between pulses.

Optical Fibers

Normal fiber optic cable is called multimode because photons can take different paths along it. The more expensive monomode fibre optic overcomes dispersion by having a core so thin that the light can only take one path along it.

Overcoming Dispersion

Fiber Optic Networks

A fiber optic ring with active repeaters.

Fiber Optic NetworksStar connection

Fiber Cables

A comparison of semiconductor diodes and

LEDs as light sources.

Characteristics of Cable Media

Wireless Media

Uses the earth’s atmosphere as a conducting media

Main typesradio wavemicrowave (including satellite)infrared

The Electromagnetic Spectrum

The electromagnetic spectrum and its uses for

communication.

Radio Transmission

(a) In the VLF, LF, and MF bands, radio waves follow the curvature of the earth.

(b) In the HF band, they bounce off the ionosphere.

Radio Wave

Most radio frequencies are regulated

Must obtain a license from a regulatory board (CRTC, FCC)

A range of radio frequencies are unregulated

Radio Wave

Low power single frequencyuses one frequencylimited range 20 to 30 metersusually limited to short open

environments

Radio Wave

High power single frequencylong distance may use repeaters

to increase distanceline of sight or bounced of the

earth’s atmosphereuses a single frequency

Radio Wave

Spread Spectrum

Maintains security of the radio

transmission by:

1. Spreading the carrier signal frequency

2. Modulating the carrier frequency by a Pseudo Random signal

Radio Wave

Spread Spectrum

Two methods:1. Frequency hopping – carrier

frequency keeps on changing according to the Pseudo Random code

2. Direct Sequence Spread Spectrum- Pseudo Random digital signal modulates the carrier directly

Transmitter and receiver must use the

same Pseudo Random code

Radio Wave

Spread Spectrum

Direct Sequence Spread Spectrum (DSSS) is used as the Physical Layer for the Mobile networking (IEEE 802.11)

Radio Wave

Microwave

Terrestrial line of sightuse relay towersuses license frequencies

Communication Satellites

Communication Satellites

One disadvantage of satellite transmission is the delay that occurs because the signal has to travel out into space and back to Earth (propagation delay).

One problem associated with some types of satellite transmission is raindrop attenuation (some waves at the high end of the spectrum are so short they can be absorbed by raindrops).

Communication Satellites

The principal satellite bands.

Communication Satellites

VSATs using a hub.

Globalstar

Infrared

Uses same technology as remotes for T.V.

signals can not penetrate objects

Can be point to point or broadcast

Point to point requires precise alignment of devices

Point less immune to eavesdropping

Multiplexing

several lines (one for each device) enter a multiplexer (mux) at the host side

the host side mux combines all incoming signals

combined signals are transmitted to a mux on the receiving side

Types of Multiplexers

Frequency Division Multiplexing (FDM)

Time Division Multiplexing (TDM)

Statistical Time Division Multiplexing (STDM)

Frequency Division Multiplexing

• The frequency spectrum is divided up among the logical channels - each user hangs on to a particular frequency. The radio spectrum (and a radio) are examples of the media and the mechanism for extracting information from the medium.

Frequency division multiplexed circuit

Frequency Division Multiplexing

One problem with FDM is that it cannot utilise the full capacity of the cable.

It is important that the frequency bands do not overlap. Indeed, there must be a considerable gap between the frequency bands in order to ensure that signals from one band do not affect signals in another band.

FDM is usually used to carry analogue signals although modulated digital signals can also be sent using this technique.

Wavelength Division Multiplexing

• The same as FDM, but applied to fibers. There's great potential for fibers since the bandwidth is so huge (25,000 GHz).

• Fibers with different energy bands are passed through a diffraction grating prism• Combined on the long distance link• Split at the destination

• High reliability, very high capacity

Time Division Multiplexing (TDM)

Like FDM, time division multiplexing (TDM) saves money by allowing more than one telephone call to use a cable at the same time.

Instead of dividing the cable into frequency bands, TDM splits cable usage into time slots. Each channel is given a regular time slot in which to send a PCM signal.

Terminal#1

Terminal#2

Terminal#3

Multiplexer

Device#1

Device#2

Device#3

Demultiplexer

#3 #2 #1#3 #2 #1

Time Division Multiplexing

• In TDM, the users take turns, each one having exclusive use of the medium in a round robin fashion. TDM can be all digital.

Time Division Multiplexing

Statistical Time Division Multiplexing

(STDM) allows connection of more

nodes to the circuit than the capacity of the circuit

works on the premise that not all nodes will transmit at full capacity at all times

must transmit a terminal identification

may require storage

Digital Technology in Telephone Networks

Over the past 30 years, much of the traditional analogue telephone network has been replaced by digital technology.

A device called a codec (coder/decoder) is used to convert analogue voice signals into digital information that can be handled by the digital technology.

WAN Transmission Media

Public Switched Telephone Network

High speed, High bandwidth dedicated leased circuits

High speed fiber optic cable

Microwave transmission links

Satellite links

Services provided by PSTN

Voice – Plain Old Telephone Service (POTS)

Based on the Voice Channel (3600 Hz)

Data transmission servicesconsist of services such as :

- switched 56- X.25- T1, T3 circuits- Frame relay- ISDN- ATM

Switching

Switching send data across different routes

Three types of switchingCircuit switchingMessage switchingPacket switching

Switching

Circuit Switching: A connection (electrical, optical, radio) is established from the caller phone to the callee phone. This happens BEFORE any data is sent.

Message Switching: The connection is determined only when there is actual data (a message) ready to be sent. The whole message is re-collected at each switch and then forwarded on to the next switch. This method is called store-and-forward. This method may tie up routers for long periods of time - not good for interactive traffic.

Packet Switching: Divides the message up into blocks (packets). Therefore packets use the transmission lines for only a short time period - allows for interactive traffic.

Circuit Switching

Connects the sender and receiver by a single physical path for the duration of the session

PSTN uses circuit switching

Before transmission a dedicated circuit must be established

Circuit Switching

Advantagesguaranteed data rateonce connected no channel

access delay

Disadvantagesinefficient use of the

transmission media (idle time)long connection delays (first

time)

Message Switching

Each message is treated as an independent unithas its own source and

destination address

Each is transmitted from device to device

Each intermediate device stores the message until the next device is readystore and forward

Message Switching

Route messages along varying paths for more efficiency

Switching devices are often PCs with special software

Message Switching

Advantagesefficient traffic managementreduces network congestionefficient use of network mediamessages can be sent when

receiver down

Disadvantagesdelay of storing and forwardingcostly intermediate storage

Packet Switching

Packet switching breaks messages into packets

Packets travel different routes (independent routing)

Each packet has its own header information

Packets small enough to be stored in RAM thus quicker than message switching

Packet Switching Advantages

improves the use of bandwidth over circuit switching

can adjust routes to reflect network conditions

shorter transmission delays than message switching (stored in RAM)

less disk spacesmaller packets to retransmit

Packet Switching

DisadvantagesMore RAMMore complex protocols more processing power for

switching deviceGreater number of packets

greater chance for packet loss or damage

Packet Switching

Two methods of packet switchingDatagram packet switchingVirtual circuit packet switching

Datagram Packet Switching

Message divided into a stream of packets

Each packet has it’s own control instructions

Switching devices route each packet independently

Datagram Packet Switching

Switching devices can route packets around busy network links

Require sequence numbers to reassemble

Small packet size facilitates retransmission due to errors

Virtual Circuit Packet Switching

Similar to circuit switching Before transmission of the

sending and receiving device agree on:maximum message sizenetwork pathestablish a logical connection

(virtual circuit)

All packets travel on the same virtual path

CIRCUIT SWITCHED versus PACKET SWITCHED NETWORKS

(a) circuit switching(b) message switching(c) packet switching

NARROWBAND ISDN - WHAT IS IT?

Integrated Services Digital Network: A completely digit, circuit-switched phone system. Integrates voice and non-voice services.

ISDN allows integration of computers and voice. It means that caller ID can be used to look up your account on the computer so that by the time a human answers the phone, a screen has your information already available.

Integrated Services Digital Network

ISDN SYSTEM ARCHITECTURE:

• ISDN uses TDM to handle multiple channels. For home use, the NT1 (Network Terminator) connects the twisted pair going to the phone company with the house wiring. Various ISDN devices can be connected to this NT1.

• Businesses may have more channels active than the home configuration internal bus can handle. So a PBX ( Private Branch eXchange ) is used to provide the internal bus containing more switching capacity. This in turn is connected to NT1.

Integrated Services Digital Network

THE ISDN INTERFACE:

• Typically a number of channels are combined together. In the USA, Primary Rate ISDN contains 23 channels (each 64 kbps carrying voice or data) + 1 channel for signaling and control (16 kbps digital channel.) In Europe, instead of 23 channels, 30 are used.

• The primary Rate is designed to connect to a business with a PBX. As it turns out, most companies now need far more capacity than 64 kbps for the many uses beyond voice. So this is less than adequate.

• N-ISDN may have a life as a connection to homes for people wanting to download images etc. But it's not useful for serious business applications.

B-ISDN (Broadband Integrated Services

Digital Network)

Broadband transmission A type of data transmission in which a single medium (wire) can carry several channels at once (ex. Cable TV).

Baseband transmission one signal at a time (most communication between computers).

B-ISDN will offer video on demand, live television from many sources, full motion multimedia email, CD-quality music, LAN interconnection, high-speed data transport for science and industry and many more, ALL over the telephone line.

A digital virtual circuit capable of 155 Mbps

Underlying technology that makes B-ISDN possible ATM (Asynchronous Transfer Mode)

Comparing Virtual Circuits and Circuit Switching

• Permanent virtual circuits that remain in place for long periods of time.

• Switched virtual circuits that are set up and torn down with each request.

• The method for establishing these circuits is shown in the Figure. The circuit is really entries in a series of switches, each mapping a circuit number onto a forwarding line.

A Virtual Circuit A kind of circuit-switched path implemented with packet switching

The service offered is connection oriented (from the customer's point of view) but is implemented with packet switching. Services offered include:

Why all the interest in ATM? These days it is more common

for companies to want to interconnect networks. There is one international standard for ATM networks and so interconnection is easier.

ATM can be used to provide both LAN and WAN networks.

ATM can be made to behave like other standard networks and so you do not have to throw away all your old equipment.

ATM provides a high speed network.

ATM Technology

ATM is based on Cell Relay technology. This uses virtual circuits to carry small packets (just 53 bytes long) over a predetermined path through the network.

Section of the path can be shared by other virtual circuits, thus ensuring that the network is used more efficiently than in the case of circuit switching.

Little error checking is performed by the network (this keeps overheads down). Instead, the transmitting and receiving hosts are responsible for error checking.

Establishing a Connection

When information needs to be communicated, the sender NEGOTIATES a "requested path" with the network for a connection to the destination.

When setting up this connection, the sender specifies the type, speed and other attributes of the call, which determine the end-to-end quality of service.

The network then determines a path through the network, and sets up a virtual circuit along this path.

An analogy for this is sending mail. One can choose to send 1st class, overnight, 2 day delivery, etc. and can ask for certified mail.

ATM Cells

The format of the ATM cell looks The format of the ATM cell looks like this:like this:

5 bytes (40 bits) are used to carry 5 bytes (40 bits) are used to carry header data and the remaining 48 header data and the remaining 48 bytes carry data.bytes carry data.

The Virtual Channel ID is used The Virtual Channel ID is used identify which virtual circuit the identify which virtual circuit the cell is to be routed along.cell is to be routed along.

Bits:

VCI Payload48 bytes

3

T

T

12

VPI

VCI

P

H CRC

16

P

1

HCRC

8

VPI

Payload Type

Virtual Path ID

Virtual Channel ID

Priority

Header Checksum

ATM Cell Transmission

– – Unlike in synchronous Unlike in synchronous communication, there is no communication, there is no requirementrequirement that cell input rigidly that cell input rigidly alternate among sourcesalternate among sources

– – Continuous cell stream is not Continuous cell stream is not requiredrequired

– – ATM cells can be transmitted on ATM cells can be transmitted on many types of transmissionmany types of transmission technologiestechnologies

– – ATM current common speed is ATM current common speed is 155Mbps (OC3c), 622Mbps155Mbps (OC3c), 622Mbps

(OC12) and 45Mbps (T3) also getting (OC12) and 45Mbps (T3) also getting common.common.

– – Fiber media preferred. Short runs Fiber media preferred. Short runs may be coax or twisted pair.may be coax or twisted pair.

ATM Speed

ATM can deliver data at rates of either 155.52 Mbps or 622.08 Mbps and higher data rates are likely to follow.

ATM can operate at these high speed because specialised switching mechanisms have been developed that can switch the short 53 byte cells very quickly through the network.

ATM can work over a variety of media. Coaxial cable and optical fibre are the most commonly used.

ATM technology is currently being used to develop the next generation of high-bandwidth telephone systems.

ATM Switches

Operation

– Cells arrive at 360000 cells/sec per 155Mbps connection• Cycle time of 2.7 ms

– Many ports per switching fabric• Fixed size cells facilitate fast switching

– Switch cells• Keep discard rate very low• Never reorder cells on virtual circuit

– Problem: Two cells destined for same output– Problem: Head-of-line blocking