Computer Networks

Dr. Małgorzata LangerB9 - Institute of Electronics;

The 3rd floor, r.No. 310

Recommendedreferences

Computer Networks – Protocol, Standards, & Interface; Uyless Black, Ed. Prentice-Hall International, Inc.Enhanced IP Services For CiscoNetworks; Donald C. Lee, CiscoSystems, Cisco PressManaging IP Networks with CiscoRouters; Scott M. Ballew, O’Reilly andAssociates, Inc.

What is a computer

network?

� A number of computers (and/orterminals) interconnected by one ormore transmission paths.

The network exists to meet one goal: totransfer and to exchange data betweencomputers and terminals

Let’s define some terms

APB1

APB2

APA1

APA2

DTE ADCE A DCE B DTE B

Data Bases

Files

Data Bases

Files

Local Connections

Communications channel

Local Connections

Applicationprocess

Let’s define some terms

APB1

APB2

APA1

APA2

DTE ADCE A DCE B DTE B

Data Bases

Files

Data Bases

Files

Local Connections

Communications channel

Local Connections

Data terminal equipment

The DTE can be:

A work stationA personal computerA sampling device to measure thequality of air or any other ‘smart’ sensorA point-of-sale terminal in a super market.......

Let’s define some terms

APB1

APB2

APA1

APA2

DTE ADCE A DCE B DTE B

Data Bases

Files

Data Bases

Files

Local Connections

Communications channel

Local Connections

Data circuit-terminatingequipment

OrData

communicationsequipment

The primary function of the DCE is to

provide an interface of the DTE into the

communications network

The interfaces are specified andestablished through protocols

Adapted common interfaces and protocolsthat are vendor- and product- independent, are recommended standards

Logical and PhysicalCommunications

APB1

APB2

APA1

APA2

DTE ADCE A DCE B DTE B

Data Bases

Files

Data Bases

Files

Local Connections

Communications channel

Local Connections

DCEs and DTEs are

connected:

Point-to-point (Only two DTEs are on the channel)In Multidrop Circuits (More than twoDTEs are on the same channel)

DTE

DTE DTE DTE

DCE

DCEDCE

DCE

The methods to send the traffic

(data flow):

Simplex – one direction only (TV, radio)Half-Duplex – both directions, but one direction at a time – TWA (two-way-alternate)Full-Duplex (or Duplex) –simultaneously in both directions – TWS (two-way simultaneous)

DSE – Data Switching Equipment

DTE

DTE

DTEDTE

DCE

DCE

DCEDCE

DCE

DCE

DCEDCE

DSE

And other DSEs, DCEs, DTEs .....

Network Topologies:

Provide maximum possible reliability to assure proper receipt of all traffic(alternative routing)Route the traffic across the least-costpath within the network between thesending and receiving DTEsGive the end user the best possibleresponse time and throughput

The more common network

topologies:

The hierarchical topology (TREE)The horizontal topology (BUS)The star topologyThe ring topology (HUB)The mesh topology

The hierarchical topology(Tree Topology, Vertical Topology)

Advantages:- a simplicity of control- a concentration point for error resolution- clear lines of authority- subordinate DTEs canbe added relativelyeasily

A

B C D

E F G

The hierarchical topology(Tree Topology, Vertical Topology)

Disadvantages:

- A BOTTLENECK!!

- reliability problems

A

B C D

E F G

Horizontal Topology (Bus)

A single stationbroadcasts to multiplestations. All the stationsreceive everytransmission

Advantages: relativelysimple and cheap

A B C

D E

Horizontal Topology (Bus)

Disadvantages:

Only one channel exists to service all the devices on the network – A RISK OF FAILURE!Difficulties in isolatingfaults

Remedies: fully redundantchannels, bypass switches

A B C

D E

Star Topology

Historically the firstone, - easy to control; - the software isn’tcomplex, - the traffic flow issimple- the fault isolation israther simple

AB

C D

E

FG

Star Topology

Disadvantages:Limited distributedprocessingcapabilities, otherslike the hierarchicaltopology

Remedies: fullyredundant hub node

AB

C D

E

FG

Ring Topology

Data flow in one direction only, withone single stationreceiving the signaland relaying it to thenext station on thering.No bottlenecksThe logic isrelatively simple

A

B

CD

E

Ring Topology

Disadvantages: One channel ties all thecomponents (one failing part causesthe entire net is lost)Remedies: switchesto route the data around the failednode; the use ofdual rings

A

B

CD

E

Mesh Topology

Due to themultiplicity of pathsfrom DTEs andDSEs, traffic can be routed around failedcomponents or busynodes.The approach isattractive thoughexpensive

AB

CD

E

F G

DNHR – DynamicNonhierarchical Routing

The technology allows a choice of thepath based on heavy overflow trafficfrom the fixed topology.

Major portions of DNHR have replacedmuch of the hierarchical network

WAN & LAN

WAN – Wide Area NetworkLAN – Local Area Network

Applied options strongly depend on thedivision above

Switched and Nonswitched Options‘Does the channel belong to us, only?’switched nonswitched

Advantages:FlexibleInexpensive for lowvolume

Disadvantages:Slow response, Blocking possibleLow qualityExpensive for highvolume

Advantages:Supports higher volumeHigher quality possibleNo blockage

Disadvantages:Expensive for lowvolumeLack of flexibility whenline is inoperable

Some

telecommunicationbasics

The first electronic transmission ofinformation was in the form of Morse codeSignals can vary over time or space

Continuous time domain – analogue signals; discrete time domain – discrete, digital, binarysignalsIf the dependent variable is processed insome way – it is an aggregate signal

Definitions:

Digital technique –uses discrete signals

Discretisation in time -sampling

Discretisation of signal values –quantisation and coding of the obtained discrete signal

Discrete bivalent signal („1”; „0”) –digital signal; numerical signal, binary signal, logical signal

Maximum frequency: the highest allowable frequency of input signal changes at which the system works correctlyMargin of interferences (noise margin): such a value of interfering signal that added to the input signal value will not cause any change in the logical value of the signal, stillPower of losses (rated dissipation): the difference between the applied power and the power output;It can be heat emmissionin devices

Power ratios

The earliest measurement used to categorizethe quality of transmission on a circuit (to definethe gain or loss in power) was the bel (B):

( )

II

I

PPwhenisdBP

PdB

BP

PB

=

=

−=−=

=

00

10

1010010

100

10

0log101

110logloglog1

Physical limits

As a signal propagates down a transmissionmedium its phase can become distorted withrespect to its frequency (SHIFT IN PHASE)NOISE: two types of noise affect the ability ofa signal to be recognized: impulse noise andthermal noiseFREQUENCY SHIFT (or PHASE JITTER -the short deviation in or displacement of the pulses in a high-frequency signal)

HOW TO PUT INFORMATION TO THESIGNAL? There are three primary waveform

characteristics, than are vital to data communications:

- amplitude, frequency, phase

The Analog Signal A:900 Phase; B:1800Phase…..

Time

An amplitude – a measurement in relationto its voltage, which can be zero, a plus, orminus valueA cycle – the complete oscillationA frequency (‘Hertz’) – the number ofoscillations per secondA baud – the rate of signal change on thechannelfor example: an 1800 Hz signal can be changed at 1200 times per second – 1800 Hz describes ‘the carrier’; 1200 changes describe the baud

The signals use subchannels

Several transmissions at differentfrequencies occupy the same physicalmedium (for example, a wire)

The modem –

modulation/demodulation

Amplitude 0

Binary 0 Binary 1 Binary 0 Binary 0

Cycle 1 Cycle 2 Cycle 3 Cycle 4

Modulation Techniques

Channel Speed

The communications channel isdescribed by its capacity in the numberof bits per second transmitted

Abbreviations are: bit/s; bps; bs

9600 bs; 14400 bs; ... 64kbs, 1.544Mbs, 2.048Mbs

Bandwidth and the

Frequency Spectrum

A telephone channel in Europe occupies a band from 300 Hz to 3400 Hz (in NorthAmerica: 300-3300Hz)

For example: the bandwith between thefrequency spectrum 103 – 104 is 9000 Hz, but 104-105 is 90000 Hz already and isequivalent to more channels by 3100Hz

Synchronization

Self-clocking codeNon-self-clocking code

Clocking signals (1) synchronize thereceiver into the transmission before thedata actually arrive, and (2) they keepthe receiver synchronized with theincoming data

Synchronization Codes

Unipolar code – no signal below zero, orno signal abovePolar code – opposite algebraic signsidentify logic statesBipolar code – the signal varies amongthree levelsAlternate mark inversion (AMI) code –‘1’ is coded with the alternate polarity

Digital codes - examples

0 1 0 1 1 0 0

NRZ

RZ

Manchester

Bipolar AMI

Real pulse stream

Bipolar AMI

NRZ – non-return-to-zero code

Very efficient use of bandwidth (a bit witheach signal change); requires no ecodingor decoding; relative simplicity, low costThe lack of self-clocking capabilities

Widely used

RZ – return-to-zero code

Entails the changing of the signal stateat least once in every bit cell – verygood synchronization characteristicsBut it requires two signal transitions for each bit (twice the baud of NRZ code)

Used in lightwave technologies, opticfibers

Manchester code

The signal state is valid during the firsthalf of every bit cell – very goodsynchronization characteristicsBut it requires two signal transitions for each bit (twice the baud of NRZ code); the interface devices are expensive

Used in local area networks, optic fibersetc.

Bipolar AMI

Used by AT&T, The Bell OperatingCompanies, and many other companiesin USAProblems occur when a long series ofzeros should be placed in the system –there is no way to synchronize with zero bit cells

Asynchronous and

Synchronous Transmission

In NRZ (non-return-to-zero) codeclocking becomes a major considerationfor DTEs and DCEsIn asynchronous transmission eachdata byte (each character) has START and STOP signalsAsynchronous transmission is widelyused because the interfaces in DTEsand DCEs are inexpensive

Byte N

stop

start

stop

start

ByteN-1

…Byte3 Byte2 Byte1

Asynchronous format

Sync Byte N …. . Byte 3 Byte 2 Byte 1 Syn

Synchronous Format

Control Error Check User Data Control Address Sync

A typical synchronous transmission

The preliminary signals are called:

Synchronization bytesSync bytesFlagsPreambles

And their principal function is to alert thereceiver, that user data are arriving.

This process is called FRAMING

MESSAGE FORMATS

Sync bytes

An identification (address) of the dataControl field(s)

User dataFCS field (frame check sequence)

A FRAME is the basic unit of informationtransmitted across the communicationschannel

The data field in a frame remainstransparent to the frame level protocol

A substantial amount of network traffic is non-user, overheadframes

The identification (ID) field usuallyprovides a number or a name for thereceiver as well as for the transmitter.Either the ID or control fields containsequence numbers which are used to further identify the specific frames fromeach sender

The error-check field is appended by thetransmitting site. Its value is derived froma calculation on the contents of the otherfields.At the receiving site an identical processcomputes another error-check field. Thetwo are then comparedThis process is called CRC (cyclicredundancy check)The field is called FCS (frame checksequence)

Additional Network Components

Front-end-processor (FEP)Multiplexer (MUX)Data service unit (DSU)Data switch (DSE)Private branch exchange (PBX)Cluster controller (CC)

A TypicalComputer/CommunicationStructure

Several protocols cooperate to manage thecommunications, among others:

Data link control (link protocol)Switching protocol (routing protocol)

The DTE in San Francisco is to transmit data

to a remote computer in Atlanta. The

transmission goes through an intermediatepoint, a computer in Dallas (as DSE)

Events:

1 – San Francisco DTE sends the data2 – Dallas sends an acknowledgmentACK – a positive acknowledgmentNAK – a negative acknowledgment

The processes 1 and 2 areCOMPLETED BEFORE event 3 occurs

3 and 4 – continuing…5 (if added) give the provision for anend-to-end acknowledgment

Checking for Errors

Cyclic redundancy checking (CRC) method (an algebraic expression to divide the constant into a binaryrepresentation of a data field. Thequotient is discarded, but the remainderis retained and used at the receiver to check for transmission errors)

Echoplex – the remote site sends back(echoes) each character to the sendingPCDesigned for two-way, full-duplex circuitsLocal echo – when a half-duplex modem is connected to the telephone channel, the data are echoed back from the localmodemParity checking – a single bit (a parity bit) is added to each string (for example to give the character bits an odd number ofbits that are 1s)

Double parity – this technique places a parity (odd or even) on a block ofcharacters – (two dimensional paritycheck codeChecksum – the transmitting PC addstogether the numeric values of all thecharacters in the transmission. The totalis placed into a 16-bit block-check countby using the least significant 16 bits ofthe calculation. This value is transmittedwith the user data to the receiving DTE

WAN and LAN

Wide area networkThe DSE acts as the PAD (packetassembly/disassembly) into and out ofthe networkThe NCC (network control center isresponsible for the efficient, reliableoperations of the network

A DSE/

PAD

WAN

LAN – local area

networkSignificantly different from a WAN:Links are usually owned by the userorganizationLinks operate on very high-speed lines (to 400 Mbit/s)DTEs are located closely together (a plant, a building..)Links are of better quality than WAN channelsTopology tends to be more ordered andstructured

CONNECTION-ORIENTED &

CONNECTIONLESS NETWORKS

Connection- oriented – no logicalconnection initially exists between theDTEs and the network (idle state)In order to communicate a ‘handshake’ is neededThe user data are exchanged through a preestablished protocolThe DTEs perform a connection release– a return to the idle condition

Idle – no connection

Connection established

Data transfer

Connection release

Idle – no connection

Connectionoriented

ACKs

flow control

error recovery

Connectionless (datagram) networkgoes directly from an idle condition into a data-transfer mode, followed directly by theidle condition

Comparable to mailing a letter (a message isplaced into the postal system with theassumption it will arrive at its destination)

Idle – No connection

Data Transfer

Idle – No connection

Connectionless

No ACKs, no flow control, no error control

But one can push error control up into the application process (or a higher-level protocol)

Data link protocols

(DLC)

They follow the steps:

-Link establishment (DLC ‘handshake’ with the remote DLC logic to ensure both systems are ready to exchange user data

- Information transfer (DLC checks all data for possibletransmission erors and sends acknowledgments back to thetransmitting machine

- Link termination (DLC must relinquish the control of the link)

Types of DLC

Primary/secondary (master/slave)one DTE, DCE or DSE is designates as the primary station on the channel andcontrols all the other stations anddictates when and if the devices cancommunicatePeer-to-peer – equal status to allstations

POLLING

Primary Site 1 Secondary Site 2

POLL

DATA

ACK DATA

EOT

Poll sequence

Selection (some systems use pollingonly!)

Primary Site 1 Secondary Site 2

SEL

ACK

DATA

ACK DATA

EOT

Select command

Definitions:

Digital technique –uses discrete signals

Discretisation in time -sampling

Discretisation of signal values –quantisation and coding of the obtained discrete signal

Discrete bivalent signal („1”; „0”) –digital signal; numerical signal, binary signal, logical signal