Local Area Networks

ContentChapter 14: Advanced Review (Part I)

Anatomy of a network

A set of interconnected resources

Hosts that run network applications software

– Clients and servers– Set of peers

The network infrastructure that interconnects the hosts

– The networking hardwareand software

Network node devices such as routers and switches

Links: cables, connectors, network interfaces

Transmission links

Convey bits, bytes, packets Physical medium

– Copper (or aluminium)– Optical fibre

Glass, plastic– Free-space optical

Infrared (IR): light-emitting diode or laser Visible red or green: laser

– Radio Satellite, microwave link, mobile, wireless LAN, ‘Bluetooth’

Mode– Point-to-point– Shared medium

Multidrop, multicast– Broadcast

Transmission & interconnection media & devices

The main types of interconnection media can be divided into:– Guided media: copper twisted pair, coaxial and optical fibre cables– Unguided media (wireless):

Infrared, microwave (point to point and satellite), radio and laser

Bits travel (propagate) at the speed of light (3 x 108 m/sec) in unguided media and roughly two-thirds this (5µs/km [what speed is this?]) In guided media

Rate at which connection operates (bits are put onto medium) know as transmission rate and depends on interface and properties of medium (measured in bit/sec)

The main types of interconnection device are:– Modems

For digital connection over analogue networks– Multiplexers , repeaters, hubs, switches, routers

For digital connection over digital networks

Network node devices

Router– Determines route packet takes through network & switches

packet onto correct output link Switch, bridge

– Switches packet, byte, bit from one transmission link to anotherunder software control

Hub, repeater– Repeats digital data stream received on one input link to all

output links Patch panel

– Physical, manually re-configurable wiring / cabling interconnect Wiring closet

– Cupboard where any of the above may be sited:distribution point for corridor, building, site network cabling

H

H

Concept of interconnection in a communication network

Shared medium– No hubs, switches, or routers– Examples:

Broadcast radio: wireless LAN, satellite

Shared links and / or hubs/repeaters: basic Ethernet, Token Ring

– Interconnects hosts directly Switches & hubs

– Examples: 10BaseT Switched ethernet:

fast ethernet, gigabit ethernet– Interconnect links

Routers– Interconnect networks

Hence internetwork (or just internet)

H

H

H

HH

H

HH Hub orSwitch

Hub orSwitch

net

netnet

net

net

netR

R

R

R

R

R

Main types:– WAN - wide area network– LAN - local area network– Ownership can be either public or

private– Clientele can be either public or

private Also

– MAN - metropolitan area network– Global networks

Global network examples– Computer networks

The internet, companies’ private networks

– The telephone network Virtual & overlay networks

– Subsets of participants

Types of network

Is this a LAN or a WAN?

The Internet:a set of interconnected overlay networks?

Structure and Infrastructure

OverviewOverview

Basic Techniques

LAN Structure

Circuit-switching and Packet-

switching

LAN Interconnection Services

Course Structure

The first interconnection networks

Inside the computer room– Connect peripheral devices via

device controllers to CPU Controllers used simple hierarchical

master/slave protocols– CPU polls controllers– Controllers poll devices– Devices respond to controller– Controller responds to CPU

Protocol simplicity arises from– Assumed reliability– Use of high-speed parallel links

Facilitates high-speed data transfer

remotemux

computerSystem

remote card/printer

time-sharing

computersystem

localmux

Terminal networks

First WANs connected local and remote terminals to central mainframe

Often called multi-access or time-sharing systems

Developed out of early operating systems work at MIT

– Used PSTN for remote links– Installed/owned local links– Communications need to be serial

(bit-oriented) over these distances– Link error rate typically rather high

Typically between 10-3 and 10-4

Required new protocols that could detect, and perhaps correct, transmission errors

Link layer hardware and software

Most hosts attach to LAN using network interface card (NIC)– Ethernet NICs dominate

WLAN becoming popular– Other possibilities include Token Ring, cable modem, ADSL

modem Routers attach LANs to WAN

– Include a mixture of interface types For example, Ethernet, Frame Relay, X.25, ATM

Network interfaces include– Physical layer (layer 1) components

Timing, coding, and ‘line driver’ chips– Link layer components

Link protocol controller chip Data memory for frame buffers

– Link layer protocols provided as operating system ‘drivers’

Multiplexing: sharing the bandwidth

Historically, the term ‘bandwidth’ refers to the range of allowable frequencies on an analogue link

– Today, we usually mean the available transmission capacity LAN links tend to use the whole of the bandwidth for a single transmission

– We sometimes see the term baseband in this context WAN links tend to be shared in some way, for several reasons

– Chief one is economics (service provider charges a lot for use of total capacity)

– They tend to have greater capacity than is required by one transmission/customer

At the signal level (Layer 1) we call this sharing multiplexing Two basic multiplexing techniques

– Frequency division multiplexing (FDM) for use over analogue links– Time division multiplexing (TDM) for use on digital links– Wavelength division multiplexing is a variation of FDM used on optical

fibre links

FDM: analogue multiplexing

Bandwidth shared on basis of frequencies Normally full-duplex transmission

– Two-way simultaneous transmission– Uses four separate frequencies

For 0 & 1 in each direction of the transmission

Cables with a wide frequency range(e.g. coaxial cables) can have multiple channels, each with own sub-range

– Each channel uses separate frequencyfor 0 & 1 in each direction

FDM used over unguided media– Wireless, microwave, satellite

Can also be used on guided media

T1 T2 T3 T4

0-4kHz 4-8kHz 8-12kHz 12-16kHz

remotemux

WDM: Wave Division Multiplexing

Used on optical fibre Comes in two main flavours

– Colour multiplexing– Dense WDM

Colour multiplexing usually refers to multiplexing of a few wavelengths on to fibre

– For example red and blue light in the visible spectrum Dense WDM refers to multiplexing of many slightly

different optical wavelengths onto a high-quality fibre– Current technology allows up to 120 wavelengths

Each separated by only a few (typically 4-6) nanometers– Each wavelength can carry a digital 40Gbit/s data stream– Equates to 4.8Tbit/s

TDM: digital multiplexing

Usually full-duplex– But if bandwidth is limited, can use

Half-duplex transmission Asymmetric bit rates

Multiplexer allocates each end-system atransmission time-slot

Each device in turn gets all of the line capacity for a small fraction of time

– For example, 3.90625 µsec for a 64 Kbit/s interfaceon a 32-way multiplex onto a 2.048 Mbit/s circuit

TDM used over guided media– Copper, fibre

Can also be used on unguided media

labels

remotemux

Statistical multiplexing

Multiplexing can be based on two techniques– Dedicated– Shared

Non-statistical multiplexing is dedicated– End-system gets part of total capacity (frequency or time slot)

all the time Whether or not the end-system is able to use it

– Demultiplexing performed on basis of known frequency or timeslot

Also called positioned multiplexing Statistical multiplexing is shared

– End-system gets total capacity some of the time, as needed Sometimes called “bandwidth on demand”

– Requires some sort of channel label to identify end-system– Demultiplexing performed on basis on label

Also called labelled multiplexing FDM and TDM systems can use both methods of multiplexing

– WDM uses positioned multiplexing

Structure and Infrastructure

OverviewOverview

Basic Techniques

LAN Structure

Circuit-switching and Packet-

switching

LAN Interconnection Services

Course Structure

Relays: Interconnectionat Different Layers

Router– Layer 3 relay– Lower layers can be LAN or

WAN protocol stacks e.g. Ethernet, PPP,

X.25 Bridge/switch

– Layer 2 MAC sublayer relay– Layer 1 LAN protocol stack

e.g. Ethernet, wireless LAN

Repeater/hub– Layer 1 relay– Can interconnect different

media e.g. copper twisted

pair, optical fibre

Protocol stack A Protocol stack B

Layer nLayer n

Layer n-1 Layer n-1

Layer nrelay

ISO or IETF protocol stacks

Relays above Layer 3

Called Application layer gateways

– Or just “gateways” Interconnect applications

of same generic type, but which use different message formats

– Possibly also different protocols

Examples– IETF-X.400 mail gateway– IP-PSTN gateway– Proxy server (firewall)

Protocol stack A Protocol stack B

Layer nLayer n

Layer n-1 Layer n-1

Layer nrelay

ISO or IETF protocol stacks

PSTN = public switchedtelephone network

Media type

Data rate

(Mbit/s)

Max. cable length

(metres)

Max. number

of stations

per cable

StructuredCabling term

Examples of Ethernet terminolog

y

Twisted

Pair

copper

10, 100, 1000

100 TwoUTP-5,

UTP-5e, UTP-6

10BASE-T,100Base-T,1000Base-T

Optical

Fibre

10, 100, 1000, 10000

Depends on fibre type and data rate

TwoMultimode

(MMF), Single mode (SMF)

10BASE-F,100BASE-

FX,1000BASE-

SX

Notes: (1) UTP-5 is, more correctly, referred to as ‘Category 5 unshielded twisted pair’ cable

(2) Coaxial cable very rarely found in modern LAN cabling

Repeaters and Hubs

Operate at the Physical Layer– Physical Layer relays– Unit of transfer is the bit

Extend domain of MAC protocol– The collision domain– Repeat incoming bits to other

ports MAC frames seen by all systems Systems contend for extended

communication channel

Support a variety of media types– Allows old style shared coaxial

segments to be connected to modern twisted pair segments

Most hubs are just multi-port repeaters

relay logicPhys

1

Phys

2

Coaxial LANsegment

hub

Ph1 Ph1 Ph1 Ph1 Ph1 Ph1 Ph1

relay logic

Ethernet Hubs

Have separate ports for each system – Enhances LAN resilience

Operate over structured cabling systems Can be cascaded (to a limited degree) to

interconnect multiple LAN segments– 10BASE-T: no more than four hubs

Same rules that applied to coax installations)

– 100BaseT: no more than 2 with twisted pair cable

– 1000BaseT: only one hub And even that is very rare

Are becoming increasingly rare as switches get cheaper

– Many users now connect to the LAN via a switch

Coaxial LANsegment

Shared vs. Switched Bandwidth

Example: Twelve users and fourservers share 100Mbit/s LAN

– All in same collision domain– Access time to shared channel

increases as usage increases Solution to increasing congestion:

replace shared LAN with 10/100Mbit/s switch

– Users divided into smallercollision domains

Each receives larger portionof bandwidth

– Switch throughput at leastport speed ½ number of ports

Eight-port switch supports up to400Mbit/s throughput

Switches commonly used for LAN-LAN interconnect– Usually interconnect same technologies

For example, Ethernet to Ethernet– Falling switch prices have killed off the hub market

Switches available for all versions of Ethernet– 10, 100, 1000 and 10000Mbit/

But support for 10Mbit/s only is increasingly rare

Rapidly vanishing support for older technologies– Token Ring (16 & 100Mbit/s), FDDI and ATM (25, 155 &

622Mbit/s) Ethernet switches have become widespread due to

– Their versatility Support of different bit rates and media types

– Their lower per-port cost than alternative technologies

Switched LANs

Evolving Technologies forEthernet LAN Interconnection

1980 – 1984

Shared Ethernet (CVSMA/CD) deployedInternational LAN standards developed

1985 – 1989

Bridges used for LAN interconnection to limit size of collision domains, with Spanning Tree facilitating bridge redundancy; routers used for LAN-WAN interconnection

1990 - 1994 High-speed, low-cost routers become alternatives to bridges‘Backbone’ routers developed for site interconnections

1995 - 1999 VLAN-capable switches replace bridges and LAN routers100Mbit/s Ethernet becomes common, GbE developed

2000 - Dedicated switched access and VLAN deployment become common10GbE developed, Ethernet switches become QoS-enabled

The Rise and Fallof the LAN Router

In early 1990s, small routers introduced to limit sizeof broadcast domains

– Became cheap, and fast, enough to use in LANs But routers operate at Network Layer

– Require configuration (are not plug-and-play)– Have higher per-port cost than equivalent bridge

LAN switches began to replace bridges in mid-1990s– Still operate at Layer 2– Have much lower per-port cost than routers– Can be operated in plug-and-play mode or

configured For example with management and VLAN

information Routers still required for inter-site and inter-VLAN

communication– Particularly suitable for interconnecting different

technologies For example, CSMA/CD & Frame Relay,

CSMA/CD & Token Ring

Bridges and Switches

Bridges and LAN switches– Used to interconnect LANs of same

type– Are Layer 2 devices

Operate on MAC frames

802.3,5,11, etc

Physical: to matchData Link Protocol

Bridge/LAN switch

Layer 2relay

Routers

Routers– Used for LAN–WAN and VLAN

interconnection– Layer 3 devices

Operate on packets

WAN

TokenRing LAN

CSMA/CD LAN

Hub

IP

PPP, 802.3,5,11, etc

Physical: to matchData Link Protocol

router

Layer 3relay

VLAN 2

VLAN 3

VLAN 3

Multilayer Switches

Multilayer switches have both switching and routing modules

– Operate at Layer 2 and Layer 3– Often very high-speed and rather expensive devices

Typically equipped with hardware acceleration– Used in backbone (or ‘distribution’) networks

Multilayer switch

Modern LAN Structure

Wiringcloset

Workgroupservers

Fiberlinks

10/100switch

Multilayerswitch

Site backboneGigabit Ethernet/ATM

Hub

Hub

Hub

Workgroups connected to small switches

– Workgroup servers get dedicated ports

– 10 and 100Mbit/s connections

Workgroup switches interconnected by multilayer switches

– The backbone or distribution network

– 100 and 1000Mbit/s connections used

Structure and Infrastructure

OverviewOverview

Basic Techniques

LAN Structure

Circuit-switching and Packet-switching

LAN Interconnection Services

Course Structure

Interconnection overview

LANs on same site typically linked by higher-speed LANs– For example, Ethernet LANs can be linked by higher-speed

Ethernet links Offsite connections usually provided by a service provider

– Often referred to as a public network operator (PNO)– For example, telephone company, cable TV operator, satellite

communications company Type of service provided divides into two main categories

– Dedicated inter-site capacity– Shared public network

It is rare for company to provide own inter-site links– Due mainly to installation cost

Dedicated capacity services

Can be “always on” and charged one basis of permanent availability *

– Physical inter-site point-to-point links Often referred to as leased lines Can be implemented over cables (copper, fibre), satellite,

microwave, high-speed wireless Or can be provided on demand, charged for duration of connection

– Also called a “switched service” or a “dial-up” service– Special call control (signalling) protocols used to set-up and clear

down connection– Examples are PSTN, ISDN

Essentially a Physical Layer service– Attached systems run own Link layer protocols end-to-end

across service

* Means permanently available; also referred to as “24/7”, meaning 24 hours a day, 7 days a week

PSTN = public switched telephone networkISDN = integrated services digital network

Shared capacity services

Can also be “always on”– Virtual point-to-point links over a shared public network

Referred to as virtual private networks (VPNs)– ADSL and cable modem for domestic and small offices– Connection over public frame relay *, ATM, SMDS or IP

network– Again, charged on basis of permanent connection

Or can be provided as a switched service– Again, signalling protocols used to set-up and clear down

connection– Main examples is X.25

Provided either as Layer 2 service: FR, ATM, SMDS– Or Layer 3 service: IP, X.25

ADSL = asymmetric digital subscriber lineATM = asynchronous transfer modeSMDS = switched multimegabit data serviceIP = internet protocol

* Frame Relay and ATM standards define a demand service, but it is rarely, if ever, used for site-interconnection

Dedicated vs. shared

Dedicated Layer 1 service appears to be point-to-point circuit– Could be non-switched – i.e. a real circuit– But is often a dedicated part of larger circuit belonging to the PNO

Usually provided as a feed into a public switched network– Hence the term circuit switched network

Circuit switched means– Dedicated capacity– Fixed path through network– Fixed inter-site communications delay

Shared Layer 2 or 3 services referred to as packet switched services– Always cheaper than comparative circuit switched service

Packet switched means– Shared capacity– Fixed or varying path through network– Variable inter-site communications delay

Packet switched services

Term “packet switching” has historical significance– Shared, as opposed to dedicated service– End-systems chop up (‘segment’) messages before transmission– Network interleaves packets from different users on an as-needed

basis Connection oriented or virtual circuit services use fixed network path

– Guarantee delivery sequentiality– Often include built-in safety checks– Connections can be permanently set up (PVC) or demand (SVC)– Examples include X.25, Frame Relay, ATM

Connectionless services use any available path– Do not offer delivery sequentiality– Have no built-in safety checks– Examples include IP, SMDS

Some service comparisons

Circuit switchedConnection

oriented packet switched

Connectionless packet switched

Layer 1 service only:

Layer 2 (FR, ATM)

or Layer 3 (X.25) service

Layer 2 (SMDS)or Layer 3 (IP)

service

Fixed network path means service lost if path fails

Variable network path means

tolerance to failure

Message blocks associated with fixed circuit/virtual circuit identifier

Message blocks need to carry full

end-system addresses

No buffer storage required in

network switchesBuffer storage required in switches

Network capacity set aside for circuit

Network capacity shared between service subscribers

Interconnection topologies

The topology (interconnection shape) is implicit in many network types– Growth and the need for redundancy blurs the topology somewhat

Here are some common topologies– You should be able to name them, and the basic network types to

which they usually apply

Connecting Multiple Sites:Circuit or Packet-switched Solution?

1

54

3

2

1

54

3

2

Not shared, gives predictable end-to-end delay,but needs n(n-1)/2 long-distance, leased lines.

Expensive!

Shared, so end-to-end delay is less predictable,but needs only n local access lines.

Cheaper than comparable circuit-based service.

1

54

3

2

Public network

?Circuit-based solution Packet-based solution

Hierarchical Network Architecture

Hierarchy typically local region & national– Highest level of hierarchy often referred to

as backbone Repeated at international level

– International backbones connect to national backbones

NationalNetLocal

Regional Net

Internal routers or switches only connect to other routers or switches

in same net

Edge routers or switches connect toother nets, or to subscriber access

Structure and Infrastructure

OverviewOverview

Basic Techniques

LAN Structure

Circuit-switching and Packet-

switching

LAN Interconnection Services

Course Structure

Services and Interfaces*

Public data networks offered for public use by– PNO: Public Network Operator

Often telecommunications service provider– ISP: Internet Service Provider

Typically IP only Usually standardized as interface specifications

– Specify the service(s) offered– The access protocol(s) for connecting to the service– But not the internal operation of network providing the service

Common to offer multiple services over a single infrastructure Level of Service depends protocol structure of interface specification

– Layer 1 : Physical connection only– Layer 2: Frame or cell-based– Layer 3: Packet-based

Circuit-based services

Packet-switched services

Two things tend to get standardized: (i) the service offered by the network, and (ii) how customers interface to the network

Circuit-Based Services

Dial-up digital connections ISDN

Dial-up telephone connections using modems

Fixed leased lines

1975 1985 1995 2005

ServiceNon-

Switched

Circuit-Switche

d

Leasedlines

PSTN ISDN

Leased Lines

Non-switched end-to-end digital connections provided by PNOs– No features added

Really just a connection, not strictly a service– Early offerings were analogue, requiring connection by modem– Digital services in use since early 1980s

Digital bit-rates based on 64kbit/s voice channel Higher bit rates are multiples of these

Customer access based on Plesiochronous Digital Hierarchy (PDH) Core network uses Synchronous Digital Hierarchy (SDH)

– Replacement of PDH Overcomes many problems of older PDH

– Optimized for use on optical fibre infrastructure– American equivalent called Synchronous Optical Network (SONET)

24/7 operation is expensive because– It is charged on basis of bit-rate and distance– Service provider must perform per-customer bandwidth

reservation

Multiplexing Level

NameBit Rate(Mbit/s)

0 STM-1 155.52

1 STM-4 622.08

2 STM-16 2488.32

3 STM-64 9953.28

4 STM-256 39893.12

Digital Hierarchies(Outside US – Reference Only)

Multiplexing Level

NameBit Rate(Mbit/s)

0Basic

channel64kbit/s

1 E-1 2.048

2 E-2 8.448

3 E-3 34.368

4 E-4 139.264

PDH

SDH

Service

Layer 2

Layer 3

X.25 FR Frame Relay

SMDS

Switched Multimegabi

t Data Service

ATMAsynchronou

s Transfer Mode

VPNVirtual Private

Network

ADSL

Asymmetric Digital

Subscriber Line

Packet-Switched Services

SMDS

Frame Relay

X.25

ATM

VPN

ADSL

1975 1985 1995 2005

DSL Overview

Always-on, dedicated broadband service– Operates over ‘subscriber lines’ (lines to local telephone

exchange) DSL – one term, many variations

– ADSL – Asymmetric DSL – 8 Mbps down, 640 kbps up– EDSL – Enhanced DSL – up to 1Mbps total (2-wire)– G.Lite – Slower version of ADSL that is easier to install

Also called UDSL - Universal DSL – 1.5Mbps down, 512kbps up

– HDSL – High bit-rate DSL – up to 2Mbps total (4-wire or 2-wire)– IDSL – Integrated DSL – ISDN 2B+0D access (128/144 kbps)– SDSL – Symmetric DSL – (= HDSL2)– VDSL – Very high speed DSL – 52 Mbps down, 2 Mbps up

Maximum data rates and range depend on individual installations– Quality and thickness of copper and line installation quality– Range generally 3km or more (less for VDSL)

ADSL vs. Other Modems

0 1000 2000 3000 4000 5000 6000 7000 8000 Kbit/s

14.4K

28.8K

56K

ISDN

G.lite ADSL

Full ADSL

Source: ADSL Life www.adsllife.com Maximum Speed - Actual speed will vary

Structure and Infrastructure

OverviewOverview

Basic Techniques

LAN Structure

Circuit-switching and Packet-

switching

LAN Interconnection Services

Course Structure

Structure of rest of course

5. Using sharedmedia networks

9. Routers androuting protocols

1. Networking devices:hubs, switches and

routers

3. Communicationsprotocols

6. Ethernet bridging,STP and VLANs

2. The IP Suite7. Operating at the

Network Layer

10 Queuing systems

11. Switches, routersand interconnection

networks

4. SwitchedEthernet LANs

8. Transmissionand coding

Summary

We've discussed the following practical issues– How networks evolved– Data link interfaces and protocols– Interconnection media and devices– Multiplexing– LAN structure– Circuit-switching and packet switching – Interconnection topologies– Interconnection services

Tutorial Questions

1. Approximately, what is the speed of propagation in a telephone line?

2. Give two examples which illustrate the idea of the Internet being a set of interconnected overlay networks.

3. When bits are to be transmitted onto the network medium, two of the choices that must be made are: (i) Whether to use digital or analogue communication, and (ii) How to represent 0 and 1 (in other words, the line coding technique).

a) Which type of transmission does a modem use: analogue or digital? What is the main difference between the two? How might it represent 0’s and 1’s?

b) Is it possible for a modem to transmit more than one bit at a time?

c) Which type of transmission does ISDN use? Analogue or digital? How many channels are multiplexed together on a ‘Basic Rate Interface’ and what multiplexing scheme is used: TDM or FDM?

d) Verify the arithmetic for the example of 32 x 64 kbit/s channels multiplexed onto 2.048 Mbit/s channel given on slide 34 of week 1 lecture.

Tutorial Questions(continued)

4. Give an example of a standard that defines how:a) You connect a computer to a modem?b) You connect a computer to a LAN?

5. Give an example of:a) An FDM technique used on guided media;b) A TDM technique used on unguided media.

6. List the network topologies and, for each topology, say what it is called and whether it is exclusive to a particular LAN or WAN technology or, if not, to which LAN and WAN technologies it applies.

7. Many operating systems provide some sort of terminal emulator program that turns your PC into a dumb terminal so that you can connect to a remote computer system. What is the protocol that provides this facility on an IP network?