Wilamowski B.M., Irwin J.D. The Industrial Electronics Handbook. Second Edition: Industrial Communication Systems

Подождите немного. Документ загружается.

16-12 Industrial Communication Systems

16.5 agents and Multi-agent Systems in Industry: Conclusions

 e.use.of.agent.technology.in.industrial.applications.provides.several.important.strengths,.namely.in.terms.

of.modu

larity,

.adap

tability,

.exi

bility,

.robu

stness,

.reus

ability,

.and.recon

gurability

.(cf..[7,1

6,39,41,42]).

.cont

rast

.to.conv

entional,

.cent

ralized,

.top-

down,

.rigi

.and.stat

.cont

rol

.arch

itectures,

.agen

t-based

syst

ems

.are.deve

loped

.unde

.the.fund

amental

.prin

ciples

.of.auto

nomy

.and.coop

eration,

.expl

oring

.the.dis-

ribution

.and.decen

tralization

.of.enti

ties

.and.func

tions,

.over.a.bott

om-up

.appr

oach.

. e.agen

t-based

.solu-

ion

.can.lead.to.more.simp

licity

.in.the.debu

gging,

.main

tenance,

.adap

tability,

.and.scala

bility

.of.the.syst

em.

.robu

stness

.of.the.cont

rol

.syst

.is.esse

ntially

.achi

eved

.since.this.appr

oach

.does.not.consi

der

a.cent

ral

.deci

sion

.elem

ent,

.whic

.mean

.that.the.loss.of.one.deci

sion

.comp

onent

.will.not.caus

.any.

fata

.fail

ure

.of.any.othe

.deci

sion

.comp

onent.

.In.fact

.if.prod

uction

.is.rest

ructured,

.exam

ple,

.due.to.the.

occu

rrence

.of.dist

urbances,

.the.same.nego

tiation

.proc

ess

.cont

inues

.to.be.exec

uted

.for,.in.spit

.of.the.

pres

ence

.of.di

erent

.acto

rs,

.mak

ing

.the.sys

tem

.rob

ust

.to.cha

nges.

Agent-based

.syst

ems

.are.plug

gable

.syst

ems,

.allo

wing

.chan

ges

.in.prod

uction

.faci

lities,

.as.the.addi

tion,

.remo

val,

.or.modi

cation

.of.hard

ware

.equi

pment

.as.well.as.sow

are

.modu

les,

.with

out

.the.need.to.

stop,.repr

ogram,

.and.rein

itialize

.the.syst

em.

. is.feat

ure

.is.cruc

ial

.to.supp

ort

.the.curr

ent

.requ

irements

impo

sed

.by.custo

mized

.proc

essing,

.allo

wing

.dyna

mic

.syst

.reco

ngurability

.to.face.the.vari

ability

.of.

the.dema

nd.

. e.migr

ation

.or.upda

.of.old.tech

nologies

.or.syst

ems

.by.new.ones.can.also.be.perf

ormed

in.a.smo

oth

.way.wit

hout

.the.nee

.to.stop.the.sys

tem

.(cf

.[7]

However,

.in.spit

.of.the.huge.pote

ntial

.of.agen

t-based

.solu

tions,

.its.indu

strial

.adopt

ion

.has.yet.fall

shor

.of.expe

ctations.

.Seve

ral

.reas

ons

.can.be.iden

tied

.for.this.slow.indu

strial

.adopt

ion,

.grou

ped

.in.two.

categories:.conceptual.limitations.and.technical.limitations.

.term

.of.conc

eptual

.limi

tations,

.the.agen

t-based

.para

digm

.is.a.new.way.of.thin

king

.that.requ

ires

.a.

bit.of.a.para

digm

.shi.from.the.way.manu

facturing

.syst

ems

.have.been.real

ized

.in.the.last.100.year

.In.

fact

.as.comp

anies

.have.inve

sted

.a.lot.of.time

.eor

.and.mone

.impl

ementing

.cent

ralized

.appr

oaches,

they.do.not.want.to.chan

.them

.Addi

tionally,

.one.of.the.cons

ensuses

.that.peop

.have.abou

.agen

tech

nology

.is.the.missi

.cent

ralized

.comp

onent

.for.deci

sion

.maki

ng,

.whic

.also.caus

.some.obst

acles

for.the.acc

eptance

.of.the

.con

cepts.

.terms.of.techn

ical

.limi

tations,

.inte

roperability,

.robu

stness,

.scala

bility,

.and.recon

guration

.mecha

nisms

.rema

.impo

rtant

.issu

.that.may.inhi

bit

.the.wide

.use.of.agen

t-based

.solut

ions

.by.indu

stry.

.Taki

the.scala

bility,

.for.exam

ple,

.the.curr

ent

.repo

rted

.labo

ratorial

.prot

otypes

.deal.with.doze

.or.hund

reds

of.agen

ts,

.but.indu

strial

.appl

ications

.usua

lly

.requ

ire

.syst

ems

.that.comp

rise

.thou

sands

.of.agen

ts.

.Curr

ent

plat

forms

.cann

.hand

.syst

ems

.of.that.size.with.the.robu

stness

.requ

ired

.by.indu

stry

.(cf..[7,4

2]).

.prev

iously

.iden

tied

.limi

tations

.cons

titute

.rese

arch

.oppo

rtunities

.and.chal

lenges

.from.whic

espe

cially

.the.foll

owing

.ones.can.be.poin

ted

.out:.benc

hmarking,

.inte

roperability,

.and.deve

lopment

.of.

engi

neering

.fra

meworks.

.orde

.to.proo

.the.matu

rity

.and.meri

.of.the.tech

nology

.and.to.conv

ince

.indu

stry

.abou

.its.appl

cability

.in.industrial.scenarios,.a.benchmarking.framework.is.required.that.provides.realistic.test.cases.

that.allo

.eval

uating

.agen

t-based

.solu

tions

.and.comp

aring

.them.with.trad

itional

.ones

.Some.benc

marking

.issu

.rema

.unde

ned,

.name

.the.sele

ction

.of.prop

.perf

ormance

.indi

cators,

.espe

cially

thos

.that.allo

.to.eval

uate

.qual

itative

.indi

cators,

.the.den

ition

.of.eval

uation

.crit

eria,

.the.stora

.and.

main

tenance

.of.the.bes

.pra

ctices,

.and.an.eas

.acc

ess

.to.thi

.ser

vice.

Interoperability

.is.prob

ably

.the.main.tech

nical

.prob

lem

.asso

ciated

.to.agen

.tech

nology

.that.shou

be.addr

essed.

. e.solu

tion

.for.the.inte

roperability

.ques

tion

.requ

ires

.more.rese

arch

.on.ontol

ogies

.and.

Sema

ntic

.Web.doma

ins

.and.open

.a.door.to.the.use.of.serv

ice-oriented

.syst

ems

.(cf..[39]

.combi

ning

.its.

best.fea

tures

.wit

.the.age

.tec

hnology.

.spec

ication

.of.form

al,

.stru

ctured,

.and.inte

grated

.deve

lopment

.engi

neering

.fram

eworks

.is.abso

lutely

.need

.for.the.indu

strial

.prac

tice

.in.orde

.to.supp

ort

.the.spec

ication,

.desig

.veri

cation,

.and.

impl

ementation

.of.agen

t-based

.cont

rol

.appl

ications,

.allo

wing

.an.easy

.modu

lar,

.and.rapi

.deve

lopment/

reconguration

.of.control.solutions.

Industrial Agent Technology 16-13

abbreviations

AARIA. Autonomous.Agents.at.Rock.Island.Arsenal

ADACOR

. .ADA

ptive

.hol

onic

.COn

trol

.aRc

hitecture

.for.dis

tributed

.man

ufacturing

.sys

tems

ADRENALIN

. Adv

anced

.Fr

actal

.Co

mpanies

.Us

.In

formation

.Su

pply

.Ch

ain

CMU

. Coo

perative

.ma

nufacturing

.un

ERP

. Ent

erprise

.re

source

.pl

anning

HMS

. Hol

onic

.ma

nufacturing

.sy

stems

LS/ATN

. Liv

ing

.sy

stems/adaptive

.tr

ansport

.ne

twork

MAS

. Mul

ti-agent

.sy

stem

MASCADA

. .Man

ufacturing

.Con

trol

.Sys

tems

.Cap

able

.of.Man

aging

.Pro

duction

.Cha

nge

.and.

Dis

turbances

MES

. Man

ufacturing

.ex

ecution

.sy

stem

OAS

. Ord

.ag

ent

.su

pervisor

PABADIS

. Pla

.au

tomation

.ba

sed

.on.di

stributed

.sy

stems

PABADIS’PROMISE

. .PAB

ADIS-based

. pro

duct-oriented

. man

ufacturing

. sys

tems

. for. rec

ongurable

ent

erprises

PROSA

. Pro

duct–resource–order–sta

.ar

chitecture

. Resi

dential

.ag

ent,

.re

source

.ag

ent

RAS

. Res

ource

.ag

ent

.su

pervisor

SOA

. Ser

vice

.or

iented

.ar

chitecture

VLCC

. Very.large.crude.carrier

. Vir

tual

.or

ganizations

. Wor

.or

der

UAV

. “Un

inhabited”

.or.“r

emotely

.pi

loted”

.ae

rial

.ve

hicles

references

. 1.. Wooldridge,. M.. and. Jennings,. N.. R.,. Intelligent. agents:. eory. and. practice.. e Knowledge

Engineering Review,

.10(2):115–152,.1995.

. 2.. Koes

tler,

.A.,.e G

host in the Machine,

.Ar

kana

.Pr

ess,

.Lo

ndon,

.U.K.,.1967.

. 3.. Mar

ík,

.V.,.Fle

tcher,

.M.,.and.Pěc

houček,

.M.,.Hol

ons

.and.age

nts:

.Rec

ent

.dev

elopments

.and.mutu

imp

acts,

.in.Pro

ceedings of the Ninth ECCAI-ACAI/EASSS 2001, AEMAS 2001, HoloMAS 2001,

Lec

ture Notes in Articial Intelligence,

. Vol

. 232

. Spr

inger

. Pub

lishing

. Com

pany,

. New. Yor

pp.

3–43,

.200

. 4.. Schi

llo,

.M.,.Self-o

rganization

.and.adju

stable

.aut

onomy:

.Two.sides.of.the.sam

.coin?.Con

nection

Science,

.14(4),.345–359,.2003.

. 5.. Fisc

her,

. K.,. Agen

t-based

. design. of. holo

nic

. man

ufacturing

. system

. Jou

rnal of Robotics and

Autonomous Systems,

.27,.3–13,.1999.

. 6.. Shen,.W.. and. Nor

rie,

.D.. H.,.Agen

t-based

.app

roaches

.for.inte

lligent

.man

ufacturing:

.A. sta

te-of-

the-art

.surv

ey,

.in.Pro

ceedings of DAI’98, Fourth Australian Workshop on Distributed Intelligence,

Bri

sbane,

.Au

stralia,

.Ju

.13,.1998.

. 7.. Mar

ik,

. V.. and. McFa

rlane,

. D.,. Ind

ustrial

. adopt

ion

. of. agen

t-based

. tec

hnologies,

. IEEE In

telligent

Systems,

.20(1),.27–35,.2005.

. 8.. Luder,.A.,.Klost

ermeyer,

.A.,.Pesc

hke,

.J.,.Brat

oukhine,

.A.,.and.Saut

er,

.T.,.Dist

ributed

.aut

omation:

PABAD

.ver

sus

.HMS,.IEEE Tr

ansactions on Industrial Informatics,

.1(1),.31–38,.2005.

. 9.. Par

unak,

.H..V..D.,.Wha

.can.agen

.do.in.indu

stry,

.and.why?.An.over

view

.of.indu

strially-oriented

R&D.at.CEC,.in.Klu

sch,

.M..and.Weis

.G..(Eds.),.Coo

perative Information Agents II—Learning,

Mobility and Electronic Commerce for Information Discovery on the Internet,

. Lec

ture Notes in

Computers Science,

.Vo

.1435,.Sp

ringer,

.Pa

ris,

.Fra

nce,

.pp

.1–18,.1998.

16-14 Industrial Communication Systems

. 10.. Karim,. S.. and. Heinze,.C.,. Experiences. with. the.design. and. implementation. of. an. agent-based.

aut

onomous

.UAV.cont

roller,

.in.Pěc

houček,

.M.,.Stein

er,

.D.,.and. om

pson,

.S..G..(Eds.),.Pro

ceedings

of Fourth International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS

2005),

.Utr

echt,

.the.Neth

erlands,

.Jul

.25–29,.2005—Spe

cial

.Track.for.Ind

ustrial

.App

lications,

.ACM.

Pres

.Ne

.Yo

rk,

.2005.

. 11.. Baxter

.J..W..and.Hor

.G..S.,.Con

trolling

.tea

.of.uninha

bited

.air.vehic

les,

.in.Pěc

houček,

.M.,.

Stein

er,

.D.,.and. om

pson,

.S..G..(Eds.),.Pro

ceedings of Fourth International Joint Conference on

Autonomous Agents and Multiagent Systems (AAMAS 2005),

.Utr

echt,

.the.Neth

erlands,

.Jul

.25–29,.

2005—Spe

cial

.Trac

.fo

.In

dustrial

.Ap

plications,

.ACM.Pr

ess,

.Ne

.Yo

rk,

.2005.

. 12.. Par

unak,

.H..V..D.,.Brue

ckner,

.S.,.and.Saut

er,

.J.,.Digit

.pher

omone

.mec

hanisms

.for.coor

dination

.of.

unmann

.vehic

les,

.in.Pro

ceedings of First International Joint Conference on Autonomous Agents and

MultiAgent Systems (AAMAS): Part 1,

.Bo

logna,

.It

aly,

.Ju

.15–19,.2002.

. 13.. Mar

son,

. P.-E.,. Sou

lignac,

. M.,. and. Tai

llibert,

. P.,. AERIAL:. Hypo

thetical

. traj

ectory

. pla

nning

. for.

mul

ti-UAVs

.coor

dination

.and.cont

rol,

.in.Pro

ceedings of Seventh International Joint Conference on

Autonomous Agents And MultiAgent Systems: Demo Papers,

.Es

toril,

.Po

rtugal,

.Ma

.12–16,.2008.

. 14.. Cam

arinha-Matos,

.L..M..and.Afsa

rmanesh,

.H.,.Virt

ual

.enter

prise

.mode

ling

.and.sup

port

.infras

truc-

tures:

.App

lying

.mul

ti-agent

.system.app

roaches,

.in.Pro

ceedings of the Ninth ECCAI Advanced Course

ACAI 2001 and Agent Link’s ird European Agent Systems Summer School, EASSS 2001,

.Lec

ture

Notes in Articial Intelligence,

.Vol

.2086,.Prague

.Czech.Repu

blic,

.Jul

.2–13,.2001,.Sele

cted

.Tut

orial

Papers;.Springer.Publishing.Company,.New.York,.2001.

. 15.. Pěc

houček,

.M.,.Šiš

lák,

.D.,.Pav

líček,

.D.,.and.Uller

.M.,.Aut

onomous

.agen

.for.air-t

rac

.deconic-

ion,

.in.Pro

ceedings of Fih International Joint Conference on Autonomous Agents and MultiAgent

Systems (industry track),

.Ha

kodate,

.Ja

pan,

.Ma

.2006.

. 16.. Bele

cheanu,

.S..M.,.Mich

ael,

.L.,.and.Terr

.P.,.Comm

ercial

.app

lications

.of.agen

ts:

.Less

ons,

.experi-

ces

.and.cha

llenges,

.in.Pro

ceedings of Fih International Joint Conference on Autonomous Agents

and MultiAgent Systems (industry track),

.Ha

kodate,

.Ja

pan,

.Ma

.2006.

. 17.. Scerr

.P.,.Von.Gon

ten,

.T.,.Fudge

.G.,.Owen

.S.,.and.Syc

ara,

.K.,.Tran

sitioning

.mul

tiagent

.tec

hnology

to.UAV.app

lications,

.in.Pro

ceedings of Seventh International Conference on Autonomous Agents and

Multiagent Systems (AAMAS)—Industry and Applications Track,

.Est

oril,

.Por

tugal,

.May.12–16,.2008.

. 18.. Reich

el,

.K.,.Hoc

hgeschwender,

.N.,.and.Voos,.H.,. OpCog:. An. indu

strial

.deve

lopment

.app

roach

for.cognit

ive

.agen

.system

.in.milit

ary

.UAV.app

lications,

.in.Pro

ceedings of Seventh International

Conference on Autonomous Agents and Multiagent Systems (AAMAS)—Industry and Applications

Track,

.Es

toril,

.Po

rtugal,

.Ma

.12–16,.2008.

. 19.. Him

o,

.J.,.Sko

belev,

.P.,.and.Woo

ldridge,

.M.,.MAG

ENTA

.tec

hnology:

.Mul

ti-agent

.sys

tems

.for.ind

ustrial

log

istics,

.in.Pěc

houček,

.M.,.Ste

iner,

.D.,.and. om

pson,

.S..G..(Ed

s.),

.Pro

ceedings of Fourth International

Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS 2005),

.Utr

echt,

.the.Net

herlands,

Jul

.25–

29,

.200

5—Special

.Tra

.for.Ind

ustrial

.App

lications,

.ACM.Pre

ss,

.New.Yor

.200

. 20.. Himo

,

.J.,.Rzevsk

.G.,.and.Sko

belev,

.P.,.MAGENTA.tec

hnology:

.Mul

ti-agent

.logis

tics

.i-sch

eduler

for.road.tran

sportation,

.in.Pro

ceedings Fih International Joint Conference on Autonomous Agents

and MultiAgent Systems (industry track),

.Ha

kodate,

.Ja

pan,

.2006.

. 21.. Sko

belev,

.P.,.Gla

schenko,

.A.,.Grach

ev,

.I.,.and.Ino

zemtsev,

.S.,.MAGENTA.tec

hnology

.cas

.studies.of.

mag

enta

.i-sch

eduler

.for.road.tran

sportation,

.in.Pro

ceedings of Sixth International Joint Conference

on Autonomous Agents and MultiAgent Systems,

.Ho

nolulu,

.HI,.2007.

. 22.. Obi

tko,

.M.,.Vrba,.P.,.Mař

ík,

.V.,.and.Rada

kovič,

.M.,.Sema

ntics

.in.indu

strial

.dist

ributed

.system

.in.

Pro

ceedings of the 17th World Congress,

. e.Int

ernational

.Federa

tion

.of.Aut

omatic

.Con

trol,

.Seo

ul,

Kor

ea,

.Ju

.6–11,.2008.

. 23.. Vya

tkin,

.V.,.Chr

istensen,

.J.,.Las

tra,

.J..L..M.,.and.Aui

nger,

.F.,.OOO

NEIDA:

.An.ope

.obj

ect-oriented

kno

wledge

.eco

nomy

.for.int

elligent

.ind

ustrial

.aut

omation,

.IEE

E Transactions on Industrial Informatics,

1(1

.4–1

.200

Industrial Agent Technology 16-15

. 24.. Deen,.S..M.,.Agent Based Manufacturing—Advances in the Holonic Approach: Advanced Information

Processing,

.Sp

ringer

.Ver

lag,

.Ber

lin,

.Ger

many,

.2003.

. 25.. Klo

stermeyer,

.A.,.Age

ntengestützte Navigation wandlungsfähiger Produktionssysteme,

.Dis

sertation,

.Fac

ulty

of.Mec

hanical

.Eng

ineering,

.Ott

o-v.-Guericke

.Uni

versity,

.Mag

deburg,

.Ger

many,

.ISB

.3-0

0-011618-4.

. 26.. Heik

kila,

.T.,.Kol

lingbaum,

.M.,.Valc

kenaers,

.P.,.and.Bluemin

.G..J.,.manA

ge:

.An.agen

.arc

hitec-

ture

.for.man

ufacturing

.cont

rol,

.in.Pro

ceedings of the Second International Workshop on Intelligent

Manufacturing Systems,

.Leu

ven,

.Be

lgium,

.Sep

tember

.1999.

. 27.. Bel

lifemine,

.F.,.Cair

.G.,.and.Gre

enwood,

.D.,.Dev

eloping Multi-Agent Systems with JADE,

.Wile

Chich

ester,

.U.K.,.2007.

. 28.. Vrba,.P.,.JAV

A-based

.agen

.pla

tform

.eval

uation,

.in.Mar

ík,

.V..et.al..(Eds.),.Mul

ti-Agent Systems and

Applications III, LNAI 2691,

.Sp

ringer-Verlag,

.Ber

lin/Heidelberg,

.Ger

many,

.pp

.47–58,.2003.

. 29.. Foun

dation

.fo

.In

telligent

.Ph

ysical

.Ag

ents

.(FIPA),.2009,.ht

tp://www.pa.org.

. 30.. Gru

ber,

.T..R.,.A.tran

slation

.app

roach

.to.por

table

.ont

ology

.spe

cications,

.Kno

wledge Acquisition,

5(2),.199–220,.Jun

.1993.

. 31.. Hak

en,

. H.,. Inf

ormation and Self-Organization: A Macroscopic Approach to Complex Systems,

Spr

inger-Verlag,

.Ber

lin,

.Ger

many,

.1988.

. 32.. Klir,.G.,.Fa

cets of System Science,

.Plen

.Pr

ess,

.Ne

.Yo

rk,

.1991.

. 33.. De. Wolf

. T.. and.Hol

voet,

. T.,. Emerg

ence

. and. self-o

rganisation:

. A. sta

tement

. of. simila

rities

. and.

dieren

ces,

. in. Wor

kshop on Engineering Self-Organizing Applications (ESOA); as Part of ird

International Joint Conference on Autonomous Agents and MultiAgent Systems (AAMAS),

.New.York,.

Jul

.21–24,.2004.

. 34.. Valc

kenaers,

. P.. et. al.,.On. the. design. of. emerg

ent

. system

. An. inves

tigation

. of. integ

ration

. and.

inter

operability

.is

sues,

.En

gineering Applications of Articial Intelligence,

.16,.377–393,.2003.

. 35.. Ulieru

.M..and.Unl

and,

.R.,.Enab

ling

.tec

hnologies

.for.the.crea

tion

.and.rest

ructuring

.pro

cess

.of.

emerg

ent

.enter

prise

.alli

ances,

.Int

ernational Journal of Information Technology and Decision Making,

3(1),.33–60,.Ma

rch

.2004.

. 36.. Bra

tman,

.M..E.,.Int

entions, Plans, and Practical Reason,

.Har

vard

.Uni

versity

.Pre

ss,

.Cam

bridge,

.MA,.198

. 37.. Mat

urana,

. F.. and. Nor

rie,

. D.,. Mul

ti-agent

. medi

ator

. arc

hitecture

. for. dist

ributed

. man

ufacturing,

Jou

rnal of Intelligent Manufacturing,

.7,.257–270,.1996.

. 38.. Leit

ão,

.P..and.Rest

ivo,

.F.,.ADAC

OR:

.A.holo

nic

.arc

hitecture

.for.agi

.and.adap

tive

.man

ufacturing

cont

rol,

.Co

mputers in Industry,

.57(2),.121–130,.Fe

bruary

.2006.

. 39.. Leit

ão,

.P.,.Agen

t-based

.dist

ributed

.man

ufacturing

.cont

rol:

.A.sta

te-of-the-art

.surv

ey,

.Eng

ineering

Applications of Articial Intelligence,

.22(7),.979–991,.2009.

. 40.. Nea

gu,

.N.,.Dor

er,

.K.,.Gre

enwood,

.D.,.and.Cal

isti,

.M.,.LS/

ATN:

.Rep

orting

.on.a.suc

cessful

.age

nt-based

sol

ution

.for.tra

nsport

.log

istics

.opt

imization,

.in.Pro

ceedings of the IEEE Workshop on Distributed Intelligent

Systems: Collective Intelligence and Its Applications (DIS’06),

.Was

hington,

.DC,.Jun

.15–

16,

.200

. 41.. Shen,.W.,.Wan

.L.,.and.Hao,.Q.,.Agen

t-based

.dist

ributed

.man

ufacturing

.pro

cess

.pla

nning

.and.

sch

eduling:

.A.sta

te-of-the-art

.surv

ey,

.IEEE Tr

ansactions on Systems, Man, and Cybernetics—Part C:

Applications And Reviews,

.36(4),.Ju

.2006.

. 42.. Pěc

houček,

. M.. and. Mar

ík,

. V.,. Ind

ustrial

. deploy

ment

. of. mul

ti-agent

. tec

hnologies:

. Rev

iew

. and.

sele

cted

.cas

.studies,.Aut

onomous Agents and Multi-Agent Systems,

.17(3),.397–431,.Decem

ber

.2008.

17-1

17.1 Introduction on real-time Systems

17.1.1 Denition

A.real-time.system.is.a.system.in.which.the.correctness.of.the.results.does.not.only.depend.on.the.logical.

corr

ectness

.of.the.comp

utations/operations

.perf

ormed

.to.obta

.them.but.also.on.the.time.the.resu

lts

are.obta

ined

.[Kop9

7,Sta88].

.Logi

cally

.corr

ect,

.but.late.resu

lts

.can.be.eith

.not.usef

.or.even.harm

ful

.to.

the.syst

em,

.depe

nding

.on.the.natu

.of.the.appl

ication

.consi

dered.

.Exam

ples

.of.real

-time

.appl

ications

are.foun

.in.indu

strial

.plan

.cont

rols,

.autom

otive,

.avio

nics,

.indu

strial

.autom

ation,

.robo

tics,

.moni

tor-

ing

.syst

ems,

.mult

imedia

.syst

ems,

.tele

communications,

.inte

ractive

.syst

ems,

.and.cons

umer

.elec

tronics.

Unli

.conv

entional,

.non-

real-time

.comp

uter

.syst

ems,

.real

-time

.comp

uter

.syst

ems

.are.clos

ely

.coup

led

with.the.phy

sical

.pro

cess

.or.the.env

ironment

.bei

.mon

itored

.and.con

trolled.

.real

-time system

.(Fig

ure

.17.1

)

.consi

sts

.of.a.real

-time

.comp

uter

.syst

em,

.a.cont

rolled

.objec

.and.may.

also.feat

ure

.an.oper

ator.

. e.real

-time

.comp

uter

.syst

.must.reac

.to.stim

uli

.from.the.cont

rolled

.objec

(or.the.oper

ator)

.with

.a.time.inte

rval

.spec

ied

.by.the.dead

line.

. e.conc

ept

.of.dead

lines

.must.not.be.

conf

used

.with.fast

.or.high

-performance

.comp

uting.

.Real

-time

.comp

uting

.is.die

rent

.since.the.objec

tive

.of.fast.comp

uting

.is.to.mini

mize

.the.aver

age

.resp

onse

.time.of.a.task

.whil

.real

-time

.comp

uting

.is.

conc

erned

.abou

.the.maxi

mum

.or.wors

t-case

.resp

onse

.time.and.the.die

rence

.betw

een

.mini

mum

.and.

maxi

mum

.res

ponse

.tim

.the.so-

called

.jitt

.[Sta

88].

.most.impo

rtant

.prop

erty

.of.a.real

-time

.syst

.is.not.high.spee

.but.pred

ictability

.[Sta

90],

.whic

mean

.the.abil

ity

.to.dete

rmine

.in.adva

nce

.whethe

.the.task.can.be.comp

leted

.with

.its.dead

lines

orn

ot.

Real-Time Systems

17.1. Introduction.on.Real-Time.Systems............................................. 17-1

Denition. •. Real-Time.Constraints.Characterization. •. Typical.

Application.Domains. •. Real-Time.Scheduling.and.Relevant.Metrics

17.2. Real-Time.Communication...........................................................17-4

Deterministic.vs..Statistical.Communication. •. Best.Eort.vs..

Guaranteed.Service. •. Performance.Metrics. •. Analytical.Methods.

to.Assess.Performance.of.Real-Time.Networks

17.3. Design.Paradigms.for.Real-Time.Systems................................... 17-7

Centralized.vs..Distributed.Architectures. •. Composability.

and.Scalability. •. Time-Triggered.vs..Event-Triggered.

Systems. •. Comparison.of.the.Real-Time.Support.Provided.

byNotable.Approaches

17.4. Design.Challenges.in.Real-Time.Industrial.

Comm

unication

.Sys

tems................................................................ 17-9

Real-Time.and.Security. •. Real-Time.and.Flexibility. •. Oering.

Real-Time.Support.to.Wireless.Communication

References....................................................................................................17-10

Lucia Lo Bello

University of Catania

José Alberto Fonseca

Universidade of Aveiro

Wilfried Elmenreich

University of Klagenfurt

17-2 Industrial Communication Systems

17.1.2 real-time Constraints Characterization

A.criterion.commonly.used.to.classify.the.timing.constraints.of.real-time.tasks.is.the.usefulness.of.their.

results.(value).as.functions.of.the.tasks’.completion.time.and,.in.particular,.of.their.lateness.. e.lateness.

of.a.task.is.dened.as.the.dierence.between.the.task’s.actual.completion.time.and.its.deadline..If.the.

task.is.on.time,.its.lateness.will.be.zero.or.negative..If.the.task.is.late,.a.positive.lateness.will.be.found..

To.what.extent.a.deadline.miss.will.compromise.the.system.depends.on.the.real-time.nature.of.the.

deadlines.imposed.on.the.system..When.missing.a.deadline.can.have.catastrophic.consequences.to.the.

system,.the.deadline.is.called.hard..If.a.deadline.miss.only.entails.a.performance.degradation,.but.does.

not.jeopardize.the.correct.system.behavior,.the.deadline.is.called.so.

According

.to.this.criterion,.three.types.of.tasks.can.be.identied,.i.e.,.hard,.so,.and.rm.real-time..All.

these.kinds.of.tasks.give.a.positive.value.to.the.system.if.they.complete.on.time.. e.dierence.between.the.

three.categories.becomes.signicant.when.the.tasks.are.late..If.the.late.task.is.hard.real-time,.the.value.of.

the.result.produced.by.the.task.aer.its.deadline.is.negative,.thus.indicating.that.a.damage.on.the.system.

occurs..As.a.result,.hard.real-time.tasks.have.to.meet.their.deadlines.under.any.circumstances..Conversely,.

the.value.of.the.result.produced.by.a.so.real-time.task.starts.decreasing.as.the.lateness.of.the.task.becomes.

positive.and.will.eventually.become.null.at.a.given.point.of.time.aer.the.deadline..So,.the.late.completion.

of.a.so.real-time.task.can.be.occasionally.tolerated,.and.it.is.advisable.to.minimize.the.lateness.of.the.task,.

as.the.value.of.a.late.task.decreases.with.time..Finally,.the.value.of.the.result.produced.by.a.rm.real-time.

task.drops.abruptly.to.zero.as.soon.as.the.lateness.becomes.zero.and.remains.null.with.increasing.lateness.

values..As.it.brings.no.value.to.the.system,.a.rm.real-time.task.should.be.either.completed.on.time.or.

dropped.if.it.is.late,.as.there.is.no.point.to.continue.its.execution.aer.the.deadline.

.the.case.of.hard.deadlines,.suitable.validation.techniques.are.required.to.demonstrate.that.the.sys-

tem

.adheres.to.the.intended.timing.behavior..If.the.deadlines.are.so,.in.general.no.validation.is.required.

and.it.is.sucient.to.prove.that.it.is.possible.to.meet.a.timing.constraint.specied.in.terms.of.statistical.

averages.(e.g.,.that.the.average.deadline.miss.rate.is.below.a.given.application-dependent.threshold).

Real-time

computer system

Operator

Controlled object

Actuators

Sensors

Input devices

Display

Man–machine

interface

Instrumentation

interface

Continuous

phenomenon

Non-deterministic

behavior

Deterministic

behavior in value

and time domain

Real-time system with user interface

Real-time control system

FIGURE 17.1 Elements.of.a.real-time.system.

Real-Time Systems 17-3

17.1.3 typical application Domains

Real-time. systems. can. be. found. in. many. dierent.application. domains.. However,. unlike. personal.

com

puters

.(PC

.and.wor

kstations

.whe

.use

.are.ful

.awa

.of.(no

n-real-time)

.app

lications

.suc

.as.

E-m

ail,

.Int

ernet

.bro

wser,

.and.tex

.pro

cessing,

.rea

l-time

.app

lications

.are.oe

.imp

lemented

.as.emb

ed-

ded systems

.pro

viding

.the

.ser

vice

.hid

den

.fro

.our.vie

.[El

m09].

. e.typ

ical

.app

lication

.dom

ains

.for.

rea

l-time

.sy

stems

.ar

•

. Dig

ital

.con

trol

.sys

tems

.imp

lementing

.a.fee

dback

.loo

.fro

.one.or.mor

.sen

sor

.rea

dings

.to.one.

or.mor

.act

uators.

. us

.the.con

trolled

.par

ameter

.is.kep

.at.a.des

ired

.val

.eve

.whe

.con

di-

tions

.cha

nge

.in.the.sys

tem.

.Sin

.suc

.con

trol

.loo

.are.typ

ically

.ver

.sen

sitive

.to.jit

ter

.and.

del

ay,

.a.dig

ital

.con

trol

.sys

tem

.has.to.be.imp

lemented

.in.a.way.tha

.the.del

.bet

ween

.mea

surement

and.actuating.is.small.in.comparison.to.the.timing.parameters.of.the.controlled.system..Jitter.

is.esp

ecially

.cri

tical

.in.suc

.app

lications.

.Dig

ital

.con

trol

.sys

tems

.are.har

.rea

l-time

.sys

tems.

Vio

lation

.of.the.tim

ing

.req

uirements

.can.res

ult

.in.ins

tabilities

.of.the.con

trol

.loo

.and

.the

re-

fore,

.oe

.in.dam

age

.of.phy

sical

.sys

tems.

.An.exa

mple

.is.the.ign

ition

.con

trol

.sys

tem

.in.ac

ar’s

com

bustion

.eng

ine.

. e.req

uirement

.of.thi

.app

lication

.is.bas

ically

.to.mak

.the.fue

l–air

.mix

ture

.ign

ite

.in.the.cyl

inder

.so.tha

.the.pis

ton

.is.acc

elerated

.on.its.way.dow

.Now.ima

gine

.ab

adly

des

igned

.con

trol

.sys

tem

.tha

.may.eve

ntually

.del

.the.ign

ition

.ins

tant

.by.a.few.mil

liseconds—the

fue

l–air

.mix

ture

.is.now.ign

ited

.whi

.the.pis

ton

.is.in.the.wro

.pos

ition.

. is.sma

.tim

.del

cha

nges

.the.for

ces

.to.pis

ton,

.con

nection

.rod

.and.cra

nksha

.in.a.way.tha

.may.per

manently

dam

age

.the.eng

ine.

•

. Man

–machine

.bas

.rea

l-time

.sys

tems.

. e.rea

l-time

.con

trol

.sys

tem

.inc

orporates

.the.ins

trumen-

tation

.int

erface

.con

sisting

.of.inp

.dev

ices

.and.out

put

.dev

ices

.tha

.int

erface

.to.a.hum

.ope

rator.

Acr

oss

.thi

.int

erface,

.the.rea

l-time

.com

puter

.sys

tem

.is.con

nected

.to.the.pro

cess

.env

ironment

.via.

sen

sors

.and.act

uators

.tha

.tra

nsform

.the.phy

sical

.sign

als

.of.the.con

trolled

.obje

.into.a.pro

cess-

able

.fo

.an

.vi

.ve

rsa.

•

. Hig

h-level

.con

trols

.inv

olve

.ig

.man

agement

.sys

tems,

.fac

tory

.auto

mation,

.rob

.con

trol,

.etc

 e.rea

l-time

.tas

.req

uire,

.for.exa

mple,

.pat

.and.tra

jectory

.pla

nning.

. e.rea

l-time

.con

trol

.sys

tem

.is.req

uired

.to.pro

vide

.its.res

ults

.wit

hin

.a.giv

.max

imum

.res

ponse

.tim

.Man

.of.the

app

lications,

.e.g

.in.tra

nsportation

.or.pla

.auto

mation,

.are.cri

tical

.req

uiring

.a.hig

hly

.dep

end-

able

.implementation.of.the.real-time.control.system.

•

. Man

. sign

. pro

cessing

. app

lications

. com

. wit

. rea

l-time

. req

uirements,

. nee

ding

. to. pro

vide

thr

oughput

.at.a.giv

.sam

pling

.rat

.Voi

.pro

cessing

.app

lications

.for.tel

ephony

.fur

ther

.req

uire

tha

.the.pro

cessing

.del

.doe

.not.exc

eed

.a.giv

.max

imum.

. is.kin

.of.app

lication

.oe

.has.

so.rea

l-time

.req

uirements,

.sinc

.a.vio

lation

.of.the.tim

ing

.mer

ely

.dim

inishes

.the.qua

lity

.of.the.

ser

vice.

.Oth

.so.rea

l-time

.app

lications

.inc

lude

.vid

.con

ference

.app

lications,

.onl

ine

.gam

ing,

and

.to.so

.ex

tent,

.in

stant

.me

ssaging.

17.1.4 real-time Scheduling and relevant Metrics

A.schedule.is.a.mapping.of.task.executions.on.a.resource..For.example,.if.the.tasks.are.processes,.the.

res

ource

.wil

.be.the.pro

cessor,

.whi

.if.the.tas

.to.be.per

formed

.are.dat

.pac

ket

.tra

nsmissions,

.the.

res

ource

.wil

.be.the.net

work

.ban

dwidth.

.A.sch

edule

.is.fea

sible

.if.all.the.tas

.com

plete

.the

.exe

cution

und

.a.set.of.spe

cied

.con

straints.

.A.tas

.set.is.sch

edulable

.if.the

.exi

sts

.a.fea

sible

.sch

edule

.for.it..In.

the.fol

lowing,

.lik

.in.[Li

u00],

.we.de

.a.tas

.as.a.seq

uence

.of.job

.tha

.join

tly

.pro

vide

.a.sys

tem

.fun

ction,

whi

.a.job.is.the.uni

.of.wor

.tha

.is.sch

eduled

.and.exe

cuted

.by.the.sys

tem.

. is.not

ation

.all

ows

.us.to.

dea

.wit

.di

erent

.sys

tem

.act

ivities,

.suc

.as.the.exe

cution

.of.a.pro

cess

.on.a.pro

cessor

.or.the.tra

nsmis-

sion

.of.a.da

.pa

cket

.on.a.ne

twork,

.in.a.un

iform

.wa

17-4 Industrial Communication Systems

While.in.conventional.non-real-time.systems,.the.commonly.used.metrics.are.throughput,.fairness,.

and.aver

age

.resp

onse

.time

.in.real

-time

.sche

duling

.all.the.targ

.metr

ics

.are.rele

vant

.to.the.syst

time

liness.

.Typ

ical

.met

rics

.are.the

refore

•

. Resp

onse

.time

.den

.as.the.die

rence

.betw

een

.the.comp

letion

.time.and.the.rele

ase

.time.of.

a.job

•

. Abso

lute

.jitt

er,

.den

.as.the.die

rence

.betw

een

.the.mini

mum

.and.the.maxi

mum

.resp

onse

.time.

for.a.job

•

. Maxi

mum

.late

ness

.for.so.real

-time

.task

.den

.as.the.maxi

mum

.die

rence

.betw

een

.the.

nis

hing

.tim

.of.a.job.and.its.dea

dline

•

. Deadline.miss.ratio.for.so.real-time.tasks,.i.e.,.the.ratio.between.the.number.of.jobs.that.missed.

thei

.dea

dlines

.and.the.ove

rall

.numb

.of.so.rea

l-time

.jobs.rel

eased

•

. rou

ghput

.on.tim

.de

ned

.as.the.amo

unt

.of.jobs.com

pleted

.by.the.dea

dline

.per.tim

.uni

•

. Cumu

lative

.valu

.for.a.set.of.so.or.rm.real

-time

.task

.den

.as.the.sum.of.all.the.valu

calc

ulated

.at.each.job.comp

letion

.time

.give

.thei

.valu

.func

tions

Since

.the.semi

nal

.pape

.of.Liu.and.Layl

and

.[Liu

73],

.a.very.larg

.amou

.of.work

.deal

.with.real

-time

sche

duling,

.addr

essing

.it.in.mult

iple

.scen

arios

.and.from.die

rent

.pers

pectives.

.Inte

rested

.read

ers

.may.

refe

.to.[Li

u00]

.and.[Sh

a04].

17.2 real-time Communication

17.2.1 Deterministic vs. Statistical Communication

In.hard.real-time.systems,.deadline.miss.should.never.happen,.as.deadline.misses.severely.aect.the.

appl

ication

.beha

vior

.and.dete

rmine

.non-

recoverable

.erro

rs.

.In.so.real

-time

.syst

ems,

.occa

sional

.viol

tions

.of.dead

lines

.can.be.toler

ated,

.alth

ough

.at.the.expe

nse

.of.a.degr

adation

.of.the.qual

ity of service

(QoS

)

.prov

ided

.to.the.appl

ication.

.Oer

ing

.real

-time

.supp

ort

.on.netw

orks

.mean

.that.a.pred

ictable

time.beha

vior

.of.comm

unications

.can.be.guar

anteed,

.eith

.in.a.dete

rministic

.or.in.a.stoch

astic

.way..

Pred

ictable

.time.beha

vior

.is.not.nece

ssarily

.a.syno

nym

.of.cons

tant

.time.beha

vior.

.What.real

.coun

.is.

the.exis

tence

.of.an.uppe

.boun

.on.some.rele

vant

.QoS.inde

.such.as.end-

to-end

.pack

.deli

very

.time

pack

.chan

nel

.acce

.time

.roun

dtrip

.time

.etc.. e.exis

tence

.of.such.an.uppe

.boun

.make

.it.possi

ble

to.asse

.whet

her

.the.pack

.dead

lines

.will.be.met.or.not,.in.eith

.a.dete

rministic

.or.a.stoch

astic

.way..

It.is.the

refore

.pos

sible

.to.de

•

. Dete

rministic

.real

-time

.chan

nel

.as.one.that.prov

ides

.a.prio

.dete

rministic

.guar

antees

.for.the.

time

.del

ivery

.of.pac

kets

.bet

ween

.two.end

-points

•

. Stati

stical

.real

-time

.chan

nel

.as.one.that.guar

antees

.the.time

.deli

very

.of.pack

ets

.betw

een

.two.

end.poin

.in.stat

istical

.term

.i.e.,.that.the.prob

ability

.that.a.pack

.miss

.its.dead

line

.is.less.than.

a.cer

tain

.los

.tole

rance

.Z,.i.e

.Pr(pa

cket delay

.>.del

ay bound)

.≤.Z

Stat

istical

.guar

antees

.on.dead

line

.meet

ing

.sati

sfy

.the.requ

irements

.of.real

-time

.non-

mission-critical

appl

ications,

.such.as.peri

odic

.cont

rol

.or.indu

strial

.mult

imedia,

.whic

.can.toler

ate

.a.smal

.viol

ation

prob

ability

.with.dela

.boun

ds.

.In.an.autom

ated

.manu

facturing

.syst

em,

.for.exam

ple,

.real

-time

.peri

odic

control.messages.may.need.to.be.delivered.within.20.ms.of.their.generation.with.a.98%.probability,.

whil

.voic

.pack

ets

.may.requ

ire

.deli

very

.with

.40.ms.with.a.94%.prob

ability,

.and.so.on..Dete

rministic

and.stat

istical

.anal

ysis

.tech

niques

.for.wire

.indu

strial

.netw

orks

.have.been.thor

oughly

.inve

stigated

.in.

the.lite

rature.

.For.inst

ance,

.in.[Kwe

03]

.Shin.et.al..anal

ytically

.demo

nstrate

.that

.in.orde

.to.stat

istically

boun

.the.medi

.acce

.time.for.a.real

-time

.pack

.tran

smitted

.over.Ether

net

.netw

orks

.in.the.pres

ence

.of.both.real

-time

.(RT).and.non-

real-time

.(NRT

)

.tra

.it.is.suc

ient

.to.keep.the.total.arri

val

.rate.

for.new.pack

ets

.gene

rated

.by.stat

ions

.belo

.a.thre

shold

.call

.the.netw

ork-wide input limit.

.To.enfo

rce

such.a.limit.on.a.per-station.basis,.each.station.is.assigned.a.local.threshold,.called.a.station input limit,.

Real-Time Systems 17-5

and.a.middleware.called.a.traffic smoother.is.entrusted.with.the.regulation.of.the.outgoing.non-

real-time

.str

eam

.on.eac

.nod

.Sev

eral

.tra

c

.smo

others

.hav

.bee

.pro

posed

.in.the.lit

erature,

.for.eit

her

sha

red

.or.swi

tched

.Ethe

rnet,

.whi

.di

.in.the.way.the.sta

tion

.inp

.lim

.is.enf

orced.

.Mor

.det

ails

can.be.fou

.in.[Kw

e04]

.and.[Lo

B05].

.Amo

.rec

ent

.app

roaches

.to.obt

ain

.rea

l-time

.per

formance

.ove

an.Ethe

rnet,

.the

.are.the.FTT

-Ethernet

.[Pe

d05],

.the.PEA

.pro

tocol

.[Bo

n03],

.and.the.Tim

e-Triggered

Ethe

rnet

.(T

TE)

.[K

op05].

17.2.2 Best Effort vs. Guaranteed Service

Guaranteed service.means.that.the.user.is.given.a.guarantee.on.the.system.timing.behavior.(for.instance,.

on.a.mes

sage

.end

-to-end

.del

.and

/or

.jit

ter,

.med

ium

.acc

ess

.tim

.etc

.and.tha

.the.mee

ting

.of.the.

rel

evant

.tim

ing

.con

straints

.has.to.be.val

idated,

.thr

ough

.for

mal

.met

hods

.or.exh

austive

.simu

lation

.and.

tes

ting.

. is.ser

vice

.is.req

uired

.for.har

.rea

l-time

.tra

c.

.On.the.con

trary,

.bes

t-eort service

.mea

.tha

alt

hough

.the.use

.req

uires

.the.bes

.QoS.the.sys

tem

.can.pro

vide,

.the.sys

tem

.is.all

owed

.to.del

iver

.a.low

qua

lity

.th

.ex

pected.

. i

.se

rvice

.is.su

itable

.fo

.so



.re

al-time

.tr

ac.

17.2.3 Performance Metrics

In.real-time.communication,.the.primary.performance.metrics.are.related.to.the.timeliness.of.data.exchange.

ove

.the.net

work.

.As.rea

l-time

.ow

.hav

.to.be.pro

vided

.wit

.a.di

erent

.QoS.tha

.non

-real-time

.one

.whe

di

erent

.typ

.of.tra

c

.hav

.to.be.tra

nsmitted

.ove

.the.sam

.cha

nnel,

.tra

c

.pri

oritization

.is.req

uired.

.In.

par

ticular,

.end

-to-end

.rea

l-time

.per

formance

.of.rea

l-time

.tra

c

.is.ver

.imp

ortant,

.as.usu

ally

.net

worked

rea

l-time

.sys

tems

.han

dle

.mon

itoring

.and.con

trol

.app

lications,

.whic

.typ

ically

.req

uire

.a.res

ponse

.wit

hin

bou

nded delay,

.oe

.com

bined

.wit

.low

jitter

.val

ues.

Typically,

.the.end-t

o-end

.del

.is.com

posed

.of.sev

eral

.sta

ges

.of.dat

.pro

cessing

.and.tra

nsmission.

To.be.abl

.to.prov

ide

.a.boun

ded

.del

ay,

.the.del

.of.eac

.sta

.mus

.be.boun

ded.

.For.exa

mple,

.the.que

ing

.del

.in.the.net

work,

.whi

.in.tur

.dep

ends

.on.the.que

uing

.del

.in.the.loc

.que

ues

.and.on.the.

num

ber

.of.hop

.in.the.pat

.fro

.the.sour

.to.the.des

tination

.nod

.of.the.pac

ket

.in.pac

ket

.swi

tched

net

works.

.In.eac

.net

work

.dev

ice

.trav

ersed

.by.the.pac

ket,

.mul

tiple

.que

ues

.wit

.di

erent

.pri

orities

are.nee

ded

.to.han

dle

.di

erent

.tra

c

.cla

sses

.and.rea

l-time

.sch

eduling

.alg

orithms,

.suc

.as.wei

ghted

fai

.que

uing

.(WF

.[De

m90],

.or.non-pr

eemptive

.ver

sions

.of.rat

.mono

tonic

.or.ear

liest

.dea

dline

.rs

[Li

u73]

.can.be.exp

loited

.to.dea

.wit

.rea

l-time

.tra

c

.and.cal

culate

.del

.boun

.for.it..How

ever,

.WFQ.

int

roduces

.a.sig

nicant

.com

putational

.over

head,

.whi

.the.ee

ctiveness

.of.pri

ority

.que

uing

.dep

ends

on.the.num

ber

.of.pri

orities,

.whi

.can.be.de

ned

.(th

.ne

.the.gra

nularity,

.the.bet

ter

.the.res

ulting

tra

c

.di

erentiation).

.Eve

.the.cap

acity

.of.the.que

ues

.of.the.net

work

.dev

ices

.the.pac

ket

.trav

erses

.on.

its.pat

.and.the.tot

.net

work

.loa

.have.to.be.tak

.int

.acc

ount,

.as.pac

ket

.dro

pping

.may.occ

.if.

any.of.the

.que

ues

.ll

.up..Sui

table

.ow.cont

rol

.alg

orithms,

.sel

ective

.dis

carding,

.and.pac

ket

.mar

ing

.policies.are.needed.to.support.QoS.for.real-time.trac..Another.signicant.contribution.to.the.

end-t

o-end

.del

.is.giv

.by.the.tra

nsmission

.del

ay,

.whi

.is.the.tim

.spa

.bet

ween

.the.emi

ssion

.of.

the.rs

.bit.and.the.emi

ssion

.of.the.las

.bit.of.a.fra

me.

. e.tra

nsmission

.del

.dep

ends

.on.the.pac

ket

siz

.and.the.net

work

.tra

nsmission

.rat

. e.pro

pagation

.tim

.de

ned

.as.the.tim

.bet

ween

.the.emi

sion

.of.the.rs

.bit.on.the.tra

nsmitter

.sid

.and.the.rec

eption

.of.suc

.a.bit.on.the.rec

eiver

.sid

.als

cont

ributes

.to.the.end-t

o-end

.del

ay,

.alt

hough

.its.imp

act

.for.loc

.are

.net

works

.is.neg

ligible.

.Cha

nnel

err

ors

.occ

urring

.dur

ing

.the.tra

nsmissions

.of.a.pac

ket

.may.ent

ail

.the.ret

ransmission

.of.the.cor

rupted

pac

ket.

.How

ever,

.whe

.dea

ling

.wit

.rea

l-time

.tra

c,

.esp

ecially

.wit

.per

iodic

.cont

rol

.pac

kets

.in.fac-

ory

.com

munication,

.the.adv

antage

.of.ret

ransmissions

.has.to. be.car

efully

.eva

luated,

.as.the

.is. a.

non-n

ull

.pro

bability

.tha

.the.ret

ransmitted

.dat

.bec

omes

.obs

olete

.dur

ing

.its.way.to.the.des

tination,

due.to.the.gen

eration

.of.a.“fr

esh”

.val

ue.

.How

ever,

.a.lar

.par

.of.the.end-t

o-end

.del

.at.the.app

lica-

tion

.lev

.is.due.to.the.lat

ency

.at.the.end

nodes.

.In.[Sk

e02],

.wit

.refe

rence

.to.swi

tched

.Eth

ernet,

.it.

17-6 Industrial Communication Systems

was.demonstrated.that.the.use.of.priorities.for.packets.in.the.communication.infrastructure.is.not.

on.its.own.su

cient

.to.gua

rantee

.tha

.app

lication-to-application

.dea

dlines

.wil

.be.met.and.tha

.the.

conce

.of.pri

ority

.has.to.be.ext

ended

.to.the.end.nod

.in.the.pro

tocol

.sta

ck.

.Oth

.wor

.dea

.wit

lim

itations

.of.swi

tched

.Eth

ernet

.as.far.as.rea

l-time

.com

munication

.is.conce

rned

.[Ja

s02].

.In.[Lo

B06],

.a.

sig

nicant

.red

uction

.in.the.roun

dtrip

.del

.for.hig

h-priority

.tra

c

.is.obt

ained

.tha

nks

.to.an.app

roach

bas

.on.bot

.a.pri

oritization

.mec

hanism

.in.the.pro

tocol

.sta

.and.mul

tiple

.rec

eption

.and.tra

nsmis-

sion

.que

ues

.in.the.end.nod

es.

Many

.net

worked

.rea

l-time

.sys

tems

.fea

ture

.app

lications

.wit

.fur

ther

.req

uirements

.tha

.tim

.con

straints

.onl

.for.ins

tance,

.mob

ility,

.dep

endability,

.com

posability,

.sca

lability,

.ex

ibility,

.sec

urity,

.and.

saf

ety.

. e.way.of.dea

ling

.wit

.rea

l-time

.con

straints

.in.the.pre

sence

.of.suc

.req

uirements

.dep

ends

.on.

the.pa

rticular

.sc

enario.

17.2.4 analytical Methods to assess Performance of real-time Networks

 e.analysis.of.real-time.networks.can.be.done.with.methods.that.have.emerged.either.in.the.eld.

of.rea

l-time

.sys

tems

.or.in.the.el

.of.com

munication

.net

works.

.Not

able

.exa

mples

.are.res

ponse-time

ana

lysis

.(R

TA)

.an

.ne

twork

.ca

lculus

.(N

C).

RTA

.is.a.met

hod

.to.ver

ify

.the.sche

dulability

.of.a.tas

.set.in.a.pro

cessor

.or.of.a.mes

sage

.set.in.a.

com

munication

.net

work,

.whi

.is.bas

.on.the.res

ponse

.tim

.com

putation

.and.com

parison

.wit

.the.

dea

dline

.for.eac

.tas

.or.mes

sage

.in.the.set

.If.the.dea

dlines

.are.met.for.eve

.ins

tance

.of.the.tas

.or.

mes

sages

.tha

.fea

ture

.rea

l-time

.con

straints,

.then.the.sche

dulability

.is.gua

ranteed.

.So,.RTA.pro

vides

an.exa

.sche

dulability

.tes

.RTA.has.bee

.rs

.pre

sented

.in.[Jo

s86],

.but.man

.evol

utions

.of.RTA.then.

foll

owed.

.In.[Au

d91],

.the.res

triction

.of.the.dea

dline

.bei

.equa

.to.the.per

iod

.was.rel

eased

.and.RTA.

was.ext

ended

.to.tas

.wit

.dea

dlines

.les

.or.equa

.to.the.per

iod.

.[Le

h90]

.ext

ended

.RTA.to.dea

dlines

gre

ater

.tha

.the.per

iod,

.thu

.ena

bling

.sev

eral

.act

ive

.ins

tances

.of.the.sam

.mes

sages.

.One.of.the.mos

use

.res

ults

.in.ind

ustrial

.com

munications

.is.the.wor

.wher

.RTA.was.ext

ended

.to.con

troller

.are

.net

work

.(CA

.[Ti

n94],

.thu

.acc

ounting

.for.non

-preemptive

.xe

d-priority

.sche

duling.

.Usi

.RTA

.the.

queu

ing

.del

.of.a.mes

sage

.can.be.obt

ained

.usi

.a.rec

ursive

.equa

tion.

.Foll

owing

.thi

.res

ult,

.sev

eral

spe

cic

.RTA

-based

.ana

lyses

.hav

.bee

.der

ived

.for.di

erent

.rea

l-time

.net

works,

.in.the.el

.of.ind

ustrial

aut

omation.

.In.[Al

m02],

.an.RTA.for.bot

.per

iodic,

.usi

.xe

.pri

orities,

.and.ape

riodic

.exch

anges

.of.

var

iables

.(me

ssages)

.is.pre

sented

.for.Wor

ldFIP,

.lea

ding

.to.a.sche

dulability

.con

dition

.nec

essary

.and.suf

cient

.for.the.per

iodic

.tra

c

.alt

hough

.jus

.su

cient

.for.the.ape

riodic

.one

.A.rec

ent

.wor

.by.Bri

.[Br

i06]

rev

isited

.the.RTA.for.ide

.CAN.net

work

.sho

wing

.by.mea

.of.exa

mples

.wit

.a.hig

.loa

.(≈9

8%)

.tha

the.analysis.as.presented.in.[Tin94].is.optimistic.. ey.proved.that,.assuming.discrete.scheduling,.the.

pro

blem

.can.be.res

olved

.by.app

lying

.the.ana

lysis

.for.xe

d-priority

.non

-preemptive

.sche

duling

.pre

sented

.in.[Ge

o96].

Network

.cal

culus

.was.int

roduced

.by.Cru

.in.[Cr

u91a,Cru91b]

.to.per

form

.det

erministic

.ana

lysis

.in.

net

works

.whe

.the.inc

oming

.tra

c

.is.not.ran

dom

.but.unk

nown,

.bei

.lim

ited

.by.a.kno

.arr

ival

cur

ve.

.Net

work

.cal

culus

.di

ers

.fro

.que

uing

.the

ory

.in.the.sen

.tha

.it.dea

.wit

.wor

.cas

.ins

tead

of.ave

rage

.cas

es.

. e.par

ameters

.of.int

erest

.in.thi

.met

hodology

.are.the.del

ay,

.the.bu

.all

ocation

req

uirements,

.and.the.thr

oughput.

.Usin

.the.arr

ival

.cur

ve,

.oe

.nam

.α,.and.a.ser

vice

.cur

.oe

red

by.the.net

work,

.nam

.β,.and.con

sidering

.a.tra

c

.ow.R(t),.it.is.pos

sible

.to.det

ermine

.upp

.bou

nds

.for.

the.ba

cklog

.(t

.nu

mber

.of.bi

.in.“t

ransit”

.in.th

.sy

stem

.[B

ou04])

.an

.th

.de

lay.

Network

.cal

culus

.has.bee

.use

.in.sev

eral

.rea

l-time

.dom

ains

.suc

.as.ind

ustrial

.auto

mation,

.avi

on-

ics,

.and.wir

eless

.sen

sor

.net

works.

.Con

sidering

.ind

ustrial

.auto

mation,

.in.[Ge

o04]

.net

work

.cal

culus

.was.

use

.to.mod

.a.swi

tched

.Ethe

rnet

.arc

hitecture

.in.ord

.to.eva

luate

.the.max

imum

.end

-to-end

.del

ays

tha

.are.cri

tical

.for.ind

ustrial

.app

lications.

.In.[Ri

d07],

.a.stoc

hastic

.net

work

.cal

culus

.app

roach

.to.eva

lu-

ate

.the.distribution.of.end-to-end.delays.in.the.AFDX.(avionics.full-duplex.switched.Ethernet,.ARINC.

664

)

.net

work

.use

.in.air

cras

.suc

.as.Air

bus

.A38

.is.pre

sented.

.Net

work

.cal

culus

.has.als

.bee

.app

lied

to.th

.di

mensioning

.of.wi

reless

.se

nsor

.ne

tworks

.[S

ch05].