Fusco Andrzej Pelc and Rossella Petreschi Computer Science Department Sapienza University of Rome 00198 Rome Italy fusco petreschi diuniroma1it Departement dinformatique Universite du Quebec en Outaouais Gatineau Quebec J8 ID: 8496 Download Pdf
Allow colocation of ISP server inside our network Winning ISP decided on WuG in 1999 Probably today it would be Nagios 1000 Euromonth The ISPs have 24 7 monitoring and we become just one additional screen in their NOC brPage 3br How to Build a 24 7
How to remove Steiner points. Lior. . Kamma. , . Weizmann Institute of Science. Joint work with . Robert . Krauthgamer. and . Huy. L. . Nguyễn. TexPoint. fonts used in EMF. . Read the . TexPoint.
Understand how to do it. Michael Niehaus. Director of Product Marketing. mniehaus@microsoft.com. Brian McNeill. Principal . PM Manager. brianm@microsoft.com. BRK3136. Why Windows as a service?. What customers are telling us.
Oliver Schulte. Zhensong. Qian. Arthur. Kirkpatrick. Xiaoqian. . Yin. Yan. Sun. Relational Dependency Networks. Neville, J. & Jensen, D. (2007), 'Relational Dependency Networks', . Journal of Machine Learning Research .
Les Mayhew . Cass Business School. International Geographical Union (IGU) conference . Leeds University, 2013. lesmayhew@googlemail.com. 1. Geographical theory and urban analogues. Classical geographical theory.
Les Mayhew . Cass Business School. International Geographical Union (IGU) conference . Leeds University, 2013. lesmayhew@googlemail.com. 1. Geographical theory and urban analogues. Classical geographical theory.
Feedback from the Working Group. Scott and Trevor. What is Fast-track?. What happened at Winterbourne View shocked everyone. . ‘Transforming Care’ is a plan to change things.. Nottinghamshire is one of the Fast-track areas who will try the plan first..
Mathematic. al Modeling by Implementing ‘Lesson Study’ in the Case of Volcanoes Eruption?. Subanar. Fadjar Shadiq. SEAMEO QITEP in Mathematics. PowerPoint Presented on . The 8th APEC - Tsukuba International Conference: Innovation of Mathematics Education through Lesson Study. Challenges to Emergency Preparedness for Mathematics (III). Focus on Fire and Volcanic Eruption.
Understanding how cost impacts where things are grown.. Commercial Agriculture . Commercial Agriculture:. Element of 2. nd. Agricultural Revolution. Farmers & Ranchers sell output for money . Money used to buy stuff at stores.
How you learn and. w. hy . some things stick…. and some things don’t!. Kindergarten was fun…. Later, . did you think that… . . . School was hard work?. Learning to . read wasn’t so easy.
Fusco Andrzej Pelc and Rossella Petreschi Computer Science Department Sapienza University of Rome 00198 Rome Italy fusco petreschi diuniroma1it Departement dinformatique Universite du Quebec en Outaouais Gatineau Quebec J8
Download Pdf - The PPT/PDF document "Learning a ring cheaply and fast Emanuel..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
thatnodescansolvethemsimulatingacentralmonitor.Weareinterestedintheeciencyofdeterministicalgorithmsforlabeledmapconstruction.InthispaperweusetheextensivelystudiedLOCALmodelofcommunication[12].Inthismodel,communicationproceedsinsynchronousroundsandallnodesstartsimultaneously.Ineachroundeachnodecanexchangearbitrarymessageswithallitsneighborsandperformarbitrarylocalcomputations.Thetimeofcompletingataskisthenumberofroundsittakes.Ourgoalistoinvestigatetradeosbetweenthetimeofconstructingalabeledmapanditscost,i.e.,thenumberofmessagesneededtoperformthistask.Toseeextremeexamplesofsuchatradeo,considerthemapconstructiontaskonann-nodering.ThefastestwaytocompletethistaskisintimeD,whereD=bn=2cisthediameterofthering.Thiscanbeachievedby ooding,butthenumberofmessagesusedisthen(n2).Ontheotherhand,cost(n)(whichisoptimal)canbeachievedbyaversionofthetimeslicingalgorithm[11],butthentimemaybecomeverylargeanddependsonthelabelsofthenodes.Thegeneralproblemoftradeosbetweentimeandcostoflabeledmapcon-structioncanbeformulatedasfollows.ForagiventimeT,whatisthesmallestnumberofmessagesneededforconstructingalabeledmapbyeachnodeintimeT?Fortreesthisproblemistrivial:leavesofann-nodetreeinitiatethecom-municationprocessandinformationabouteverlargersubtreesgetsrsttothecentralnode(orcentralpairofadjacentnodes)andthenbacktoallleaves,usingtimeequaltothediameterofthetreeandO(n)messages,bothofwhichareop-timal.However,assoonastherearecyclesinthenetwork,thereisnocanonicalplacetostartinformationexchangeoneachcycleandproceedingfastseemstoforcemanymessagestobesentinparallel,whichinturnintuitivelyimplieslargecost.Thisphenomenonispresentalreadyinthesimplestsuchnetwork,i.e.,thering.Indeed,ourstudyshowsthatmeaningfultradeosbetweentimeandcostoflabeledmapconstructionalreadyoccurinrings.Weconsiderringswhosenodeshaveuniquelabelsthatarebinarystringsoflengthpolynomialinthesizeofthering.(Ourresultsarevalidalsoformuchlongerlabels,butthesecanbedismissedforpracticalityreasons.)Inthebegin-ning,everynodeknowsonlyitsownlabel,theallowedtimeTandthediameterDofthering.Equivalently,weprovideeachnodewithitslabel,withthedi-ameterDandwiththedelay=TD,whichistheextratimeallowedontopoftheminimumtimeDinwhichlabeledmapconstructioncanbeachieved,knowingDapriori.Knowingitsownlabelisanobviousassumption.Withoutanyadditionalknowledge,nodeswouldhavetoassumetheleastpossibletimeandhencedo oodingatquadraticcost.InsteadofprovidingnodeswithDand,wecouldhaveprovidedthemonlywiththealloweddelayovertheleastpossibletimeoflearningtheringwithoutaprioriknowledgeofthediameter.Thiswouldnotaectourasymptoticbounds.However,itwouldresultinmorecumbersomeformulationsbecause,withoutknowingDapriori,theoptimaltimeoflabeledmapconstructionvariesbetweenDandD+1,dependingonwhethertheringis eterD.Weprovethelowerboundontheclassoforientedringsofevensize.(Restrictingtheclassonwhichthelowerboundisprovedonlyincreasesthestrengthoftheresult.)Weformalizeorientationbyassigningportnumbers0and1intheclockwiseorderateachnode.Foreverynodev,let`(v)beitslabel.WerstdenethehistoryH(v;t)ofnodevattimet.IntuitivelyH(v;t)rep-resentstheentireknowledgethatnodevcanacquirebytimet.Sincewewanttoprovealowerboundoncost,itisenoughtoassumethatwheneveranodevsendsamessagetoaneighborinroundt+1,thecontentofthismessageisitsentirehistoryH(v;t).Wedenehistoriesofallnodesbysimultaneousinductionont.DeneH(v;0)astheone-elementsequenceh`(v)i.Intheinductivedeni-tion,wewillusetwosymbols,s0ands1,correspondingtothelackofmessage(silence)onport0and1,respectively.Assumethathistoriesofallnodesaredeneduntilroundt.WedeneH(v;t+1)as:{hH(v;t);s0;s1i,ifvdidnotgetanymessageinroundt+1;{hH(v;t);s0;H(u;t)i,ifvdidnotgetanymessageinroundt+1onport0butreceivedamessageonport1fromitsclockwiseneighboruinthatround;{hH(v;t);H(w;t);s1i,ifvdidnotgetanymessageinroundt+1onport1butreceivedamessageonport0fromitscounterclockwiseneighborwinthatround;{hH(v;t);H(w;t);H(u;t)i,ifvreceivedamessageonport0fromitscounter-clockwiseneighborwandamessageonport1fromitsclockwiseneighboru,inroundt+1.WedeneacommunicationpatternuntilroundtforthesetEofalledgesoftheringasafunctionf:Ef1;:::;tg!f0;1g,wheref(e;i)=0,ifandonlyifnomessageissentonedgeeinroundi.ExecutingamapconstructionalgorithmAonagivenringdeterminesacommunicationpattern,whichinturndetermineshistoriesH(v;t),forallnodesvandallroundst.Foranypathk=hu0:::ukibetweennodesu0andukwedene,byinduc-tiononk,thecommunicationdelay(k;f)inducedonkbythecommunica-tionpatternf.Fork=1,(1;f)=d,ifandonlyif,f(fu0;u1g;i+1)=0,forallid,andf(fu0;u1g;d+1)=1.Inparticular,iff(fu0;u1g;1)=1then(1;f)=0.Supposethat(k1;f)hasbeendened.Wedene(k;f)=(k1;f)+d,ifandonlyif,f(fuk1;ukg;(k1;f)+k+i)=0,forallid,andf(fuk1;ukg;(k1;f)+k+d)=1.Inparticular,iff(fuk1;ukg;(k1;f)+k)=1then(k;f)=(k1;f).Intuitivelythecommunicationdelayonapathbetweenuandvindicatestheadditionaltime,withrespecttothelengthofthispath,thatitwouldtakenodevtoacquireanyinformationaboutnodeu,alongthispath,ifnoinformationcouldbecodedbysilence.Infactsomeinformationcanbecodedbysilence,andanalyzingthisphenomenonisthemainconceptualdicultyofourlowerboundproof.Inparticular,wewillshowthatifmapcon-structionhastobeperformedquickly,thenthenumberofcongurationsthatcanbecodedbysilenceissmallwithrespecttothetotalnumberofpossibleinstances,andhencemanymessageshavetobeusedforsomeofthem.Wedenethecommunicationdelayinducedbyacommunicationpatternfbetweenanodexanditsantipodalnode xastheminimumofthedelays pairscomposedofanedgeandaroundnumber,oftheform(fxh+p;xh+p+1g;rp),for0pD,whererp=(hxh:::xh+p+1i;g)+p+1.Bythedenitionofcommunicationdelayinducedonapath,wehaveg(fxh+p;xh+p+1g;rp)=1,forall0pD.WenowshowthatsetsZyarepairwisedisjoint.PicktwonodesfromY:anodey=xhandanodey0=xh+datclockwisedistancedDfromxh.Consideranodexh+p,withdpD.Considertwopairs,(fxh+p;xh+p+1g;rp)2Zxhand(fxh+p;xh+p+1g;rpd)2Zxh+d.Since(hxh+d:::xh+p+1i;g),wehaverpd=pd+(hxh+d:::xh+p+1i;g)+1pd++1.BydenitionofY,wehaved-353;,hencerp=p+(hxh:::xh+p+1i;g)+1p+1-353;pd++1rpd.Thisimpliesthatrp6=rpdandhencesetsZyandZy0aredisjoint.Noticethatify0isatdistanceDfromy,theny0istheantipodalnodeofyandhenceZyandZy0aredisjointbecausetheedgesintheirelementsaredierent.ItfollowsthatallsetsZyarepairwisedisjoint,hence[y2YZyhasatleastDbD=(+1)celements.Sinceeachelementcorrespondstoatleastonemessagesent,weconcludethatthealgorithmusesatleastDbD=(+1)c2 (D2=)messages.ut3ThealgorithmThegeneralideaofourlabeledmapconstructionalgorithmistospendthealloweddelayinapreprocessingphasethatdeactivatessomenodes,usingtheresidualtimeDforaphasedevotedtoinformationspreading.Thisresultsinareductionoftheoverallcostofthealgorithm,withrespectto ooding,sincenon-activenodesareonlyresponsibleforrelayingmessagesoriginatedatnodesthatremainedactiveafterthepreprocessingphase.Hence,thisapproachrequirestodeactivateasmanynodesaspossible.However,withindelay,wecannotaordtodeactivatesequencesofconsecutivenodesoflengthlargerthan2.Indeed,deactivatingsuchlongsequenceswouldimplythatthelabelofsomenon-activenodeisunknowntoallactivenodes,whichwouldmakethetimeoftheinformationspreadingphaseexceedtheremainingDrounds.Wereconciletheseoppositerequirementsbydeninglocalrulesthatallowustodeactivatealmosthalfofthecurrentlyactivenodes,withoutdeactivatingtwoconsecutiveones.Thisprocessistheniteratedasmanytimesaspossiblewithindelay.Thepreprocessingphaseofouralgorithmisdividedintostages,eachofwhichisinturncomposedofmultiplesteps.Intherststage,allnodesareactive.Nodesthatbecomenon-activeattheendofastagewillneverbecomeactiveagain.Inordertosimplifythedescriptionofthealgorithm,wewillusetheconceptofresidualring.Insucharing,thesetofnodesisasubsetoftheoriginalsetofnodes,andedgescorrespondtopathsofconsecutiveremovednodes.Inparticular,stageiisexecutedontheresidualringRicomposedofnodesthatremainedactiveattheendofthepreviousstage.CommunicationbetweenconsecutivenodesRiissimulatedbyamulti-hopcommunicationintheoriginalring,wherenon-activenodesrelaymessagesofactivenodes.Eachsimulatedmessageexchangeduringstageiisallotted2i1rounds. {IfSisofevenlength,i.e.,S=hlak:::a1b1:::bkri,nodesatandbtbecomenon-active,forallevenvaluesoft.Thismeansthateverysecondnodeisdeactivated,startingfromtheneighborsofthetwocentralnodes.Wearenowreadytoprovideadetaileddescriptionofourlabeledmapcon-structionalgorithm.Foreachtaskthatcannotbecarriedoutlocally,weallotaspecicnumberofroundstomaintainsynchronizationbetweentheexecutionofagivenpartofthealgorithmbydierentnodes.Intheanalysiswewillshowthattheallottedtimesarealwayssucient.AlgorithmRingLearningInput:D;;and".Phase1{preprocessingsetallnodesasactive{(locally);fori 1toblogc2dlog(8=")edlog((logD+3))e//STAGEconstructtheresidualringRiofactivenodes{(locally);electallnodesinRias(i;0)-leaders{(locally);forj 1todlog(8=")e//STEPconstructtheresidualringRi;jof(i;j1)-leaders{(locally);assigncolorc(u)toallnodesuinRi;jwithprocedureRTC(i;j);(allottedtime2i14j1(logD+1))elect(i;j)-leaderswithprocedureElect(i;j);(allottedtime2i14j1)runprocedureDeactivate(i,")inRi;(allottedtime2i14dlog(8=")e)Phase2{informationspreadinginround+1eachnodethatisstillactiveconstructslocallyalabeledmapofthepartoftheoriginalringconsistingofnodesfromwhichitreceivedmessagesduringPhase1,andsendsthismaptoitsneighbors;bothactiveandnon-activenodesthatreceiveamessagefromoneneighbor,sendittotheotherneighbor;attimeD+,allnodeshavethelabeledmapoftheringandstop.WenowprovethecorrectnessofAlgorithmRingLearningandanalyzeitbyestimatingitscostforagivendelay.Thersttwolemmasshowthatthetime2i1allottedformulti-hopcommunicationbetweenconsecutiveactivenodesinstagei,andthetime2i14j1allottedformulti-hopcommunicationbetweenconsecutive(i;j1)-leadersinstepjofstagei,aresucienttoperformtherespectivetasks.Lemma2.Thedistancebetweentwoconsecutive(i;j)-leadersisatmost2i14j.ThenexttwolemmaswillbeusedtoprovethecorrectnessofAlgorithmRing-Learning.Lemma3.AllcallstoproceduresRTC,Elect,andDeactivatecanbecarriedoutwithintimesallottedinAlgorithmRingLearning. Proof.AsshownintheproofofLemma4,thetimeusedforstageiisatmost2i14s(logD+3),wheres=dlog(8=")eisthenumberofstepsineachstage.InviewofLemma5,duringstageithereareatmostn(("=2+1)=2)i1activenodesinaringofsizen.Hencethecostofstageiisatmost2i14s(logD+3)n"=2+1 2i1:Sincethenumberofstagesislessthanlog,theoverallcostofthepreprocessingphaseislessthanblogcXi=12i14s(logD+3)n"=2+1 2i1:BoundingeachsummandwiththelastonewhichisthelargestweobtainblogcXi=12i14s(logD+3)n"=2+1 2i1n(logD+3)log(1+"=2)log8 "2;whichisO(DlogDlog(1+")log="2).utLemma7.ThecostoftheinformationspreadingphaseofAlgorithmRingLearningwithinputparametersD;,and",where0"1,isO(D2logD=("21")).Theorem3.ThecostofAlgorithmRingLearning,executedintimeD+inaringofdiameterD,isO(D2logD=1"),foranyconstantparameter0"1andanyD.Proof.Lemmas6and7implythatthecostofAlgorithmRingLearning,ex-ecutedwithparametersD,,and",inaringofdiameterD,isoftheor-derO(DlogDlog(1+"=2)log+D2logD=(1"));foranyconstant0"1.Sincelog(1+"=2)"isnegativeforall"]TJ/;ø 9;.962; Tf; -31;.75; -1;.95; Td; [00;0,andD,wehave1+log(1+"=2)"logD,forsucientlylargeD.HenceD 1"log(1+"=2)log;whichimpliesODlogDlog(1+"=2)log+D2logD 1"=OD2logD 1":ut4DiscussionandopenproblemsWeprovedalmostmatchingupperandlowerboundsforthetradeosbetweentimeandcostofthelabeledmapconstructiontaskintheclassofrings.Canthesetradeosbegeneralizedtoalargerclassofnetworks?Sincelowerboundsarestrongerwhenestablishedonamorerestrictedclassofgraphs,thechallenge suchtradeosbeestablishedforsomeotherclassesofnetworks(suchasboundeddegreenetworksorevenjustgridsandtori),similarlyaswedidforrings?Finally,noticethatforringstheinformationspreadingphasecanbeper-formedintime2D(insteadof3D)bylettingeachactivenodeinitiatetwosequencesofmessages(oneclockwise,andtheothercounterclockwise),eachcontaininglabelsofallalreadyvisitednodes.Moreover,theoverallcostofthedoublingalgorithm,executedintime2D+onaringofdiameterDandsizen,isO(nlog+nD=)=O(Dlog+D2=).ThisshouldbecomparedtothecostofAlgorithmRingLearning,thatcanbeassmallasO(D1+"logD)fortotaltime2Dandanyconstant"0.Thecostofthedoublingalgorithmbecomesasymptoticallysmallerwhentheoveralltimeislargerthan2D+D1"=logD.Closingthesmallgapbetweenourboundsonthetimevs.costtradeosforlabeledmapconstructiononringsisanotheropenproblem.References1.H.Attiya,A.Bar-Noy,D.Dolev,D.Koller,D.Peleg,andR.Reischuk,Renaminginanasynchronousenvironment,JournaloftheACM37(1990),524{548.2.B.Awerbuch,Optimaldistributedalgorithmsforminimumweightspanningtree,counting,leaderelectionandrelatedproblems,Proc.19thAnnualACMSymposiumonTheoryofComputing(STOC1987),230{240.3.J.Chalopin,S.Das,andA.Kosowski,Constructingamapofananonymousgraph:Applicationsofuniversalsequences,Proc.14thInternationalConferenceonPrin-ciplesofDistributedSystems(OPODIS2010),119{134.4.R.ColeandU.Vishkin,Deterministiccointossingwithapplicationstooptimalparallellistranking,InformationandControl70(1986),32-53.5.A.Czumaj,L.Gasieniec,andA.Pelc,Timeandcosttrade-osingossiping,SIAMJournalonDiscreteMathematics11(1998),400-413.6.G.N.FredricksonandN.A.Lynch,Electingaleaderinasynchronousring,JournaloftheACM34(1987),98{115.7.L.Gasieniec,A.Pagourtzis,I.Potapov,andT.Radzik.Deterministiccommuni-cationinradionetworkswithlargelabels.Algorithmica47(2007),97{117.8.A.V.Goldberg,S.A.Plotkin,andG.E.Shannon,Parallelsymmetry-breakinginsparsegraphs,SIAMJournalonDiscreteMathematics1(1988),434{446.9.D.S.Hirschberg,andJ.B.Sinclair,Decentralizedextrema-ndingincircularcon-gurationsofprocesses,CommunicationsoftheACM23(1980),627{628.10.A.Israeli,E.Kranakis,D.Krizanc,andN.Santoro,Time-messagetrade-osfortheweakunisonproblem,NordicJournalofComputing4(1997),317{341.11.N.L.Lynch,Distributedalgorithms,MorganKaufmannPubl.Inc.,SanFrancisco,USA,1996.12.D.Peleg,DistributedComputing,ALocality-SensitiveApproach,SIAMMono-graphsonDiscreteMathematicsandApplications,Philadelphia2000.13.G.L.Peterson,AnO(nlogn)unidirectionaldistributedalgorithmforthecircularextremaproblem,ACMTransactionsonProgrammingLanguagesandSystems4(1982),758{762.14.M.YamashitaandT.Kameda,Computingonanonymousnetworks:PartI-characterizingthesolvablecases,IEEETrans.ParallelandDistributedSystems7(1996),69{89.
© 2021 docslides.com Inc.
All rights reserved.