Upload
others
View
8
Download
0
Embed Size (px)
Citation preview
1
CSCI325:DistributedSystems
ProfessorSprenkle
Objec?ves• Courseoverview• Overviewofdistributedsystems• Introduc?ontoreadingresearchpapers
Sept8,2017 Sprenkle-CSCI325 2
2
DistributedSystems?• Whatisadistributedsystem?• Knowanyexamplesofdistributedsystems?
Ø Anyused?
Sept8,2017 Sprenkle-CSCI325 3
DistributedSystems?• Whatisadistributedsystem?
Ø Collec?onsofindependent,networkedcomputersworkingtogether
• ExamplesofdistributedsystemsØ Networkedprinters,storageØ InternetØ Peer-to-peersystemsØ Gridcompu?ngØ GamesØ Sensornetworks
Sept8,2017 Sprenkle-CSCI325 4
3
Sept8,2017 Sprenkle-CSCI325 5
DistributedSystemsArchitectures• Twomainmodels
Ø Client-server• Mostcommon,andarguablythesimplest• Mul?-?erclient-server
Ø Varia?ononthesimpleclient-serverarchitectureØ Mul?plelevelsofcommunica?on
Ø Peer-to-peer• Allprocessesinvolvedinataskorac?vityplaysimilarroles
Sept8,2017 Sprenkle-CSCI325 6
4
Client-ServerModel
Sept8,2017 Sprenkle-CSCI325 7
Client
Client
Client
ClientClient
Server
request
request
requestresponse
response
response
Peer-to-PeerSystems
Peer
Peer
Peer
PeerPeer
request
request
request
reque
strequest
response
response
response
response
Sept8,2017 8Sprenkle-CSCI325
Connections between peers
5
Challenges• Whatarechallengesindealingwithdistributedsystems?
Sept8,2017 Sprenkle-CSCI325 9
DistributedSystemsChallenges• Communica?on• Naming• Distribu?onofworkload• Distribu?ontransparency• Consistency• Handlingfailure• Security• Scaling
Sept8,2017 Sprenkle-CSCI325 10
6
WhatThisCourseisAbout• Networkingfundamentals• Distributedsystems
Ø ChallengesofdistributedsystemsØ DesignprinciplesØ Learnhowtobuildlarge-scaledistributedsystems
• Severalprogrammingprojects• Emergingresearchissues
Ø Studyfundamentalresearchpapers• Life-skills
Ø Reading,wri?ng,discussion,presenta?on
Sept8,2017 Sprenkle-CSCI325 11Overall goal: Emphasize “why” and “how” over “what”
Bonus: OS
ALITTLEBITOFHISTORYWhatmadedistributedsystemspossible?
Sept8,2017 Sprenkle-CSCI325 12
7
TheInternet• Connec?onofcomputernetworksusingtheInternetProtocol(IP)Ø Allowsnetworkapplica?ons,e.g.,email,filetransfer,world
wideweb,remotelogin,…
Sept8,2017 Sprenkle-CSCI325 13
Internet
VannevarBush• EstablishedtheU.S.military/universityresearch
partnershipthatdevelopedARPANET• Wrote1stvisionarydescrip?onofpoten?aluse
forinforma?ontechnologyØ inspiredmanyofInternet'screators
Sept8,2017 Sprenkle-CSCI325 14
Source: Livinginternet.com
Could you envision the WWW years before it existed?
“Consider a future device for individual use, which is a sort of mechanized private file and library. It needs a name, and to coin one at random, ‘memex’ will do. A memex is a device in which an individual stores all his books, records, and communications, and which is mechanized so that it may be consulted with exceeding speed and flexibility. It is an enlarged intimate supplement to his memory. “It consists of a desk, and while it can presumably be operated from a distance, it is primarily the piece of furniture at which he works. On the top are slanting translucent screens, on which material can be projected for convenient reading. There is a keyboard, and sets of buttons and levers. Otherwise it looks like an ordinary desk.”
- Vannevar Bush, “As We May Think,” Atlantic Monthly, July 1945
8
J.C.R.Licklider• JosephCarlRobnef“Lick”Lickliderdevelopedideaofuniversalnetwork
• SpreadhisvisionthroughouttheIPTO(Informa?onProcessingTechniquesOffice)
• Inspiredhissuccessorstorealizehisdreambycrea?ngARPANET
Sept8,2017 Sprenkle-CSCI325 15
Source: Livinginternet.com
“It seems reasonable to envision, for a time 10 or 15 years hence, a ‘thinking center’ that will incorporate the functions of present-day libraries together with anticipated advances in information storage and retrieval.“The picture readily enlarges itself into a network of such centers, connected to one another by wide-band communication lines and to individual users by leased-wire services. In such a system, the speed of the computers would be balanced, and the cost of the gigantic memories and the sophisticated programs would be divided by the number of users.”�
- J.C.R. Licklider, Man-Computer Symbiosis, 1960.
Background• 1957:USSRlaunchesSputnik,firstar?ficialearthsatelliteØ U.S.respondsbyformingAdvancedResearchProjectsAgency(ARPA)
• 1962:Licklider’sGalac?cNetwork• 1966:MarillandRoberts(MIT)paper:“TowardaCoopera?veNetworkofTime-SharedComputers”Ø hfp://dl.acm.org/cita?on.cfm?id=1464336
• 1967:Roberts(MIT):ACMSOSP“Mul?pleComputerNetworksandIntercomputerCommunica?on”Ø hfp://dl.acm.org/cita?on.cfm?id=811680
Sept8,2017 Sprenkle-CSCI325 16
9
1969InternetMap:ARPANET
Sept8,2017 Sprenkle-CSCI325 17
1st assignment:draw today’s Internet
Stanford Research Institute
1st message: “LO” as in “Lo and Behold”(supposed to be “LOG” but failure!)From UCLA to SRIOct 29, 1969, 10:30 p.m.
SDSSigma7
http://www.nsf.gov/news/special_reports/nsf-net/1960s.jsp https://www.nsf.gov/news/special_reports/nsf-net/kleinrockvideopop.html
Sept8,2017 Sprenkle-CSCI325 18
10
InternetTimeline
Sept8,2017 Sprenkle-CSCI325 19
Year Milestone1971 Tomlinsondevelopsemailprogram,bighit1972 Telnet1973 FileTransferProtocol(FTP)
1974 TransmissionControlProtocol(TCP)
1978 TCPsplitintoTCPandIP(InternetProtocol)1979 USENET(newsgroup)established
1984 1000hostsconnectedtoInternet,DNSintroduced
1988 Internetwormbringsdown10%ofInternet1991 WAIS,Gopher,WWWreleased
InternetGrowthTrends
Sept8,2017 Sprenkle-CSCI325 20
Year Hosts on Internet1977 1111981 2131983 5621984 1000
1986 50001987 10,0001989 100,000
1992 1,000,0002001 151-175 million2002 Over 200 million2014 1.01 billion
# of computers connected directly to the Internet increased at a yearly rate >37% across 21 years
https://www.internetsociety.org/sites/default/files/Global_Internet_Report_ 2014_0.pdf
11
Sta?s?csfromtheIITFReport“TheEmergingDigitalEconomy”*• Togetamarketof50Millionpeoplepar?cipa?ng:Ø Radio:38yearsØ TV:13yearsØ Internet:4years
• Aqeropentogeneralpublic• hfp://govinfo.library.unt.edu/ecommerce/EDEreprt.pdfØ ReleasedonApril15,1998
Sept8,2017 Sprenkle-CSCI325 21
* Delivered to the President and the U.S. Public on April 15,1998 by Bill Daley, Secretary of Commerce and Chairman of the Information Infrastructure Task Force
Sept8,2017 Sprenkle-CSCI325 22
12
COURSEINFO
Sept8,2017 Sprenkle-CSCI325 23
MyResponsibili?es• Prepareuseful,interes?ngknowledge• Cometoclassprepared,on?me• Interes?ng,relevant,andchallengingassignments
• Promptfeedbackonassignments
Sept8,2017 Sprenkle-CSCI325 24
13
YourResponsibili?es• Cometoclassprepared,on?me,andPARTICIPATE
• Turninassignmentson$me• Whenyou’rehavingtrouble
Ø LookforhelpontheWeb• Find,adaptsolu?ons• Givecredittowhereyoufoundsolu?on,ifnovelenough
Ø Askmeforhelp!• Learn,absorb,synthesize
Ø ExtraCredit:takeittothenextlevelSept8,2017 Sprenkle-CSCI325 25
Textbook• Required:DistributedSystems,byvanSteenandTanenbaum,3rded.Ø Providesbackgroundforclassdiscussionsandprojects
Ø Availableonline
• Op?onal:DistributedSystems,ConceptsandDesigns,byCourlouis,Dollimore,Kindberg,5thed.
Sept8,2017 Sprenkle-CSCI325 26
14
Grading• 17%Individualprogramming,reading,wri?ngassignments
• 20%Midtermexam• 33%Programmingprojects• 25%FinalProject
Ø Includingpaperandpresenta?onØ Startthinkingaboutpossibletopics
• 5%Par?cipa?onandafendanceØ Successofclassdependsonstudentpar?cipa?on
Sept8,2017 Sprenkle-CSCI325 27
ProgrammingProjects• 3projectsspanningthesemester
Ø Hands-onconstruc?onofinteres?ngdistributedservices
Ø Approximately2.5weekstocompleteØ Workinteamsof2or3
• UseversioncontrolØ Startearly!
Sept8,2017 Sprenkle-CSCI325 28
15
READINGRESEARCHPAPERS
Sept8,2017 Sprenkle-CSCI325 29
WhattoLookForWhileReading• Overallproblem
Ø Howlarge/importantistheproblem?• Goals• Contribu?ons
Ø Keywords:new,novel• Technicalapproach
Ø Keyinsights(“leverage”,“u?lize”)• Evalua?on
Ø Answersallyourques?onsaboutapproach?• Limita?ons
Ø Maynotbeageneral-purposesolu?onØ Checkassump?ons
Sept8,2017 Sprenkle-CSCI325 30
16
SomeConcreteQues?ons• StatementoftheProblem/Goals
Ø Trytostatesuccinctlytheoverallproblembeingaddressedinthispaper.Ø Whatpar?culargoalsdotheseresearchershaveinaddressingthis
problem?Ø Whatcontribu?onaretheyseekingtomaketothestate-of-the-art?
• TechnicalApproachØ Whatisthekeyinsightofthisgroup'sapproachtotacklingthestated
problem?Whatistheiroverallapproach/strategytosolvingtheproblem?• Discussion/Cri?que
Ø Howdidtheresearchersevaluatetheirefforts?Ø Whatconclusionsdidtheymakefromtheirevalua?onresults?Ø Whatapplica?on/usefulbenefitdotheresearchers/youseeforthiswork?Ø Whatlimita?onsdotheresearchersmen?onwiththeirapproach?Ø Whataddi?onallimita?onsdoyouthinkthereare?Ø Writeoneinteres?ngques?ontoponderwithregardtothispaperbeyond
contentunderstanding.
Sept8,2017 Sprenkle-CSCI325 31
SEDA
Sept8,2017 Sprenkle-CSCI325 32
WeproposeanewdesignforhighlyconcurrentInternetservices,whichwecallthestagedevent-drivenarchitecture(SEDA).SEDAisintendedtosupportmassiveconcurrencydemandsandsimplifytheconstruc?onofwell-condi?onedservices.InSEDA,applica?onsconsistofanetworkofevent-drivenstagesconnectedbyexplicitqueues.Thisarchitectureallowsservicestobewell-condi?onedtoload,preven?ngresourcesfrombeingovercommifedwhendemandexceedsservicecapacity.SEDAmakesuseofasetofdynamicresourcecontrollerstokeepstageswithintheiropera?ngregimedespitelargefluctua?onsinload.Wedescribeseveralcontrolmechanismsforautoma?ctuningandloadcondi?oning,includingthreadpoolsizing,eventbatching,andadap?veloadshedding.WepresenttheSEDAdesignandanimplementa?onofanInternetservicesplawormbasedonthisarchitecture.WeevaluatetheuseofSEDAthroughtwoapplica?ons:ahigh-performanceHTTPserverandapacketrouterfortheGnutellapeer-to-peerfilesharingnetwork.TheseresultsshowthatSEDAapplica?onsexhibithigherperformancethantradi?onalservicedesigns,andarerobusttohugevaria?onsinload.
17
SEDA• Problem/Goals
Ø Highlyconcurrentinternetsystems• Goal:well-behavedunderload
• TechnicalApproachØ Staged,event-drivenarchitecture(SEDA)Ø Automa?ctuning,loadcondi?oning
• DiscussionØ Evalua?on:UsedSEDAarchitectureforwebserver,P2Ppacketrouter• Measuredperformance,robustnesstoloadvaria?on
Sept8,2017 Sprenkle-CSCI325 33
ReadingFeedback:Annota?ons• Perusall:Applica?onaccessiblethroughSakai
Ø Allowsyoutocommentonanar?clesuchthatallstudentsandprofessorcanviewthem
• Youwillbeexpectedtomakeacertainnumberofannota?onsoneachar?cleØ “Certainnumber”=5,typically
• Annota?onscanbeques?onsorcommentsØ mustbesubstan?ve
• Eachannota?onwillbegradedasØ 2:thoughwul;full-creditØ 1:par?al-creditØ 0:thoughtlessornotcomplete;nocredit
Sept8,2017 Sprenkle-CSCI325 34
18
TODO• SetupPerusall,throughSakai• ExploreCourseWebPage• Checkout“WelcometotheMachine”
Ø ReviewingsometermsfromCSCI210(plusmaybemore)
• ReadE2EArgumentpaperforFridayØ Skimthroughonce,reviewsec?onheadingsØ 3hoursmaxØ Reviewpaper
• Write5annota?onsinPerusallØ Wed:Discusspaperandques?ons
Sept8,2017 Sprenkle-CSCI325 35