Advanced Computer Networks cs538, Spring 2016 @ UIUC

Advanced Computer Networks cs538, Spring 2016 @ UIUC

KlaraNahrstedt

Lecture2Basedon

B.M.Leineretal,“BriefHistoryoftheInternet”,InternetSociety2014C.ParCdge,“FortyDataCommunicaConsResearchQuesCons”,CCR,2011Prior2010-2013,lecturematerialbyBrightenGodfreyandMaNCaesar

Announcements

•  SelectreviewsbyJanuary28•  FormgroupsandselectlectureasagroupbyJanuary28• ReadpapersforJanuary28lecture

•  AprotocolforpacketnetworkintercommunicaCon(CerfandKahn,1974)•  TheDesignPhilosophyoftheDARPAInternetProtocols(Clark,1988)

Outline

• HistoryofInternet

• Reviewofbasicconcepts

• Grandchallenges–fewexamples

Visions • VannevarBush,“Aswemaythink”(1945):memex

•  J.C.R.Licklider(1962):“GalacCcNetwork”

•  ConceptofaglobalnetworkofcomputersconnecCngpeoplewithdataandprograms

•  FirstheadofDARPAcomputerresearch,October1962 Bush

Licklider

Circuit switching

1920s

1967

[US Air Force]

InternaConaloperator,NewYorkAT&TSource:hNp://www.corp.aN.com/history/nethistory/switching.html

1935

1961-64: Packet switching

Circuit Switching Packet switching

Physical channel carrying stream of data from source to destination

Message broken into short packets, each handled separately

Three phase: setup, data transfer, tear-down

One operation: send packet

Data transfer involves no routingPackets stored (queued) in each router, forwarded to appropriate neighbor


• Keybenefit:StaCsCcalMulCplexing•  (whatelse?)

Circuit switching

Time

Packet switching:multiplexed

Time


• Concurrentdevelopmentatthreegroups•  LeonardKleinrock(MIT):queueing-theoreCcanalysisofpacketswitchinginPh.D.thesis(1961-63)demonstratedvalueofstaCsCcalmulCplexing

•  PaulBaran(RAND)•  DonaldDavies(NaConalPhysicalLaboratories,UK)

Kleinrock Baran Davies

Baran’s packet switching

Paul Baran, “On distributed communications networks”, Sept. 1962

There is an increasingly repeated statement made that one day we will require more capacity for data transmission than needed for voice. If this statement is correct, then it would appear prudent to broaden our planning consideration to include new concepts for future data network directions. ... New digital computer techniques using redundancy make cheap unreliable links potentially usable. ... Such a system should economically permit switching of very short blocks of data from a large number of users simultaneously with intermittent large volumes among a smaller set of points.

Baran’s packet switching

Paul Baran, “On distributed communications networks”, Sept. 1962

“

”

1965: First computer network

•  LawrenceRobertsandThomasMerrillconnectaTX-2atMITtoaQ-32inSantaMonica,CA

• ARPA-fundedproject• Connectedwithtelephoneline

•  works,butit’sinefficientandexpensive•  confirmedonemoCvaConforpacketswitching

• RobertsuClizedDavies’packetswitchingtheorylate1960’sandbuiltintoARPANET

Roberts

The ARPANET begins

• RobertsjoinsDARPA(1966),publishesplanfortheARPANETcomputernetwork(1967)

• December1968:Bolt,Beranek,andNewman(BBN)winbidtobuildpacketswitch,theInterfaceMessageProcessor(IMP)

•  FirstgeneraConofgateways•  September1969:BBNdeliversfirstIMPtoKleinrock’slabatUCLA An older Kleinrock

with the first IMP

ARPANET comes alive

Stanford Research Institute (SRI)

“LO”Oct 29, 1969

UCLA

ARPANET grows

•  Dec1970:ARPANETNetworkControlProtocol(NCP)

•  1971:Telnet,FTP•  1972:Email(RayTomlinson,BBN)

•  1979:USENET

ARPANET, April 1971

ARPANET grows

ARPANET to Internet

•  Meanwhile,othernetworkssuchasPRnet,SATNETdeveloped

•  May1973:VintonG.CerfandRobertE.KahnpresentfirstpaperoninterconnecCngnetworks

•  ConceptofconnecCngdiversenetworks,unreliabledatagrams,globaladdressing,...

•  BecameTCP/IP

Kahn

Cerf

TCP/IP deployment •  TCP/IPimplementedonmainframesbygroupsatStanford,BBN,UCL

•  DavidClarkguidesarchitecture,implementsitonXeroxAltoandIBMPC

•  1982:InternaConalOrganizaConforStandards(ISO)releasesOpenSystemsInterconnecCon(OSI)referencemodel

•  DesignbycommiNeedidn’twin•  January1,1983:“FlagDay”NCPtoTCP/IPtransiCononARPANET

Application

Presentation

Session

Transport

Network

Data Link

Physical

OSI ReferenceModel’s layers

Growth from Ethernet

•  Ethernet:R.MetcalfeandD.Boggs,July1976

•  SpanningTreeprotocol:RadiaPerlman,1985

•  Madelocalareanetworkingeasy

Metcalfe

Perlman

Growth spurs organic change •  Early1980s:Manynewnetworks:CSNET,BITNET,MFENet,SPAN(NASA),...

•  Nov1983:DNSdevelopedbyJonPostel,PaulMockapetris(USC/ISI),CraigPartridge(BBN)

•  1984:HierarchicalrouCng:EGPandIGP(latertobecomeeBGPandiBGP)

Postel

Partridge

Mockapetris

NSFNET •  1984:NSFNETforUShighereducaCon

•  Servemanyusers,notjustonefield

•  Encouragedevelopmentofprivateinfrastructure(e.g.,backbonerequiredtobeusedforResearchandEducaCon)

•  SCmulatedinvestmentincommerciallong-haulnetworks

•  1990:ARPANETends•  1995:NSFNETdecommissioned

NSFNET backbone, 1992

Explosive growth! In hosts

Explosive growth!

Inte

rnet

forw

ardi

ng t

able

siz

e

[Huston ’12]Year

(Colors correspondto measurements

from different vantage points)

In networks

Explosive growth! In applications

Morris Internet Worm (1988)World wide web (1989)

MOSAIC browser (1992)Search engines

VoiceRadio

Botnets

Streaming videoSocial networking

Peer-to-peer

The results of your class projects!

In devices & technologies

NATs, firewalls, DPI, ...Wireless everywhereMobile everywhere

Tiny devices (smart phones)

...Giant devices (data centers)

Link speeds 200,000x fasterO(100 million) times as many devices

Cloud computingMobile apps

Huge societal relevance

FridayJune 122009

SaturdayJune 13

SundayJune 14

Routing instabilities and outages in Iranian prefixesfollowing 2009 presidential election

Affe

cted

pre

fixes

[James Cowie,Renesys Corporation]

Huge societal relevance

Rea

chab

le p

refix

es

July - August 2011 [James Cowie,Renesys Corporation]

Reachability to Lybia

Top 30 invenZons of the last 30 years

16.MediaFileCompression17.Microfinance18.PhotovoltaicSolarEnergy19.LargeScaleWindTurbines20.SocialNetworkingviaInternet21.GraphicUserInterface(GUI)22.DigitalPhotography/Videography23.RFID24.GeneCcallyModifiedPlants25.Biofuels26.BarCodesandScanners27.ATMs28.Stents29.SRAM/FlashMemory30.AnC-RetroviralTreatmentforAIDS

1.  Internet/Broadband/WorldWideWeb

2.  PC/LaptopComputers

3. MobilePhones

4.  E-Mail

5.  DNATesCngandSequencing/HumanGenomeMapping

6. MagneCcResonanceImaging(MRI)

7. Microprocessors

8.  FiberOpCcs9.  OfficeSopware

10. Non-InvasiveLaser/RoboCcSurgery11. OpenSourceSopwareandServices12. LightEmiqngDiodes(LEDs)

13. LiquidCrystalDisplays(LCDs)14. GPS15. OnlineShopping/E-Commerce/AucCons

Compiled by the Wharton School @ U Penn, 2009

So we’re done! ... right?

•  CoreprotocolschangedliNle,butthecontexthas...•  CriminalsandmaliciousparCes•  Everyonetryingtogamethesystem•  Incrediblegrowth•  Constantmobility•  Extremecomplexity

•  ...andfixingthenetinvolvesfundamentalchallenges•  It’sdistributed•  Componentsfail•  Highlyheterogeneousenvironments•  HighlycomplexsystemscomponentsandinteracCons•  MustgetcompeCngparCestoworktogether

Review of Basics before Grand Challenges

Layering

• Akindofmodularity•  FuncConalityseparatedintolayers

•  Layernimplementshigher-levelfuncConalitybyinterfacingonlywithlayern-1

•  Hidescomplexityofsurroundinglayers:enablesgreaterdiversityandevoluConofmodules

Application

Transport

Network

Data Link

Physical

Application

Transport

Network

Data Link

Physical

Network

Data Link

Physical

Network

Data Link

Physical

Layering Application

Transport

Network

Data Link

Physical

Application

Transport

Network

Data Link

Physical

Network

Data Link

Physical

Network

Data Link

PhysicalTransport

Network

Data Link

Physical

Transport

Network

Data Link

Physical

Tunnel

Common funcZonality & problems

Framing, errors, addressing

Reliability, flow control, ordering, congestion, ...

Application

Transport

Network

Data Link

Physical Encoding of bits Physics

Packets on a ‘wire’

Packets across networksPackets across domains

Process-level communication

Anything you want...

Independent parties, scale, routing

Life, the universe, and everything

Addressing, heterogeneity, routing

Grand Challenges

Grand Challenges

• Widelyrecognizedasamongthemostimportantunsolvedproblemsinafield

•  Pvs.NP•  naturallanguageunderstanding•  bug-freeprograms•  movingsocietytocarbon-neutralenergy•  prevenCngcancer•  ...

Grand Challenges in networking?

• GeqnganAinthisclass?

GC’s in networking

•  “WhatI’mworkingon!”•  HighlevelobjecCves

•  Security&privacy•  Reliability•  Usability

•  DifferentthanPvs.NP:hardtoevendefine“security”;objecCvesinvolvetradeoffs

An Informal Survey

Unreliability: One Example

Internet RouZng

AS 7018AT&T

AS 36561YouTube

AS 698UIUC

eBGP

iBGP

AS11537Internet2

Border Gateway Protocol

BD

A

BD

C

routeselection

ACBD

CAD

Instability causes outages

BD X

C

•  Link state changes•  Router failures•  Config. changes•  ...

•  Loops•  Detection delay•  Black holes

AForwarding loop

FAIL

CACBDEventually, control message: Loop detected!


Internet

Destinationsite

Source sites

X

[F. Wang, Z. M. Mao, J. Wang, L. Gao, R. Bush SIGCOMM’06]


More outages Longer outages[F. Wang, Z. M. Mao, J. Wang, L. Gao, R. Bush SIGCOMM’06]

(...and higher latency, packet reordering,router CPU load during instability)

Failureinjected Outage length (sec)

Many sources of unreliability

• CongesCon•  noend-to-endbandwidthreservaConsintheInternet

• ConfiguraConorsopwarebugs•  Failuresordelays

•  innetwork,DNSservers,caches,applicaConservers,...

Insecurity: one example

Prefix hijacking

• AnyonecanadverCseroutesforanyIPprefix!• HowcanhijackergettheadverCsedroutestoactuallybeusedbyotherASes?

•  Announcemorespecific(longer)prefixthanrealowner•  Noweveryone’strafficis“blackholed”

• Canprotectagainstthis(SecureBGP),but...•  it’snotdeployedtoday•  andeventhen,cansCllcleverly(oraccidentally)aNracttrafficandeavesdrop

From hijacking to MITM • August’08,KapelaandPilosov• ManintheMiddle(MITM)aNack

•  TraffictoadesCnaConredirected(notblackholed)throughanaNacker

•  ANackercanwatcheverythingyoudowithoutyounoCcing

• What’sthekeyproblemhere?

How can attacker forward traffic to destination,if attacker is pretending to be the destination?

Hijacking + eavesdropping •  AfindslegiCmatepath

ABDfor128.2.0.0/16

D

CD

BD

CD

ECD

ECD

FCD

D

C

BA

F

E

BGP Announcements

128.2.0.0/16

Selected routes &flow of data

Hijacking + eavesdropping •  AfindslegiCmatepath

ABDfor128.2.0.0/16•  Asendssemi-bogus

announcementofpathABDfor128.2.0.0/17

•  Result:• ASes(hereB)onrealpathkeepusingrealpathbecauseofloopeliminaCon

• AllotherASesuseroutethroughA(/17beats/16)

•  AforwardstraffictoB

D

BCEA

BD

BD

EABD

ABD

EABD

D

C

BA

F

E

BGP Announcements

Selected routes &flow of data

128.2.0.0/16

Grand Challenges in networking

•  “WhatI’mworkingon!”•  NebuloushighlevelobjecCves

•  Security&privacy•  Reliability•  Usability

•  WhydoesnetworkinglackacrispGrandChallenge?•  Infrastructureneedstosupporthighlydiverseanddynamicgoals,applicaCons,andenvironments

An Informal Survey

•  Complexity

Grand Challenges in networking

Meta-challenge:

How do we make the Internetevolvable?

Advanced Computer Networks cs538, Spring 2016 @ UIUC

Documents

Transcript of Advanced Computer Networks cs538, Spring 2016 @ UIUC