Advanced Computer Networks cs 538 fall 2011
Brighten Godfrey [email protected]
Fall 2011
slides ©2010-2011 by Brighten Godfrey except photographs (from Computer History Museum) and unless otherwise noted To d ay
Course Overview Internet History Your Future This course
• is instructed by Brighten Godfrey ([email protected], 3128 Siebel) • takes place Tue & Thu, 3:30 - 4:45 pm, in 1302 Siebel • comes with FREE office hours: currently, Tuesdays after class and by appointment • has a web site: http://www.cs.illinois.edu/~pbg/ courses/cs538fa11/ Your instructor
• Ph.D. from UC Berkeley, 2009 • Dissertation on improving resilience and performance of distributed systems by taking advantage of heterogeneity • Research interests: • Reliable, flexible, and efficient networked systems • Algorithms for and analysis of distributed systems Course goal
• Prepare ourselves to perform high-quality research advancing the field of networking Course components
• Networking literature • The classics • The challenges • The latest • Research project • How to read, criticize, and present research Major topics
• Core architecture • Classic Internet architecture • Congestion control • Forwarding • Routing • Naming • Making it work well • Reliability, scalability, selfishness, security • Domain-specific networks • Enterprise, data center, P2P, wireless Requirements & grading
• Project (45%) • Midterm presentation (10%) • Final paper (20%) • Final poster presentation (15%) • Assignments, quizzes (15%) • Paper reviews (15%) • Paper presentations (15%) • Class participation (10%) 1. Project
• Research project that could be developed into a conference submission • Work alone or in groups of two • Project topics • Explore your own ideas • Next lecture: some suggestions • Steps • Project proposal (3 weeks from now) • Midterm presentation • Final poster presentation and paper 2. Readings
• Generally one or two papers per lecture • Submit a review on the wiki by 11:59pm the night before we discuss the paper • For each paper, a review is • At least 2 comments • About one paragraph (longer is not better) • Don’t just repeat what we already read in the paper! 3. Topic presentations
• 20 minute presentation on one topic in the course • 10 minutes of “depth” on one paper, including key concepts, techniques, results, and your criticism • 10 minutes of “breadth” comparing to 2-3 other papers and the required reading • 20 minutes of discussion during/after • At least 2 days before it happens, meet with me to show me your presentation 4. Assignments
• 2-3 homeworks or quizzes during the semester 5. Class participation
• Comment, question, and interact! • Discuss on the course wiki To d ay
Course Overview Internet History Your Future Visions
• Vannevar Bush, “As we may think” (1945): memex • J. C. R. Licklider (1962): “Galactic Network” • Concept of a global network of computers connecting people with data and programs Bush • First head of DARPA computer research, Licklider October 1962 Circuit switching
1920s 1967 [Getty Images] [US Air Force] 1961-64: Packet switching
Datagram packet Circuit Switching switching Physical channel carrying stream of Message broken into short packets, data from source to destination each handled separately
Three phase: setup, data transfer, One operation: send packet tear-down
Packets stored (queued) in each Data transfer involves no routing router, forwarded to appropriate neighbor 1961-64: Packet switching
• Leonard Kleinrock: queueing-theoretic analysis of packet switching in MIT Ph.D. thesis (1961-63) demonstrated value of statistical multiplexing • Concurrent work from Paul Baran (RAND), Donald Davies (National Kleinrock Physical Labratories, UK)
Circuit switching Packet switching: multiplexed
Time Time Baran 1965: First computer network
• Lawrence Roberts and Thomas Merrill connect a TX-2 at MIT to a Q-32 in Santa Monica, CA • ARPA-funded project • Connected with telephone line – it works, but it’s inefficient and expensive – confirming motivation for packet switching
Roberts The ARPANET begins
• Roberts joins DARPA (1966), publishes plan for the ARPANET computer network (1967) • December 1968: Bolt, Beranek, and Newman (BBN) wins bid to build packet switch, the Interface Message Processor • September 1969: BBN delivers first IMP to Kleinrock’s lab at UCLA
An older Kleinrock with the first IMP ARPANET comes alive
Stanford Research Institute (SRI)
“LO” Oct 29, 1969
UCLA ARPANET grows
• Dec 1970: ARPANET Network Control Protocol (NCP) • 1971: Telnet, FTP • 1972: Email (Ray Tomlinson, BBN) • 1979: USENET
ARPANET, April 1971 ARPANET grows ARPANET to Internet
• Meanwhile, other networks such as PRnet, SATNET deveoped • May 1973: Vinton G. Cerf and Robert E. Kahn present first paper on interconnecting networks • Concept of connecting diverse networks, unreliable datagrams, global addressing, ... Cerf • Became TCP/IP
Kahn TCP/IP deployment
• TCP/IP implemented on mainframes by groups at Stanford, BBN, UCL • David Clark implements it on Xerox Application Alto and IBM PC Presentation • 1982: International Organization for Session Standards (ISO) releases Open Transport Systems Interconnection (OSI) Network reference model Data Link
• Design by committee didn’t win Physical • January 1, 1983: “Flag Day” NCP to OSI Reference TCP/IP transition on ARPANET Model’s layers Growth from Ethernet
• Ethernet: R. Metcalfe and D. Boggs, July 1976 • Spanning Tree protocol: Radia Perlman, 1985 • Made local area networking easy
Metcalfe
Perlman Growth spurs organic change
• Early 1980s: Many new networks: CSNET, BITNET, MFENet, SPAN (NASA), ... Mockapetris • Nov 1983: DNS developed by Jon Postel, Paul Mockapetris (USC/ISI), Craig Partridge (BBN) Postel • 1984: Hierarchical routing: EGP and IGP (later to become eBGP and iBGP)
Partridge NSFNET
• 1984: NSFNET for US higher education NSFNET backbone, 1992 • Serve many users, not just one field • Encourage development of private infrastructure (e.g., backbone required to be used for Research and Education) • Stimulated investment in commercial long-haul networks • 1990: ARPANET ends • 1995: NSFNET decommissioned Explosive growth!
In hosts Explosive growth!
In networks (Colors correspond to measurements from different vantage points) Internet forwarding table size
Year [Huston ’10] n g J g a a k p k o o a k r B r , e t e , T i a T j , S H i a n I G D i g 0E p , 10 V e C l S i E a , N 10 e T 1 d o iv u W o l A , E s K 0 I to 12 k R , S R U A E 0 T 3 ok 1 y o, JP A E 0 I 4 1
S N yd ne y A , A E U 0 5 [Huffaker, claffy,1 E Hyun, Luckie, C Lyu, CAIDA]
E O 0 IPv4 & IPv6 INTERNET TOPOLOGY MAP JANUARY 2009
B AS-level INTERNET GRAPH S
i a n n g J g a a k p N k o I o N a k , r I B r , y e , t e T i a a , i h b T j , S H i l G a n I e m D i g 90E o p 0E D , 10 80E E V e C B A l S i E a , N 10 , e T 1 70 i d o E b iv u W a o l A h s , E t K 20 I D o 1 R 60 k S E u , b R A U A E T 0 3 o 1 5 k 0 y E o, U J A R P , E w 0 I o L 4 c I s , 1 4 o v 0 i R E M v T N A a, Sy el r d T ka A ne n U A y A A , Z , A ev , U E i ia 0 r 5 K o 1 t I E 3 re F 0 P i, E k A sin G E el B C H a, ofi , PL F S saw U r E a E O S 0 W m, 6 4755(TATACOMM) ol 1 R h 38809(NXGNET) 2 ck T 0 o A R t , 39049304 (Hutchison) E S na Vien 7575(AARNET) I , DE O lin 8395(COMSTAR) Ber C O E 10026(Asia Netcom) o, N 0 Osl 7 P t, DE 1 kfur 1 Fran 2516(KDDI) 0 A 20485(JSC) E
E Brussels, BE
W Paris, FR / E 7714(TelstraClear) 3320(Deutsche Telekom) 0 8928(Interoute)
8 London, UK
1 0 702(MCI) 1299(TeliaNet) 1273(CW)
3356(Level 3) 5459(London IX)
3549(Global Crossing) Dublin, IE
W
0 701(UUNET) 12989(HWNG) Alge
7 1239(Sprint) s, PT
1
1
0 2914(NTT) W
2828(XOXO) 6677(ICENET)
3561(Savvis)
W
0 4320(Embratel)
6
1
2
0
7843(Road Runner) W S 577(Bell Canada)
, U u lul
ono
H W
0
5
1
3
0
W
W
0
4
1 4
0
W
S
O
R
U i
W o
0
3
1 d
5 T
0 e
W
J H a n
ei
W r A
0
2 o
A
1 B
6 ,
0 M
W
u B
C
S E e R
[Huffaker, claffy, N
I
U R n
O
W
0 o I
, 1 R
1
S 7 0 R
C s e W
E
T
s A M B
U H A A
o
S W O
, 0 0
1 8 o
0 i
J W W
o S r W 0 U 9 t s T e n g S t t , U C s a e a a o o , i x , S i U s w n h r A Hyun, Luckie, h D , n n o , i a i R r o e c n U n n , e t o a a s t g v C S
g o S h t u n o A P o , o e n , C
, D H U A U
Lyu, CAIDA] S S Explosive growth!
In complexity
Autonomous System BGP router
IP router Routing protocols LAN LAN eBGP, iBGP switch ethernet MPLS, CSPF,... segment OSPF, RIP, ... hub spanning tree + learning broadcast Explosive growth!
In devices & technologies In applications
65 million times as many devices Morris Internet Worm (1988) Link speeds 200,000x faster World wide web (1989) NATs and firewalls MOSAIC browser (1992) Wireless everywhere Search engines Mobile everywhere Peer-to-peer Tiny devices (smart phones) Voice Giant devices (data centers) Radio ... Botnets Social networking Streaming video The results of your class projects! Huge societal relevance
Routing instabilities and outages in Iranian prefixes following 2009 presidential election Affected prefixes
Friday Saturday Sunday June 12 June 13 June 14 [James Cowie, 2009 Renesys Corporation] Huge societal relevance
Routing instabilities and outages in Georgian prefixes following 2008 South Ossetia War Affected prefixes (%)
Fri, Aug 8, 2008 [Earl Zmijewski, Renesys Corporation] Huge societal relevance
Reachability to Lybia Reachable prefixes
July - August 2011 [James Cowie, Renesys Corporation] Top 30 inventions of the last 30 years
Compiled by the Wharton School @ U Penn, 2009
1. Internet/Broadband/World Wide 15. Online Shopping/E-Commerce/ Web Auctions 2. PC/Laptop Computers 16. Media File Compression 3. Mobile Phones 17. Microfinance 4. E-Mail 18. Photovoltaic Solar Energy 5. DNA Testing and Sequencing/ 19. Large Scale Wind Turbines Human Genome Mapping 20. Social Networking via Internet 6. Magnetic Resonance Imaging (MRI) 21. Graphic User Interface (GUI) 7. Microprocessors 22. Digital Photography/Videography 8. Fiber Optics 23. RFID 9. Office Software 24. Genetically Modified Plants 10. Non-Invasive Laser/Robotic Surgery 25. Biofuels 11. Open Source Software and Services 26. Bar Codes and Scanners 12. Light Emitting Diodes (LEDs) 27. ATMs 13. Liquid Crystal Displays (LCDs) 28. Stents 14. GPS 29. SRAM/Flash Memory 30. Anti-Retroviral Treatment for AIDS So we’re done! ... right?
• Core protocols changed little, but the context has... • Criminals and malicious parties • Everyone trying to game the system • Incredible growth • Constant mobility • Extreme complexity • ...and fixing the net involves fundamental challenges • It’s distributed • Components fail • Highly heterogeneous environments • Must get competing parties to work together To d ay
Course Overview Internet History Your Future Your (near-term) future
• Thursday Aug. 26: Grand Challenges in computer networking; project, project topic suggestions • Next week: Internet architecture technical overview