1 / 43

cs / ee 143 Communication Networks

cs / ee 143 Communication Networks. Steven Low CMS, EE, Caltech Fall 2012. Warning. These notes are not self-contained, probably not understandable, unless you also were in the lecture They are supplement to not replacement for class attendance. TV & home theatre. Finding your way.

raymond-booker
Télécharger la présentation

cs / ee 143 Communication Networks

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. cs/ee 143 Communication Networks Steven Low CMS, EE, Caltech Fall 2012

  2. Warning These notes are not self-contained, probably not understandable, unless you also were in the lecture They are supplement to not replacement for class attendance

  3. TV & home theatre Finding your way Music Mail Library at your finger tip Cloud computing Internet applications (2006) Telephony Social networking Games

  4. Networking: 143, 144, 145 • cs/ee 143 Communication Networks • How to send information from A to B • cs/ee 144 Ideas Behind the Web • How web technologies work • cs/ee 145 Projects in Networking • Advanced project – propose your own topic • Understanding both theory and practice of networks

  5. Networking: 143, 144, 145 • 143: transport layer and down • 144: transport layer and up • 145: project course • 143/144/145 forms a project sequence for CS major • Another RSRG sequence: cs141abc

  6. 143: Communication Networks

  7. 144: Ideas behind the weba.k.a. What makes google tick? • Unique course that you can’t • find at any other school • The course will be a series of importanttopics from the web. • For each, we will learn • The theory behind the topic • Hands-on design, system building

  8. Alumni Recommendation • The Caltech experience should give engineers skills in: • Communication • Writing scientific papers, proposals, plans • Speaking to science, government, business audiences • Working in teams • Project planning and management • Division of labor; leadership • Larger, global view of the world Source: Mani Chandy

  9. RSRG courses and research RSRG courses • Learn by doing • Emphasis on creativity, problem-solving, research • Integrating courses, SURF, internships, theses, research opportunities in multiyear programs • Close interaction with faculty in and out of class • Interactive classes: “The lecture is dead.” Source: Mani Chandy

  10. Course website http://courses.cms.caltech.edu/cs143/

  11. cs/ee 143 Communication NetworksChapter 1 The Internet Text: Walrand & Parakh, 2010 Steven Low CMS, EE, Caltech

  12. Hosts, routers, links • IP addresses, domain name • Packet, header/trailer

  13. Basic mechanisms Packet switching • No dedicated resources • Packets may follow different paths • Packets may be lost, error, out-of-order Addressing • Globally unique IP address • Domain name • Mapping from DN to IP may change dynamically (DNS)

  14. Basic mechanisms Routing • Destination based • Routing decision may adapt to network condition, e.g., failure Error & loss recovery • Retransmission of lost or erroneous pkt • Typically done end-to-end Flow & congestion control • End-to-end • Implicit feedback class project focus

  15. Router

  16. CIDR notation Example (IPv4) 18.128.33.0/24 = { 18.128.33.0 – 18.128.33.255 } = subnet mask 255.255.255.0 18.128.0.0/22 = { 18.128.00000000. 00000000 – 18.128.0000 0011. 11111111 } = { 18.128.0.0 – 18.128.3.255 } subnet mask: prefix = first 24 bits #addresses = 232-24 = 28 #addresses = 232-22 = 210 CIDR = classless inter-domain routing

  17. CIDR notation 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 1 1 1 0 0 1 1 1 1 0 0 0 0 0101 1010

  18. CIDR notation 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0.0.0.0/1 0 1 1.0.0.0/1 1 1 0 0 1 1 1 1 0 0 0 0 0101 1010

  19. CIDR notation 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 1 1.0.0.0/1 0 1 1 0.0.0.0/2 0.1.0.0/2 0 1 1 1 1 0 0 0 0 0101 1010 1.1.0.0/3 Address blocks may be nested ! e.g. 1.0.0.0/1 contains 1.1.0.0/3

  20. Forwarding decision Packet’s destination IP 18.128.33.5 Output link ? entry 1: 18.128.0010 0001. 00000000 entry 2: 18.128.0000 0000. 00000000 destination: 18.128.0010 0001. 0000 0101

  21. Forwarding decision 18.128.0.0/18 0 1 18.128.0.0/22 18.128.33.0/24 18.128.00000000.00000000/22 18.128.0010 0001.00000000/24 output link 2 output link 1

  22. Forwarding decision 18.128.0.0/18 0 1 18.128.0.0/22 18.128.33.0/24 Packet dest: 18.128.33.5 output link 2 output link 1

  23. Longest prefix matching Packet’s destination IP 18.128.33.5 Output link 1 entry 1: 18.128.0010 0001. 00000000 entry 2: 18.128.0000 0000. 00000000 destination: 18.128.0010 0001. 0000 0101 It matches more with the first address block

  24. Longest prefix matching Packet’s destination IP 18.128.4.5 Output link 2 entry 1: 18.128.00100001. 00000000 entry 2: 18.128.00000000. 00000000 destination: 18.128.00000000. 00000101 It matches more with the second address block

  25. Longest prefix matching Packet’s destination IP 192.168.51.15 Output link 2 entry 1: 192.168.0011 0100. 00000000 Even though it matches both entries 2 and 3, entry 2 has a longer prefix entry 2: 192.168.0011 0010. 00000000 entry 3: 192.168.0011 0000. 00000000 destination: 192.168.0011 0011. 0010 0101 entry 3 “contains” entry 2

  26. Longest prefix matching 192.168.48.0/21 0 1 192.168.48.0/22 192.168.52.0/22 1 192.168.50.0/23 Subnets in routing tables may not be disjoint! Route to most specific address block

  27. How to construct routing table How do we make routing decisions? (Dijkstra, Bellman-Ford, BGP, ... Ch 5 Routing)

  28. cs/ee 143 Communication NetworksChapter 2 Principles Text: Walrand & Parakh, 2010 Steven Low CMS, EE, Caltech

  29. Principles • Sharing • Metrics • Scalability • Layering • Application & technology independence • Application topology

  30. Sharing Packet switching, statistical multiplexing  Queueing analysis cf circuit switching in traditional telephone network

  31. Sharing

  32. Metrics Link rate (bps) • DSL e.g. 768kbps down, 256kbps up Link bandwidth (Hz) • Size of frequency band • FM radio, wireless spectra, ... • If link can transmit over [300Hz, 1MHz], its bandwidth is ~1MHz Link capacity (bps) • Maximum link rate possible

  33. Metrics • Link capacity example • SNR = 220 • W = 1MHz • C = 106 log2(1+ 220) bit/s = 20 Mbps link “rate” is often called link “bandwidth” in networking literature

  34. Metrics Throughput (bps) • Bit rate actually achieved • Example • MP3 file of 3MB transferred in 2 mins • Throughput = 3MB / 2mins = 200 kbps • Generally less than link rate because of transmission and protocol overheads throughput <= link rate <= link capacity networking, protocol design communication, coding theories information theory

  35. RTT 1 2 W 1 2 W data ACKs 1 2 W 1 2 W Window Flow Control • ~ W packets per RTT • Lost packet detected by missing ACK Source time Destination time

  36. Metrics • Delay and delay jitter (s) • Some applications care more about throughput, some about delay, some about delay jitter • M/M/1 queue • Little’s law

  37. Queueing system random arrival process with rate random service time with average • Little’s law • Verifies directly for M/M/1, but holds much moregenerally • Extremely useful because of its generality

  38. Queueing system random arrival process with rate random service time with average

  39. M/M/1 queue Poisson arrival process with rate Exponential service time with average

  40. Scalability Location-based addressing • Example: IP is location based • Ethernet is not • Advantages • Reduce routing table size • Simpler routing operation • Disadvantages • Makes mobility hard • Makes security hard

  41. Scalability Hierarchical • Two level • Autonomous system • Same administrative and/or economic domain • Shortest path routing: OSPF, IS-IS • Inter-domain • BGP Both simplify routing

  42. Scalability Best effort service • Simple TCP/IP layer End-to-end principle • Simple network, intelligent hosts • Stateless routers • Packet carries its own state Both simplify router  fast big inexpensive routers Latest (~decade) idea: software defined networking (SDN)

  43. Scalability Hierarchical naming • Easier on human • DNS translation Layering • Application & technology independence Application topology • Client/server • CDN • P2P • Cloud computing • Social networking

More Related