1. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 1 Introduction to Telephony, Cable and Internet Technologies Based in part upon slides of S. Keshav (Ensim), J. Bellamy’s book, Prof. Raj Jain (OSU), L. Peterson (Princeton), J. Kurose (U Mass) http://www.pde.rpi.edu/ Or http://www.ecse.rpi.edu/Homepages/shivkuma/ Shivkumar Kalyanaraman Rensselaer Polytechnic Institute shivkuma@ecse.rpi.edu

2. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 2 q Connectivity: q direct (pt-pt, N-users), q indirect (switched, inter-networked) q Telephony, Internet, Cable Networks: Basic Concepts q Concepts: Topologies, Framing, Multiplexing, Flow/Error Control, Reliability, Multiple-access, Circuit/Packet- switching, Addressing/routing, Congestion control q Data link/MAC layer: SLIP, PPP, LAN technologies … q Interconnection Devices q S. Keshav book (Chapter 2), Opt Nets (Sec 11.1, 13.1, 13.2) Overview

3. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 3 Connectivity... q Building Blocks q links: coax cable, optical fiber... q nodes: general-purpose workstations... q Direct connectivity: q point-to-point q multiple access

4. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 4 Connectivity… (Continued) q Indirect Connectivity q switched networks => switches q inter-networks => routers

5. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 5 What is “Connectivity” ? q Direct or indirect access to every other node in the network q Connectivity is what you get instead of a direct physical link q Key Tradeoff: Performance characteristics worse!

6. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 6 Connectivity … q Internet: q Best-effort (no performance guarantees) q Packet-by-packet q A pt-pt link: q Always-connected q Fixed bandwidth q Fixed delay q Zero-jitter

7. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 7 Telephony

8. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 8 Telephone Network: What is It? q Specialized to carry voice traffic q Aggregates like T1, SONET OC-N can also carry data q Also carries q Telemetry, video, fax, modem calls q Internally, uses digital samples q Switches and switch controllers are special purpose computers Pieces: 1. End systems 2. Transmission 3. Switching 4. Signaling

9. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 9 Telephone Network: What is It? q Single basic service: two-way voice q low end-to-end delay q guarantee that an accepted call will run to completion q Endpoints connected by a circuit, like an electrical circuit q Signals flow both ways (full duplex) q Associated with reserved bandwidth and buffer resources

10. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 10 Telephone Network Design q Fully connected core q simple routing q telephone number is a hint about how to route a call q But not for 800/888/700/900 numbers: these are pointers to a directory that translates them into regular numbers q hierarchically allocated telephone number space

11. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 11 Telephone Network Design

12. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 12 Telephone Pieces: End Systems

13. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 13 Telephone Pieces: End Systems q Transducers: key to carrying voice on wires q Dialer q Ringer q Switch-hook

14. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 14 Last-Mile Transmission Environment q Wire gauges:19, 22, 24, 26 gauge(smaller better) q Diameters: 0.8, 0.6, 0.5, 0.4 mm (larger better) q Various forms of noise: (twisting reduces noise) q Bridged-tap noise: bit-energy diverted to extension phone sockets q Crosstalk q Ham radio q AM broadcast q Insertion loss: -140 dBm noise floor q 100 million times more sensitive than normal modems q Bandwidth range = 600 kHz q Notch effects in insertion loss due to bridged-taps q Transmission PSD = -40dBm => 90 dBm budget

15. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 15 2-wire vs 4-wire: Sidetones and Echoes q Both trans & reception circuits need two wires q 4 wires from every central office to home q Alternative: Use same pair of wires for both transmission and reception q Signal from transmission flows to receiver: sidetone  Reverse Effect: received signal at end-system bounces back to CO (esp if delay > 20 ms): echo  Solutions: balance circuit (attenuate side-tone) + echo- cancellation circuit (cancel echoes).

16. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 16 Dialing q Pulse q sends a pulse per digit q collected by central office (CO) q Interpreted by CO switching system to place call or activate special features (eg: call forwarding, prepaid- calls etc) q Tone q key press (feep) sends a pair of tones = digit q also called Dual Tone Multifrequency (DTMF) q CO supplies the power for ringing the bell. q Standardized interface between CO and end-system => digital handsets, cordless/cellular phones

17. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 17 Telephone Pieces: Transmission Muxing q Trunks between central offices carry hundreds of conversations q Can’t run thick bundles! Instead, send many calls on the same wire q Multiplexing (a.ka. Sharing) q Analog multiplexing q Band-limit call to 3.4 KHz and frequency shift onto higher bandwidth trunk q obsolete q Digital multiplexing q first convert voice to samples q 1 sample = 8 bits of voice q 8000 samples/sec => call = 64 Kbps

18. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 18 Transmission Multiplexing (contd) q How to choose a sample? q 256 quantization levels, logarithmically spaced (why?) q sample value = amplitude of nearest quantization level q Two choices of levels (  law and A law) q Time division multiplexing q Trunk carries bits at a faster bit rate than inputs q n input streams, each with a 1-byte buffer q Output interleaves samples q Need to serve all inputs in the time it takes one sample to arrive => output runs n times faster than input q Overhead bits mark end of frame (why?)

19. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 19 Transmission Multiplexing q Multiplexed trunks can be multiplexed further q Need a standard! (why?) q US/Japan standard is called Digital Signaling hierarchy (DS)

20. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 20 Telephone Pieces: Switching

21. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 21 Telephone Pieces: Switching q Problem: q each user can potentially call any other user q can’t have (a billion) direct lines! q Switches establish temporary circuits q Switching systems come in two parts: switch and switch controller

22. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 22 Switching System Components

23. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 23 Switch: What does it do? q Transfers data from an input to an output q many ports (up to 200,000 simultaneous calls) q need high speeds q Some ways to switch: q 1. space division switching: eg: crossbar q if inputs (or crosspoints) are multiplexed, need a schedule (why?)

24. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 24 Crossbar Switching Elements

25. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 25 Switching (Contd) q Another way to switch q time division (time slot interchange or TSI) q also needs a service schedule (why?) n To build larger switches we combine space and time division switching elements

26. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 26 Telephone pieces: Signaling q A switching system has a switch and a switch controller q Switch controller is in the control plane q does not touch voice samples q Manages the network q call routing (collect dialstring and forward call) q alarms (ring bell at receiver) q billing q directory lookup (for 800/888 calls)

27. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 27 Signaling q Switch controllers are special purpose computers q Linked by their own internal computer network q Common Channel Interoffice Signaling (CCIS) network q Earlier design used in-band tones, but was hacked q Also was very rigid (why?) q Messages on CCIS conform to Signaling System 7 (SS7)

28. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 28 Signaling (contd) q One of the main jobs of switch controller: keep track of state of every endpoint q Key is state transition diagram

29. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 29 Telephony Routing of Signaled Calls q Circuit-setup (I.e. the signaling call) is what is routed. q Voice then follows route, and claims reserved resources.  3-level hierarchy, with a fully-connected core  AT&T: 135 core switches with nearly 5 million circuits  LECs may connect to multiple cores

30. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 30 Telephony Routing algorithm  If endpoints are within same CO, directly connect  If call is between COs in same LEC, use one-hop path between COs  Otherwise send call to one of the cores  Only major decision is at toll switch  one-hop or two-hop path to the destination toll switch.  Essence of telephony routing problem: which two-hop path to use if one-hop path is full (almost a static routing problem… )

31. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 31 Features of telephone routing  Resource reservation aspects:  Resource reservation is coupled with path reservation  Connections need resources (same 64kbps)  Signaling to reserve resources and the path  Stable load  Network built for voice only.  Can predict pairwise load throughout the day  Can choose optimal routes in advance  Technology and economic aspects:  Extremely reliable switches  Why? End-systems (phones) dumb because computation was non-existent in early 1900s.  Downtime is less than a few minutes per year => topology does not change dynamically

32. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 32 Features of telephone routing  Source can learn topology and compute route  Can assume that a chosen route is available as the signaling proceeds through the network  Component reliability drove system reliability and hence acceptance of service by customers  Simplified topology:  Very highly connected network  Hierarchy + full mesh at each level: simple routing  High cost to achieve this degree of connectivity  Organizational aspects:  Single organization controls entire core  Afford the scale economics to build expensive network  Collect global statistics and implement global changes => Source-based, signaled, simple alternate-path routing

33. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 33 Telecommunications Regulation History q FCC regulations cover telephony, cable, broadcast TV, wireless etc q “Common Carrier”: provider offers conduit for a fee and does not control the content q Customer controls content/destination of transmission & assumes criminal/civil responsibility for content q Local monopolies formed by AT&T’s acquisition of independent telephone companies in early 20 th century q Regulation forced because they were deemed natural monopolies (only one player possible in market due to enormous sunk cost) q FCC regulates interstate calls and state commissions regulate intra-state and local calls q Bells + 1000 independents interconnected & expanded q FCC rulemaking process: q Intent to act, solicitation of public comment etc…

34. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 34 Deregulation of telephony q 1960s-70s: gradual de-regulation of AT&T due to technological advances q Terminal equipment could be owned by customers (CPE) => explosion in PBXs, fax machines, handsets q Modified final judgement (MFJ): breakup of AT&T into ILECs (incumbent local exchange carrier) and IXC (inter-exchange carrier) part q Long-distance opened to competition, only the local part regulated… q Equal access for IXCs to the ILEC network q 1+ long-distance number introduced then… q 800-number portability: switching IXCs => retain 800 number q 1995: removed price controls on AT&T

35. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 35 Telecom Act of 1996 q Required ILECs to open their markets through unbundling of network elements (UNE-P), facilities ownership of CLECs…. q Today UNE-P is one of the most profitable for AT&T and other long-distance players in the local market: due to apparently below-cost regulated prices… q ILECs could compete in long-distance after demonstrating opening of markets q Only now some ILECs are aggressively entering long distance markets q CLECs failed due to a variety of reasons… q But long-distance prices have dropped precipitously (AT&T’s customer unit revenue in 2002 was $11.3 B compared to 1999 rev of $23B) q ILECs still retain over 90% of local market q Wireless substitution has caused ILECs to develop wireless business units

36. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 36 US Telephone Network Structure (after 1984)

37. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 37 Exchange Area Network

38. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 38 Cable TV Networks

39. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 39 Cable Technology q Coaxial cable RF distribution networks. q Attributes: q Broadcast, low-band reverse channels q Mainly one-way video channels q Reasonably secure network (private conduit to home) q Free from free-space interferences q Good signal capacity (over 1 GHz) and flexibility q Multiple signaling channels q Significant attenuation that increases proportional to frequency => (active) RF amplification (every 1000 ft) q Freq responses of deployed amps and filters limit practical usage of frequencies > 1 GHz

40. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 40 Cable Building Blocks

41. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 41 Cable Spectrum: Upto 750 Mhz

42. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 42 Cable Technology & Architecture q Head-end: signal processing center q Each carrier: Baseband analog or digital modulation q Carriers multiplexed w/ freq-selective diplex filters: q allows simultaneous info transfer in both directions q Tree-and-branch architecture: q Well-suited for one-way broadcast video transmission (same signals to every customer) q Accumulates noise & distortions (amplifiers) q Affects plant reliability and received signal quality q Limits on the number of amplifiers cascaded q Limits on bandwidth in operation (few 100s of MHz): below cable potential… q Makes delivery of “switched” services (separate stream for each customer) difficult

43. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 43 Tree-and-Branch Architecture

44. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 44 Fiber Optics For Cable Networks q Key: Leave the laser ON and intensity-modulate with the analog signal q Such analog modulated lasers are very different from their digital counterparts q Low internal noise and high linearity in the range q Receiver: simple photo-detector -> back to RF spectrum q Result: Hybrid fiber-coax infrastructure, with fiber closer to headend q Coax plant serves smaller range (segmentation), but overall HFC reach dramatically increased q Also, it allows the economical support of remote, smaller clusters of homes q Each part could also provide different services to area (micro- market segmentation) q Assign different portions of HFC spectrum to diff uses: many virtual networks: sustained investments possible

45. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 45 Hybrid Fiber Coax (HFC) Networks

46. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 46 Multiple Services over HFC

47. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 47 Future Potential of HFC Broadband q Due to smaller loops, the region from 900MHz – 1 GHz can be used for data. q Reduced noise in this region => increased bit rate (200 Mbps) per segment… q Future: fiber moves closer, smaller coax- segments, reduced homes per coax run (60 homes), use of frequencies above 1 Ghz using new electronics q Latest DOCSIS 2.0 spec: 256 QAM (=> 8 bits/Hz) or S-CDMA on cable for more robust transmissions

48. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 48 Cable Regulation q Very different from telephony: not common-carrier q Able to control content AND the conduit! q Grew by providing an alternative (and extension) to broadcast TV and had initial growth troubles q Did not have to offer service on a non-discriminatory basis (unlike common carriers) q Asserted first-amendment rights to maintain control over content q Not required to provide access to their distribution system to other providers (some portion of capacity required to be offered to unaffiliated players: eg: CNN) q But they reserve rights to appropriately bundle these channels q Limited regulation: basic tier is rate-regulated by local authorities till 1999 based upon FCC rules

49. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 49 Cable regulation (contd) q Cable networks limited in horizontal expansion, and from vertically integrating w/ CNN etc q Note: ILECs like Bell Atlantic in contrast merged with IXCs like GTE q AT&T’s cable acquisitions were interesting (and will be explored later…) q Cable service is multi-faceted and varied from area to area => regulation formulation more complicated q Over-builders (satellite providers) got access to independent content providers: otherwise regulation achieved little for cable q Local authorities get revenue from cable regulation q HFC dominates franchise regulation talks, but cable providers are not obligated to provide broadband access..

50. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 50 Data Networking and the Internet

51. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 51 Recall: Indirect Connectivity… q Indirect Connectivity q switched networks => switches q inter-networks => routers

52. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 52 Inter-Networks: Networks of Networks = Internet … … …… The internet is just a big switch providing indirect connectivity

53. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 53 Recall: Connecting N users: Directly… q Pt-pt: connects only two users directly… q How to connect N users directly ? q What are the costs of each option? q Does this method of connectivity scale ? AB... Full mesh Bus

54. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 54 Point-to-Point Connectivity Issues q Physical layer: coding, modulation etc q Link layer needed if the link is shared bet’n apps; is unreliable; and is used sporadically q No need for protocol concepts like addressing, names, routers, hubs, forwarding, filtering … AB

55. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 55 Link Layer: Serial IP (SLIP) q Simple: only framing = Flags + byte-stuffing q Compressed headers (CSLIP) for efficiency on low speed links for interactive traffic. q Problems: q Need other end’s IP address a priori (can’t dynamically assign IP addresses) q No “type” field => no multi-protocol encapsulation q No checksum => all errors detected/corrected by higher layer. q RFCs: 1055, 1144

56. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 56 Link Layer: PPP q Point-to-point protocol q Frame format similar to HDLC q Multi-protocol encapsulation, CRC, dynamic address allocation possible q key fields: flags, protocol, CRC q Asynchronous and synchronous communications possible q Link and Network Control Protocols (LCP, NCP) for flexible control & peer-peer negotiation q Can be mapped onto low speed (9.6Kbps) and high speed channels (SONET)

57. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 57 Connecting N users: Directly... q Bus: Low cost vs broadcast/collisions, MAC complexity q Full mesh: High cost vs simplicity q New concept: q Address to identify nodes. q Needed if we want the receiver alone to consume the packet!... Full mesh Bus q Problem: Direct connectivity does not “scale”….

58. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 58 How to build Scalable Networks? q Scaling: system allows the increase of a key parameter. Eg: let N increase… q Inefficiency limits scaling … q Direct connectivity is inefficient & hence does not scale q Mesh: inefficient in terms of # of links q Bus architecture: 1 expensive link, N cheap links. Inefficient in bandwidth use

59. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 59 Filtering, forwarding … q Filtering: choose a subset of elements from a set q Don’t let information go where its not supposed to… q Filtering => More efficient => more scalable Filtering is the key to efficiency & scaling q Forwarding: actually sending packets to a filtered subset of link/node(s) q Packet sent to one link/node => efficient q Solution: Build nodes which focus on filtering/forwarding and achieve indirect connectivity “switches” & “routers”

60. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 60 Connecting N users: Indirectly q Star: One-hop path to any node, reliability, forwarding function q “Switch” S can filter and forward! q Switch may forward multiple pkts in parallel for additional efficiency! Star S

61. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 61 Connecting N users: Indirectly … q Ring: Reliability to link failure, near-minimal links q All nodes need “forwarding” and “filtering” q Sophistication of forward/filter lesser than switch Ring

62. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 62 Ring Star S Tree Topologies: Indirect Connectivity

63. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 63 Protocol Issues in Data Networks q Pt-Pt connectivity: q Framing q Error control/Reliability q Flow control & Windowing protocols q Multiplexing, Virtualization q Circuit vs Packet Switching: a muxing view q MAC arbitration schemes: q Random access/CSMA, TDMA, CDMA q Interconnection components: repeater, hub, bridge, switch, router

64. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 64 Reliability: Types of errors & effects q Forward channel bit-errors (garbled packets) q Forward channel packet-errors (lost packets) q Reverse channel bit-errors (garbled status reports) q Reverse channel bit-errors (lost status reports) q Protocol-induced effects: q Duplicate packets q Duplicate status reports q Out-of-order packets q Out-of-order status reports q Out-of-range packets/status reports (in window-based transmissions)

65. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 65 Temporal Redundancy Model Packets Sequence Numbers CRC or Checksum Status Reports ACKs NAKs, SACKs Bitmaps Packets FEC information Retransmissions Timeout

66. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 66 Reliability Mechanisms q Mechanisms: q Checksum: detects corruption in pkts & acks q ACK: “packet correctly received” q Duplicate ACK: “packet incorrectly received” q Sequence number: identifies packet or ack q 1-bit sequence number used both in forward & reverse channel q Timeout only at sender q Reliability capabilities achieved: q An error-free channel q A forward & reverse channel with bit-errors q Detects duplicates of packets/acks q NAKs eliminated q A forward & reverse channel with packet-errors (loss)

67. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 67 Stop and Wait Flow Control Data Ack Data t frame t prop  = t prop t frame = Distance/Speed of Signal Frame size /Bit rate = Distance  Bit rate Frame size  Speed of Signal = 1 2  + 1 U= 2t prop +t frame t frame U  Light in vacuum = 300 m/  s Light in fiber = 200 m/  s Electricity = 250 m/  s

68. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 68 Sliding Window Protocols Data Ack t frame t prop U= Nt frame 2t prop +t frame = N 2  +1 1 if N>2  +1

69. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 69 Multiplexing: The Method of Sharing Costly Resources q Multiplexing = sharing q Allows system to achieve “economies of scale” q Cost: waiting time (delay), buffer space & loss q Gain: Money ($$) => Overall system costs less Full MeshBus

70. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 70 Virtualization q The multiplexed shared resource with a level of indirection will seem like a unshared virtual resource! q I.e. Multiplexing + indirection = virtualization q We can “refer” to the virtual resource as if it were the physical resource. q Eg: virtual memory, virtual circuits… q Connectivity: a virtualization created by the Internet! q Indirection requires binding and unbinding… q Eg: use of packets, slots, tokens etc... Physical Bus = AB AB Virtual Pt-Pt Link

71. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 71 Statistical Multiplexing q Reduce resource requirements (eg: bus capacity) by exploiting statistical knowledge of the system. q Eg: average rate <= service rate <= peak rate q If service rate < average rate, then system becomes unstable!! q First design to ensure system stability!! q Then, for a stable multiplexed system: q Gain = peak rate/service rate. q Cost: buffering, queuing delays, losses. q Useful only if peak rate differs significantly from average rate. q Eg: if traffic is smooth, fixed rate, no need to play games with capacity sizing…

72. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 72 Stability of a Multiplexed System Average Input Rate > Average Output Rate => system is unstable! How to ensure stability ? 1.Reserve enough capacity so that demand is less than reserved capacity 2.Dynamically detect overload and adapt either the demand or capacity to resolve overload

73. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 73 What’s a performance tradeoff ? q R=link bandwidth (bps) q L=packet length (bits) q a=average packet arrival rate Traffic intensity = La/R A situation where you cannot get something for nothing! Also known as a zero-sum game.

74. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 74 What’s a performance tradeoff ? q La/R ~ 0: average queuing delay small q La/R -> 1: delays become large q La/R > 1: average delay infinite (service degrades unboundedly => instability)! Summary: Multiplexing using bus topologies has both direct resource costs and intangible costs like potential instability, buffer/queuing delay.

75. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 75 How to design large inter-networks? Circuit-Switching q Divide link bandwidth into “pieces” q Reserve pieces on successive links and tie them together to form a “circuit” q Map traffic into the reserved circuits q Resources wasted if unused: expensive. – Mapping can be done without “headers”. – Everything inferred from timing.

76. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 76 How to design large inter-networks? Packet-Switching q Chop up data (not links!) into “packets” q Packets: data + meta- data (header) q “Switch” packets at intermediate nodes q Store-and-forward if bandwidth is not immediately available. Bandwidth division into “pieces” Dedicated allocation Resource reservation

77. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 77 Packet Switching A B C 10 Mbs Ethernet 1.5 Mbs 45 Mbs D E statistical multiplexing queue of packets waiting for output link  Cost: self-descriptive header per-packet, buffering and delays due to statistical multiplexing at switches.  Need to either reserve resources or dynamically detect and adapt to overload for stability

78. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 78 Spatial vs Temporal Multiplexing q Spatial multiplexing: Chop up resource into chunks. Eg: bandwidth, cake, circuits… q Temporal multiplexing: resource is shared over time, I.e. queue up jobs and provide access to resource over time. Eg: FIFO queueing, packet switching q Packet switching is designed to exploit both spatial & temporal multiplexing gains, provided performance tradeoffs are acceptable to applications. q Packet switching is potentially more efficient => potentially more scalable than circuit switching !

79. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 79 Protocol Issues in Data Networks (Contd) q Pt-Pt connectivity: q Framing q Error control/Reliability q Flow control & Windowing protocols q Multiplexing, Virtualization q Circuit vs Packet Switching: a muxing view q MAC arbitration schemes: q Random access/CSMA, TDMA, CDMA q Interconnection components: repeater, hub, bridge, switch, router

80. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 80 Multi-Access LANs q Hybrid topologies: q Uses directly connected topologies (eg: bus), or q Indirectly connected with simple filtering components (switches, hubs). q Limited scalability due to limited filtering q Medium Access Protocols: q ALOHA, CSMA/CD (Ethernet), Token Ring … q Key: Use a single protocol in network q Concepts: address, forwarding (and forwarding table), bridge, switch, hub, token, medium access control (MAC) protocols

81. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 81 MAC Protocols: a taxonomy Three broad classes: q Channel Partitioning q divide channel into smaller “pieces” (time slots, frequency) q allocate piece to node for exclusive use q “Taking turns”: Token-based q tightly coordinate shared access to avoid collisions q Random Access q allow collisions q “recover” from collisions Goal: efficient, fair, simple, decentralized

82. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 82 Channel Partitioning MAC protocols. Eg: TDMA TDMA: time division multiple access q Access to channel in "rounds" q Each station gets fixed length slot (length = pkt trans time) in each round q Unused slots go idle q Example: 6-station LAN, 1,3,4 have pkt, slots 2,5,6 idle

83. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 83 “Taking Turns” MAC protocols - 1 Channel partitioning MAC protocols: q share channel efficiently at high load q inefficient at low load: delay in channel access, 1/N bandwidth allocated even if only 1 active node! Random access MAC protocols q efficient at low load: single node can fully utilize channel q high load: collision overhead “Taking turns” protocols look for best of both worlds!

84. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 84 Polling: q Master node “invites” slave nodes to transmit in turn q Request to Send, Clear to Send messages q Concerns: q polling overhead q latency q single point of failure (master) Token passing: q Control token passed from one node to next sequentially. q Token message q Concerns: q token overhead q latency q single point of failure (token) “Taking Turns” MAC protocols - 2

85. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 85 “Taking Turns” Protocols –3 Reservation-based a.k.a Distributed Polling: q Time divided into slots q Begins with N short reservation slots q reservation slot time equal to channel end-end propagation delay q station with message to send posts reservation q reservation seen by all stations q After reservation slots, message transmissions ordered by known priority

86. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 86 Random Access Protocols q Aloha at University of Hawaii: Transmit whenever you like Worst case utilization = 1/(2e) =18% q CSMA: Carrier Sense Multiple Access Listen before you transmit q CSMA/CD: CSMA with Collision Detection Listen while transmitting. Stop if you hear someone else. q Ethernet uses CSMA/CD. Standardized by IEEE 802.3 committee.

87. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 87 10Base5 Ethernet Cabling Rules q Thick coax q Length of the cable is limited to 2.5 km, no more than 4 repeaters between stations  No more than 500 m per segment  “10Base5” 2.5m 500 m Repeater Terminator Transceiver

88. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 88 10Base5 Cabling Rules (Continued) q No more than 2.5 m between stations q Transceiver cable limited to 50 m 2.5m 500 m Repeater Terminator Transceiver

89. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 89 Inter-connection Devices q Repeater: Layer 1 (PHY) device that restores data and collision signals: a digital amplifier q Hub: Multi-port repeater + fault detection q Note: broadcast at layer 1 q Bridge: Layer 2 (Data link) device connecting two or more collision domains. q Key: a bridge attempts to filter packets and forward them from one collision domain to the other. q It snoops on passing packets and learns the interface where different hosts are situated, and builds a L2 forwarding table q MAC multicasts propagated throughout “extended LAN.” q Note: Limited filtering intelligence and forwarding capabilities at layer 2

90. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 90 Interconnection Devices (Continued) q Router: Network layer device. IP, IPX, AppleTalk. Interconnects broadcast domains. q Does not propagate MAC multicasts. q Switch: q Key: has a switch fabric that allows parallel forwarding paths q Layer 2 switch: Multi-port bridge w/ fabric q Layer 3 switch: Router w/ fabric and per-port ASICs These are functions. Packaging varies.

91. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 91 Interconnection Devices H H B H H Router Extended LAN =Broadcast domain LAN= Collision Domain Network Datalink Physical Transport Router Bridge/Switch Repeater/Hub Gateway Application Network Datalink Physical Transport Application

92. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 92 Ethernet (IEEE 802) Address Format q 48-bit flat address => no hierarchy to help forwarding q Hierarchy only for administrative/allocation purposes q Assumes that all destinations are (logically) directly connected. q Address structure does not explicitly acknowledge indirect connectivity q => Sophisticated filtering cannot be done! 10111101 G/L bit (Global/Local) G/I bit (Group/Individual) OUI (Organizationally Unique ID)

93. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 93 Ethernet (IEEE 802) Address Format q G/L bit: administrative q Global: unique worldwide; assigned by IEEE q Local: Software assigned q G/I: bit: multicast q I: unicast address q G: multicast address. Eg: “To all bridges on this LAN” 10111101 G/L bit (Global/Local) G/I bit (Group/Individual) OUI (Organizationally Unique ID)

94. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 94 Ethernet & 802.3 Frame Format q Ethernet q IEEE 802.3 Dest. Address Source Address Type 662 Size in bytes Dest. Address Source Address Length Info 662 IPIPXAppleTalk LLC IPIPXAppleTalk CRC 4 4 Pad Length Info Maximum Transmission Unit (MTU) = 1518 bytes Minimum = 64 bytes (due to CSMA/CD issues)

95. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 95 Network/Transport Layer Issues q Inter-networking: heterogeneity, scale q Routing q Congestion control q Quality of Service (QoS)

96. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 96 Inter-Networks: Networks of Networks q What is it ? q “Connect many disparate physical networks and make them function as a coordinated unit … ” - Douglas Comer q Many => scale q Disparate => heterogeneity q Result: Universal connectivity! q The inter-network looks like one large switch, q User interface is sub-network independent

97. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 97 Inter-Networks: Networks of Networks q Internetworking involves two fundamental problems: heterogeneity and scale q Concepts: q Translation, overlays, address & name resolution, fragmentation: to handle heterogeneity q Hierarchical addressing, routing, naming, address allocation, congestion control: to handle scaling q Two broad approaches: circuit-switched and packet- switched

98. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 98 Scalable Forwarding, Structured Addresses q Address has structure which aids the forwarding process. q Address assignment is done such that nodes which can be reached without resorting to L3 forwarding have the same prefix (network ID) q A simple comparison of network ID of destination and current network (broadcast domain) identifies whether the destination is “directly” connected q I.e. Reachable through L2 forwarding only q Within L3 forwarding, further structure can aid hierarchical organization of routing domains (because routing algorithms have other scalability issues) Network IDHost ID Demarcator

99. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 99 Flat vs Structured Addresses q Flat addresses: no structure in them to facilitate scalable routing q Eg: IEEE 802 LAN addresses q Hierarchical addresses: q Network part (prefix) and host part q Helps identify direct or indirectly connected nodes

100. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 100 Internet Routing Drivers q Technology and economic aspects: q Internet built out of cheap, unreliable components as an overlay on top of leased telephone infrastructure for WAN transport. q Cheaper components => fail more often => topology changes often => needs dynamic routing q Components (including end-systems) had computation capabilities. q Distributed algorithms can be implemented q Cheap overlaid inter-networks => several entities could afford to leverage their existing (heterogeneous) LANs and leased lines to build inter-networks. q Led to multiple administrative “clouds” which needed to inter-connect for global communication.

101. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 101 Internet Routing Model q 2 key features: q Dynamic routing q Intra- and Inter-AS routing, AS = locus of admin control q Internet organized as “autonomous systems” (AS). q AS is internally connected q Interior Gateway Protocols (IGPs) within AS. q Eg: RIP, OSPF, HELLO q Exterior Gateway Protocols (EGPs) for AS to AS routing. q Eg: EGP, BGP-4

102. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 102 Intra-AS and Inter-AS routing inter-AS, intra-AS routing in gateway A.c network layer link layer physical layer a b b a a C A B d Gateways: perform inter-AS routing amongst themselves perform intra-AS routers with other routers in their AS A.c A.a C.b B.a c b c

103. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 103 Intra-AS and Inter-AS routing: Example Host h2 a b b a a C A B d c A.a A.c C.b B.a c b Host h1 Intra-AS routing within AS A Inter-AS routing between A and B Intra-AS routing within AS B

104. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 104 Requirements for Intra-AS Routing q Should scale for the size of an AS. q Low end: 10s of routers (small enterprise) q High end: 1000s of routers (large ISP) q Different requirements on routing convergence after topology changes q Low end: can tolerate some connectivity disruptions q High end: fast convergence essential to business (making money on transport) q Operational/Admin/Management (OAM) Complexity q Low end: simple, self-configuring q High end: Self-configuring, but operator hooks for control q Traffic engineering capabilities: high end only

105. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 105 Requirements for Inter-AS Routing q Should scale for the size of the global Internet. q Focus on reachability, not optimality q Use address aggregation techniques to minimize core routing table sizes and associated control traffic q At the same time, it should allow flexibility in topological structure (eg: don’t restrict to trees etc) q Allow policy-based routing between autonomous systems q Policy refers to arbitrary preference among a menu of available options (based upon options’ attributes) q In the case of routing, options include advertised AS- level routes to address prefixes q Fully distributed routing (as opposed to a signaled approach) is the only possibility. q Extensible to meet the demands for newer policies.

106. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 106 i  ii q If information about i, and  is known in a central location where control of i or  can be effected with zero time delays, q the congestion problem is solved! q Unfortunately, we have incomplete info, require a distributed solution with time-varying time- delays The Congestion Problem 1 n Capacity Demand Problem: demand outstrips available capacity

107. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 107 Congestion: A Close-up View q knee – point after which q throughput increases very slowly q delay increases fast q cliff – point after which q throughput starts to decrease very fast to zero (congestion collapse) q delay approaches infinity q Note (in an M/M/1 queue) q delay = 1/(1 – utilization) Load Throughput Delay kneecliff congestion collapse packet loss

108. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 108 Congestion Control vs. Congestion Avoidance q Congestion control goal q stay left of cliff q Congestion avoidance goal q stay left of knee q Right of cliff: q Congestion collapse Load Throughput kneecliff congestion collapse

109. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 109 Goals of Congestion Control q To guarantee stable operation of packet networks q Sub-goal: avoid congestion collapse q To keep networks working in an efficient status q Eg: high throughput, low loss, low delay, and high utilization q To provide fair allocations of network bandwidth among competing flows in steady state q For some value of “fair” 109

110. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 110 CC Techniques: Self-clocking PrPr PbPb ArAr AbAb Receiver Sender AsAs q Implications of ack-clocking: q More batching of acks => bursty traffic q Less batching leads to a large fraction of Internet traffic being just acks (overhead)

111. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 111 CC Techniques: Additive Increase/Multiplicative Decrease (AIMD) Policy q Assumption: decrease policy must (at minimum) reverse the load increase over-and-above efficiency line q Implication: decrease factor should be conservatively set to account for any congestion detection lags etc x0x0 x1x1 x2x2 Efficiency Line Fairness Line User 1’s Allocation x 1 User 2’s Allocation x 2

112. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 112 Quality of Service: What is it? Multimedia applications: network audio and video network provides application with level of performance needed for application to function. QoS

113. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 113 Fundamental QoS Problems q In a FIFO service discipline, the performance assigned to one flow is convoluted with the arrivals of packets from all other flows! q Cant get QoS with a “free-for-all” q Need to use new scheduling disciplines which provide “isolation” of performance from arrival rates of background traffic B Scheduling Discipline FIFO B

114. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 114 Fundamental QoS Problems q Conservation Law (Kleinrock):  (i)W q (i) = K q Irrespective of scheduling discipline chosen: q Average backlog (delay) is constant q Average bandwidth is constant q Zero-sum game => need to “set-aside” resources for premium services

115. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 115 QoS Big Picture: Control/Data Planes

116. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 116 Internet Regulation q FCC has largely had a hands-off policy q Early development of internet in part was influenced by high cost of telecom links q Packet switching developed as better multiplexing technology q Common-carriage regulation has affected Inet: q Eg: modems were like fax machine for the common carrier q Use of basic service (eg: telephony) to provide enhanced service (eg: internet access) => not subject to FCC or state jurisdiction q Led to community bulletin-boards, ISPs, value-added networks (frame-relay?)… q Home-to-ISP treated as local call (even if crossed state- boundaries) q ILECs prohibited from offering inter-LATA services q DSL viewed as basic service => must unbundle DSL to allow 3 rd parties to offer internet access over ILEC DSL

117. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 117 Summary: List of Internet Problems q Basics: Direct/indirect connectivity, topologies q Link layer issues: q Framing, Error control, Flow control q Multiple access & Ethernet: q Cabling, Pkt format, Switching, bridging vs routing q Internetworking problems: Naming, addressing, Resolution, fragmentation, congestion control, traffic management, Reliability, Network Management

118. Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 118 Additional Reading q Internet Design Philosophy: q Saltzer, Reed, Clark: "End-to-End arguments in System Design""End-to-End arguments in System Design" q Clark: "The Design Philosophy of the DARPA Internet Protocols":"The Design Philosophy of the DARPA Internet Protocols": q RFC 2775: Internet Transparency: In HTMLIn HTML