introduction to telephony, cable and internet …koushik/shivkuma-teaching/sp2003/bon...can’t run...
Post on 16-Jul-2020
3 Views
Preview:
TRANSCRIPT
Shivkumar KalyanaramanRensselaer 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 KalyanaramanRensselaer Polytechnic Institute
shivkuma@ecse.rpi.edu
Shivkumar KalyanaramanRensselaer Polytechnic Institute
2
Connectivity: direct (pt-pt, N-users), indirect (switched, inter-networked)
Telephony, Internet, Cable Networks: Basic ConceptsConcepts: Topologies, Framing, Multiplexing, Flow/Error Control, Reliability, Multiple-access, Circuit/Packet-switching, Addressing/routing, Congestion control Data link/MAC layer: SLIP, PPP, LAN technologies …Interconnection DevicesS. Keshav book (Chapter 2), Opt Nets (Sec 11.1, 13.1, 13.2)
Overview
Shivkumar KalyanaramanRensselaer Polytechnic Institute
3
Connectivity...Building Blocks
links: coax cable, optical fiber...nodes: general-purpose workstations...
Direct connectivity:point-to-point
multiple access
Shivkumar KalyanaramanRensselaer Polytechnic Institute
4
Connectivity… (Continued)
Indirect Connectivityswitched networks
=> switches
inter-networks
=> routers
Shivkumar KalyanaramanRensselaer Polytechnic Institute
5
What is “Connectivity” ?
Direct or indirect access to every other node in the network
Connectivity is what you get instead of a direct physical link
Key Tradeoff: Performance characteristics worse!
Shivkumar KalyanaramanRensselaer Polytechnic Institute
6
Connectivity …Internet:
Best-effort(no performance guarantees)Packet-by-packet
A pt-pt link: Always-connectedFixed bandwidthFixed delay Zero-jitter
Shivkumar KalyanaramanRensselaer Polytechnic Institute
7
Telephony
Shivkumar KalyanaramanRensselaer Polytechnic Institute
8
Telephone Network: What is It?Specialized to carry voice traffic
Aggregates like T1, SONET OC-N can also carry dataAlso carries
Telemetry, video, fax, modem callsInternally, uses digital samplesSwitches and switch controllers are special purpose computers
Pieces:1. End systems2. Transmission3. Switching4. Signaling
Shivkumar KalyanaramanRensselaer Polytechnic Institute
9
Telephone Network: What is It?Single basic service: two-way voice
low end-to-end delayguarantee that an accepted call will run to completion
Endpoints connected by a circuit, like an electrical circuitSignals flow both ways (full duplex)Associated with reserved bandwidth and buffer
resources
Shivkumar KalyanaramanRensselaer Polytechnic Institute
10
Telephone Network Design
Fully connected coresimple routingtelephone number is a hint about how to route a call
But not for 800/888/700/900 numbers: these are pointers to a directory that translates them into regular numbers
hierarchically allocated telephone number space
Shivkumar KalyanaramanRensselaer Polytechnic Institute
11
Telephone Network Design
Shivkumar KalyanaramanRensselaer Polytechnic Institute
12
Telephone Pieces: End Systems
Shivkumar KalyanaramanRensselaer Polytechnic Institute
13
Telephone Pieces: End Systems
Transducers: key to carrying voice on wiresDialerRingerSwitch-hook
Shivkumar KalyanaramanRensselaer Polytechnic Institute
14
Last-Mile Transmission EnvironmentWire gauges:19, 22, 24, 26 gauge(smaller better)Diameters: 0.8, 0.6, 0.5, 0.4 mm (larger better)Various forms of noise: (twisting reduces noise)
Bridged-tap noise: bit-energy diverted to extension phone socketsCrosstalkHam radioAM broadcast
Insertion loss: -140 dBm noise floor 100 million times more sensitive than normal modemsBandwidth range = 600 kHzNotch effects in insertion loss due to bridged-tapsTransmission PSD = -40dBm => 90 dBm budget
Shivkumar KalyanaramanRensselaer Polytechnic Institute
15
2-wire vs 4-wire: Sidetones and Echoes
Both trans & reception circuits need two wires
4 wires from every central office to home
Alternative: Use same pair of wires for bothtransmission and receptionSignal from transmission flows to receiver: sidetoneReverse Effect: received signal at end-system bounces
back to CO (esp if delay > 20 ms): echoSolutions: balance circuit (attenuate side-tone) + echo-
cancellation circuit (cancel echoes).
Shivkumar KalyanaramanRensselaer Polytechnic Institute
16
DialingPulse
sends a pulse per digitcollected by central office (CO)Interpreted by CO switching system to place call or activate special features (eg: call forwarding, prepaid-calls etc)
Tonekey press (feep) sends a pair of tones = digitalso called Dual Tone Multifrequency (DTMF)
CO supplies the power for ringing the bell.Standardized interface between CO and end-system => digital handsets, cordless/cellular phones
Shivkumar KalyanaramanRensselaer Polytechnic Institute
17
Telephone Pieces: Transmission MuxingTrunks between central offices carry hundreds of conversationsCan’t run thick bundles! Instead, send many calls on the same wire
Multiplexing (a.ka. Sharing)Analog multiplexing
Band-limit call to 3.4 KHz and frequency shift onto higher bandwidth trunkobsolete
Digital multiplexingfirst convert voice to samples1 sample = 8 bits of voice8000 samples/sec => call = 64 Kbps
Shivkumar KalyanaramanRensselaer Polytechnic Institute
18
Transmission Multiplexing (contd)How to choose a sample?
256 quantization levels, logarithmically spaced (why?)sample value = amplitude of nearest quantization levelTwo choices of levels (µ law and A law)
Time division multiplexingTrunk carries bits at a faster bit rate than inputsn input streams, each with a 1-byte bufferOutput interleaves samplesNeed to serve all inputs in the time it takes one sample to arrive
=> output runs n times faster than inputOverhead bits mark end of frame (why?)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
19
Transmission MultiplexingMultiplexed trunks can be multiplexed furtherNeed a standard! (why?)US/Japan standard is called Digital Signalinghierarchy (DS)
Digital Signal Number
Number of previous level circuits
Number of voice circuits
Bandwidth
DS0 1 64 KbpsDS1 24 24 1.544MbpsDS2 4 96 6.312 MbpsDS3 7 672 44.736 Mbps
Shivkumar KalyanaramanRensselaer Polytechnic Institute
20
Telephone Pieces: Switching
Shivkumar KalyanaramanRensselaer Polytechnic Institute
21
Telephone Pieces: SwitchingProblem:
each user can potentially call any other usercan’t have (a billion) direct lines!
Switches establish temporary circuitsSwitching systems come in two parts: switch and switch controller
Shivkumar KalyanaramanRensselaer Polytechnic Institute
22
Switching System Components
Shivkumar KalyanaramanRensselaer Polytechnic Institute
23
Switch: What does it do?Transfers data from an input to an output
many ports (up to 200,000 simultaneous calls)need high speeds
Some ways to switch:1. space division switching: eg: crossbarif inputs (or crosspoints) are multiplexed, need a schedule (why?)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
24
Crossbar Switching Elements
Shivkumar KalyanaramanRensselaer Polytechnic Institute
25
Switching (Contd)Another way to switch
time division (time slot interchange or TSI)also needs a service schedule (why?)
To build larger switches we combine space and time To build larger switches we combine space and time division switching elementsdivision switching elements
Shivkumar KalyanaramanRensselaer Polytechnic Institute
26
Telephone pieces: SignalingA switching system has a switch and a switch controllerSwitch controller is in the control plane
does not touch voice samplesManages the network
call routing (collect dialstring and forward call)alarms (ring bell at receiver)billingdirectory lookup (for 800/888 calls)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
27
SignalingSwitch controllers are special purpose computersLinked by their own internal computer network
Common Channel Interoffice Signaling (CCIS) network
Earlier design used in-band tones, but was hackedAlso was very rigid (why?)
Messages on CCIS conform to Signaling System 7 (SS7)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
28
Signaling (contd)One of the main jobs of switch controller: keep track of state of every endpointKey is state transition diagram
Shivkumar KalyanaramanRensselaer Polytechnic Institute
29
Telephony Routing of Signaled Calls
Circuit-setup (I.e. the signaling call) is what is routed. Voice then follows route, and claims reserved resources. 3-level hierarchy, with a fully-connected coreAT&T: 135 core switches with nearly 5 million circuitsLECs may connect to multiple cores
Shivkumar KalyanaramanRensselaer Polytechnic Institute
30
Telephony Routing algorithmIf endpoints are within same CO, directly connectIf call is between COs in same LEC, use one-hop path between COsOtherwise send call to one of the coresOnly 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… )
Shivkumar KalyanaramanRensselaer Polytechnic Institute
31
Features of telephone routingResource reservation aspects:
Resource reservation is coupled with path reservationConnections need resources (same 64kbps)Signaling to reserve resources and the path
Stable loadNetwork built for voice only.Can predict pairwise load throughout the dayCan 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
Shivkumar KalyanaramanRensselaer Polytechnic Institute
32
Features of telephone routingSource can learn topology and compute routeCan assume that a chosen route is available as the signaling proceeds through the networkComponent reliability drove system reliability and hence acceptance of service by customers
Simplified topology: Very highly connected networkHierarchy + full mesh at each level: simple routingHigh cost to achieve this degree of connectivity
Organizational aspects:Single organization controls entire coreAfford the scale economics to build expensive networkCollect global statistics and implement global changes
=> Source-based, signaled, simple alternate-path routing
Shivkumar KalyanaramanRensselaer Polytechnic Institute
33
Telecommunications Regulation HistoryFCC regulations cover telephony, cable, broadcast TV,
wireless etc“Common Carrier”: provider offers conduit for a fee and does not control the content
Customer controls content/destination of transmission & assumes criminal/civil responsibility for content
Local monopolies formed by AT&T’s acquisition of independent telephone companies in early 20th century
Regulation forced because they were deemed natural monopolies (only one player possible in market due to enormous sunk cost)FCC regulates interstate calls and state commissions regulate intra-state and local callsBells + 1000 independents interconnected & expanded
FCC rulemaking process:Intent to act, solicitation of public comment etc…
Shivkumar KalyanaramanRensselaer Polytechnic Institute
34
Deregulation of telephony1960s-70s: gradual de-regulation of AT&T due to technological advances
Terminal equipment could be owned by customers (CPE) => explosion in PBXs, fax machines, handsets Modified final judgement (MFJ): breakup of AT&T into ILECs (incumbent local exchange carrier) and IXC (inter-exchange carrier) part
Long-distance opened to competition, only the local part regulated… Equal access for IXCs to the ILEC network1+ long-distance number introduced then…
800-number portability: switching IXCs => retain 800 number1995: removed price controls on AT&T
Shivkumar KalyanaramanRensselaer Polytechnic Institute
35
Telecom Act of 1996Required ILECs to open their markets through unbundling of network elements (UNE-P), facilities ownership of CLECs….
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…
ILECs could compete in long-distance after demonstrating opening of markets
Only now some ILECs are aggressively entering long distance marketsCLECs failed due to a variety of reasons…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)
ILECs still retain over 90% of local marketWireless substitution has caused ILECs to develop wireless business units
Shivkumar KalyanaramanRensselaer Polytechnic Institute
36
US Telephone Network Structure (after 1984)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
37
Exchange Area Network
Shivkumar KalyanaramanRensselaer Polytechnic Institute
38
Cable TV Networks
Shivkumar KalyanaramanRensselaer Polytechnic Institute
39
Cable Technology Coaxial cable RF distribution networks. Attributes:
Broadcast, low-band reverse channelsMainly one-way video channelsReasonably secure network (private conduit to home)Free from free-space interferencesGood signal capacity (over 1 GHz) and flexibilityMultiple signaling channelsSignificant attenuation that increases proportional to frequency => (active) RF amplification (every 1000 ft)Freq responses of deployed amps and filters limit practical usage of frequencies > 1 GHz
Shivkumar KalyanaramanRensselaer Polytechnic Institute
40
Cable Building Blocks
Shivkumar KalyanaramanRensselaer Polytechnic Institute
41
Cable Spectrum: Upto 750 Mhz
Shivkumar KalyanaramanRensselaer Polytechnic Institute
42
Cable Technology & Architecture Head-end: signal processing center
Each carrier: Baseband analog or digital modulationCarriers multiplexed w/ freq-selective diplex filters:
allows simultaneous info transfer in both directionsTree-and-branch architecture:
Well-suited for one-way broadcast video transmission (same signals to every customer)Accumulates noise & distortions (amplifiers)Affects plant reliability and received signal qualityLimits on the number of amplifiers cascadedLimits on bandwidth in operation (few 100s of MHz): below cable potential…
Makes delivery of “switched” services (separate stream for each customer) difficult
Shivkumar KalyanaramanRensselaer Polytechnic Institute
43
Tree-and-Branch Architecture
Shivkumar KalyanaramanRensselaer Polytechnic Institute
44
Fiber Optics For Cable NetworksKey: Leave the laser ON and intensity-modulate with the analog signalSuch analog modulated lasers are very different from their digital counterparts
Low internal noise and high linearity in the rangeReceiver: simple photo-detector -> back to RF spectrum Result: Hybrid fiber-coax infrastructure, with fiber closer to headend
Coax plant serves smaller range (segmentation), but overall HFC reach dramatically increasedAlso, it allows the economical support of remote, smaller clusters of homesEach part could also provide different services to area (micro-market segmentation)Assign different portions of HFC spectrum to diff uses: many virtual networks: sustained investments possible
Shivkumar KalyanaramanRensselaer Polytechnic Institute
45
Hybrid Fiber Coax (HFC) Networks
Shivkumar KalyanaramanRensselaer Polytechnic Institute
46
Multiple Services over HFC
Shivkumar KalyanaramanRensselaer Polytechnic Institute
47
Future Potential of HFC BroadbandDue to smaller loops, the region from 900MHz –1 GHz can be used for data.
Reduced noise in this region => increased bit rate (200 Mbps) per segment…
Future: fiber moves closer, smaller coax-segments, reduced homes per coax run (60 homes), use of frequencies above 1 Ghz using new electronics
Latest DOCSIS 2.0 spec: 256 QAM (=> 8 bits/Hz) or S-CDMA on cable for more robust transmissions
Shivkumar KalyanaramanRensselaer Polytechnic Institute
48
Cable RegulationVery different from telephony: not common-carrier
Able to control content AND the conduit!Grew by providing an alternative (and extension) to broadcast TV and had initial growth troublesDid not have to offer service on a non-discriminatory basis (unlike common carriers)Asserted first-amendment rights to maintain control over contentNot required to provide access to their distribution system to other providers (some portion of capacity required to be offered to unaffiliated players: eg: CNN)But they reserve rights to appropriately bundle these channels
Limited regulation: basic tier is rate-regulated by local authorities till 1999 based upon FCC rules
Shivkumar KalyanaramanRensselaer Polytechnic Institute
49
Cable regulation (contd)Cable networks limited in horizontal expansion, and from vertically integrating w/ CNN etc
Note: ILECs like Bell Atlantic in contrast merged with IXCs like GTEAT&T’s cable acquisitions were interesting (and will be explored later…)
Cable service is multi-faceted and varied from area to area => regulation formulation more complicatedOver-builders (satellite providers) got access to independent content providers: otherwise regulation achieved little for cableLocal authorities get revenue from cable regulationHFC dominates franchise regulation talks, but cable providers are not obligated to provide broadband access..
Shivkumar KalyanaramanRensselaer Polytechnic Institute
50
Data Networking and the Internet
Shivkumar KalyanaramanRensselaer Polytechnic Institute
51
Recall: Indirect Connectivity…
Indirect Connectivityswitched networks
=> switches
inter-networks
=> routers
Shivkumar KalyanaramanRensselaer Polytechnic Institute
52
Inter-Networks: Networks of Networks
=
…
Internet
…
… …
The internet is just a big switch providing indirect connectivity
Shivkumar KalyanaramanRensselaer Polytechnic Institute
53
Recall: Connecting N users: Directly…
Pt-pt: connects only two users directly…How to connect N users directly ?
What are the costs of each option? Does this method of connectivity scale ?
A B
. . .
Full meshBus
Shivkumar KalyanaramanRensselaer Polytechnic Institute
54
Point-to-Point Connectivity Issues
A B
Physical layer: coding, modulation etcLink layer needed if the link is shared bet’napps; is unreliable; and is used sporadicallyNo need for protocol concepts like addressing, names, routers, hubs, forwarding, filtering …
Shivkumar KalyanaramanRensselaer Polytechnic Institute
55
Link Layer: Serial IP (SLIP)Simple: only framing = Flags + byte-stuffingCompressed headers (CSLIP) for efficiency on low speed links for interactive traffic.Problems:
Need other end’s IP address a priori (can’t dynamically assign IP addresses)No “type” field => no multi-protocol encapsulationNo checksum => all errors detected/corrected by higher layer.
RFCs: 1055, 1144
Shivkumar KalyanaramanRensselaer Polytechnic Institute
56
Link Layer: PPPPoint-to-point protocolFrame format similar to HDLCMulti-protocol encapsulation, CRC, dynamic address allocation possible
key fields: flags, protocol, CRC Asynchronous and synchronous communications possibleLink and Network Control Protocols (LCP, NCP) for flexible control & peer-peer negotiationCan be mapped onto low speed (9.6Kbps) and high speed channels (SONET)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
57
Connecting N users: Directly ...Bus: Low cost vs broadcast/collisions, MAC complexityFull mesh: High cost vs simplicityNew concept:
Address to identify nodes. Needed if we want the receiver alone to consume the packet!
. . .
Full meshBus
Problem: Direct connectivity does not “scale”….
Shivkumar KalyanaramanRensselaer Polytechnic Institute
58
How to build Scalable Networks?
Scaling: system allows the increase of a key parameter. Eg: let N increase…
Inefficiency limits scaling …
Direct connectivity is inefficient & hence does not scale
Mesh: inefficient in terms of # of linksBus architecture: 1 expensive link, N cheap links. Inefficient in bandwidth use
Shivkumar KalyanaramanRensselaer Polytechnic Institute
59
Filtering, forwarding …Filtering: choose a subset of elements from a set
Don’t let information go where its not supposed to…Filtering => More efficient => more scalable
Filtering is the key to efficiency & scaling
Forwarding: actually sending packets to a filtered subset of link/node(s)
Packet sent to one link/node => efficient
Solution: Build nodes which focus on filtering/forwardingand achieve indirect connectivity
“switches” & “routers”
Shivkumar KalyanaramanRensselaer Polytechnic Institute
60
Connecting N users: IndirectlyStar: One-hop path to any node, reliability, forwarding function“Switch” S can filter and forward!
Switch may forward multiple pkts in parallel for additional efficiency!
S Star
Shivkumar KalyanaramanRensselaer Polytechnic Institute
61
Connecting N users: Indirectly …Ring: Reliability to link failure, near-minimal links All nodes need “forwarding” and “filtering” Sophistication of forward/filter lesser than switch
Ring
Shivkumar KalyanaramanRensselaer Polytechnic Institute
62
RingStarS
Tree
Topologies: Indirect Connectivity
Shivkumar KalyanaramanRensselaer Polytechnic Institute
63
Protocol Issues in Data NetworksPt-Pt connectivity:
FramingError control/ReliabilityFlow control & Windowing protocols
Multiplexing, VirtualizationCircuit vs Packet Switching: a muxing view
MAC arbitration schemes: Random access/CSMA, TDMA, CDMA
Interconnection components: repeater, hub, bridge, switch, router
Shivkumar KalyanaramanRensselaer Polytechnic Institute
64
Reliability: Types of errors & effectsForward channel bit-errors (garbled packets)Forward channel packet-errors (lost packets)Reverse channel bit-errors (garbled status reports)Reverse channel bit-errors (lost status reports)
Protocol-induced effects: Duplicate packetsDuplicate status reportsOut-of-order packetsOut-of-order status reportsOut-of-range packets/status reports (in window-based transmissions)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
65
Temporal Redundancy ModelPackets • Sequence Numbers
• CRC or Checksum
Status Reports • ACKs• NAKs, • SACKs• Bitmaps
Timeout
Retransmissions• Packets• FEC information
Shivkumar KalyanaramanRensselaer Polytechnic Institute
66
Reliability MechanismsMechanisms:
Checksum: detects corruption in pkts & acksACK: “packet correctly received”Duplicate ACK: “packet incorrectly received”Sequence number: identifies packet or ack
1-bit sequence number used both in forward & reversechannel
Timeout only at senderReliability capabilities achieved:
An error-free channelA forward & reverse channel with bit-errorsDetects duplicates of packets/acksNAKs eliminatedA forward & reverse channel with packet-errors (loss)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
67
Stop and Wait Flow Control
Data
Ack
Ack
Data
tframe
tprop
α =tprop
tframe
=Distance/Speed of SignalFrame size /Bit rate
=Distance × Bit rateFrame size × Speed of Signal
=1
2α + 1
U=2tprop+tframe
tframe
U
αLight in vacuum= 300 m/µs
Light in fiber = 200 m/µsElectricity = 250 m/µs
Shivkumar KalyanaramanRensselaer Polytechnic Institute
68
Data
Ack
tframe
tprop
U=Ntframe
2tprop+tframe
=
N2α+1
1 if N>2α+1
Sliding Window Protocols
Shivkumar KalyanaramanRensselaer Polytechnic Institute
69
Multiplexing: The Method of Sharing Costly Resources
Multiplexing = sharingAllows system to achieve “economies of scale”Cost: waiting time (delay), buffer space & loss Gain: Money ($$) => Overall system costs less
Full Mesh Bus
Shivkumar KalyanaramanRensselaer Polytechnic Institute
70
VirtualizationThe multiplexed shared resource with a level of indirection will seem like a unshared virtual resource!
I.e. Multiplexing + indirection = virtualization
We can “refer” to the virtual resource as if it were the physical resource.
Eg: virtual memory, virtual circuits… Connectivity: a virtualization created by the Internet!Indirection requires binding and unbinding…
Eg: use of packets, slots, tokens etc
. . .
Physical Bus
=A B
A B
Virtual Pt-Pt Link
Shivkumar KalyanaramanRensselaer Polytechnic Institute
71
Statistical MultiplexingReduce resource requirements (eg: bus capacity) by exploiting statistical knowledge of the system.
Eg: average rate <= service rate <= peak rate
If service rate < average rate, then system becomes unstable!!
First design to ensure system stability!!
Then, for a stable multiplexed system: Gain = peak rate/service rate. Cost: buffering, queuing delays, losses.
Useful only if peak rate differs significantly from average rate.
Eg: if traffic is smooth, fixed rate, no need to play games with capacity sizing…
Shivkumar KalyanaramanRensselaer Polytechnic Institute
72
Stability of a Multiplexed SystemAverage 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
Shivkumar KalyanaramanRensselaer Polytechnic Institute
73
What’s a performance tradeoff ? • A situation where you cannot get something for nothing!
• Also known as a zero-sum game.
R=link bandwidth (bps)L=packet length (bits)a=average packet arrival rate
Traffic intensity = La/R
Shivkumar KalyanaramanRensselaer Polytechnic Institute
74
What’s a performance tradeoff ?
La/R ~ 0: average queuing delay smallLa/R -> 1: delays become largeLa/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.
Shivkumar KalyanaramanRensselaer Polytechnic Institute
75
How to design large inter-networks? Circuit-Switching
Divide link bandwidth into “pieces”Reserve pieces on successive links and tie them together to form a “circuit”Map traffic into the reserved circuitsResources wasted if unused: expensive.
– Mapping can be done without “headers”. – Everything inferred from timing.
Shivkumar KalyanaramanRensselaer Polytechnic Institute
76
How to design large inter-networks? Packet-Switching
Chop up data (not links!)into “packets”
Packets: data + meta-data (header)
“Switch” packets at intermediate nodes
Store-and-forward if bandwidth is not immediately available.
Bandwidth division into “pieces”Dedicated allocationResource reservation
Shivkumar KalyanaramanRensselaer Polytechnic Institute
77
Packet Switching
45 Mbsqueue of packetswaiting 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
10 MbsEthernet
statistical multiplexingCA
1.5 MbsB
D E
Shivkumar KalyanaramanRensselaer Polytechnic Institute
78
Spatial vs Temporal Multiplexing Spatial multiplexing: Chop up resource into chunks. Eg: bandwidth, cake, circuits…
Temporal multiplexing: resource is shared over time, I.e. queue up jobs and provide access to resource over time. Eg: FIFO queueing, packet switching
Packet switching is designed to exploit both spatial & temporal multiplexing gains, provided performance tradeoffs are acceptable to applications.
Packet switching is potentially more efficient => potentially more scalable than circuit switching !
Shivkumar KalyanaramanRensselaer Polytechnic Institute
79
Protocol Issues in Data Networks (Contd)
Pt-Pt connectivity:FramingError control/ReliabilityFlow control & Windowing protocols
Multiplexing, VirtualizationCircuit vs Packet Switching: a muxing view
MAC arbitration schemes: Random access/CSMA, TDMA, CDMA
Interconnection components: repeater, hub, bridge, switch, router
Shivkumar KalyanaramanRensselaer Polytechnic Institute
80
Multi-Access LANsHybrid topologies:
Uses directly connected topologies (eg: bus), orIndirectly connected with simple filtering components (switches, hubs).
Limited scalability due to limited filtering
Medium Access Protocols:ALOHA, CSMA/CD (Ethernet), Token Ring …Key: Use a single protocol in network
Concepts: address, forwarding (and forwarding table), bridge, switch, hub, token, medium access control (MAC) protocols
Shivkumar KalyanaramanRensselaer Polytechnic Institute
81
MAC Protocols: a taxonomyThree broad classes:
Channel Partitioningdivide channel into smaller “pieces” (time slots, frequency)allocate piece to node for exclusive use
“Taking turns”: Token-basedtightly coordinate shared access to avoid collisions
Random Accessallow collisions“recover” from collisions
Goal: efficient, fair, simple, decentralized
Shivkumar KalyanaramanRensselaer Polytechnic Institute
82
Channel PartitioningMAC protocols. Eg: TDMA
TDMA: time division multiple accessAccess to channel in "rounds" Each station gets fixed length slot (length = pkt trans time) in each round Unused slots go idle Example: 6-station LAN, 1,3,4 have pkt, slots 2,5,6 idle
Shivkumar KalyanaramanRensselaer Polytechnic Institute
83
“Taking Turns” MAC protocols - 1Channel partitioning MAC protocols:
share channel efficiently at high loadinefficient at low load: delay in channel access, 1/N bandwidth allocated even if only 1 active node!
Random access MAC protocolsefficient at low load: single node can fully utilize channelhigh load: collision overhead
“Taking turns” protocolslook for best of both worlds!
Shivkumar KalyanaramanRensselaer Polytechnic Institute
84
“Taking Turns” MAC protocols - 2Polling:
Master node “invites” slave nodes to transmit in turnRequest to Send, Clear to Send messagesConcerns:
polling overhead latencysingle point of failure (master)
Token passing:Control token passed from one node to next sequentially.Token messageConcerns:
token overhead latencysingle point of failure
(token)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
85
“Taking Turns” Protocols –3Reservation-based a.k.a Distributed Polling:
Time divided into slotsBegins with N short reservation slots
reservation slot time equal to channel end-end propagation delay station with message to send posts reservationreservation seen by all stations
After reservation slots, message transmissions ordered by known priority
Shivkumar KalyanaramanRensselaer Polytechnic Institute
86
Random Access ProtocolsAloha at University of Hawaii: Transmit whenever you likeWorst case utilization = 1/(2e) =18%CSMA: Carrier Sense Multiple Access Listen before you transmitCSMA/CD: CSMA with Collision DetectionListen while transmitting. Stop if you hear someone else.Ethernet uses CSMA/CD.Standardized by IEEE 802.3 committee.
Shivkumar KalyanaramanRensselaer Polytechnic Institute
87
10Base5 Ethernet Cabling RulesThick coaxLength of the cable is limited to 2.5 km, no more than 4 repeaters between stationsNo more than 500 m per segment ⇒ “10Base5”
2.5m500 mTransceiver
Repeater
Terminator
Shivkumar KalyanaramanRensselaer Polytechnic Institute
88
10Base5 Cabling Rules (Continued)No more than 2.5 m between stationsTransceiver cable limited to 50 m
2.5m500 mTransceiver
Repeater
Terminator
Shivkumar KalyanaramanRensselaer Polytechnic Institute
89
Inter-connection DevicesRepeater: Layer 1 (PHY) device that restores data and collision signals: a digital amplifier
Hub: Multi-port repeater + fault detectionNote: broadcast at layer 1
Bridge: Layer 2 (Data link) device connecting two or more collision domains.
Key: a bridge attempts to filter packets and forward them from one collision domain to the other.It snoops on passing packets and learns the interface where different hosts are situated, and builds a L2 forwarding tableMAC multicasts propagated throughout “extended LAN.”Note: Limited filtering intelligence and forwarding capabilities at layer 2
Shivkumar KalyanaramanRensselaer Polytechnic Institute
90
Interconnection Devices (Continued)Router: Network layer device. IP, IPX, AppleTalk. Interconnects broadcast domains.
Does not propagate MAC multicasts.
Switch:Key: has a switch fabric that allows parallel forwarding pathsLayer 2 switch: Multi-port bridge w/ fabricLayer 3 switch: Router w/ fabric and per-port ASICs
These are functions. Packaging varies.
Shivkumar KalyanaramanRensselaer Polytechnic Institute
91
Interconnection DevicesExtended LAN=Broadcast domainH H B H H
LAN=CollisionDomain
Router
Application Application
RouterBridge/SwitchRepeater/Hub
Gateway
NetworkDatalinkPhysical
TransportNetworkTransport
DatalinkPhysical
Shivkumar KalyanaramanRensselaer Polytechnic Institute
92
Ethernet (IEEE 802) Address Format
OUI(Organizationally Unique ID)
10111101
G/L bit(Global/Local)
G/I bit(Group/Individual)
48-bit flat address => no hierarchy to help forwardingHierarchy only for administrative/allocation purposesAssumes that all destinations are (logically) directly connected.
Address structure does not explicitly acknowledge indirect connectivity
=> Sophisticated filtering cannot be done!
Shivkumar KalyanaramanRensselaer Polytechnic Institute
93
Ethernet (IEEE 802) Address Format
G/L bit: administrativeGlobal: unique worldwide; assigned by IEEELocal: Software assigned
G/I: bit: multicastI: unicast addressG: multicast address. Eg: “To all bridges on this LAN”
10111101
G/L bit(Global/Local)
G/I bit(Group/Individual)
OUI(Organizationally Unique ID)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
94
Ethernet & 802.3 Frame Format
Ethernet
❑ IEEE 802.3
Dest.Address
SourceAddress
IP IPX AppleTalk
Type
6 6 2
Size in bytesInfo CRC
4
Dest.Address
SourceAddress Length
6 6 2
IP IPX AppleTalk
LLC CRC
4
Pad
Length
Info
• Maximum Transmission Unit (MTU) = 1518 bytes• Minimum = 64 bytes (due to CSMA/CD issues)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
95
Network/Transport Layer IssuesInter-networking: heterogeneity, scaleRoutingCongestion controlQuality of Service (QoS)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
96
Inter-Networks: Networks of Networks
What is it ?“Connect many disparate physical networks and make them function as a coordinated unit … ” - Douglas ComerMany => scaleDisparate => heterogeneity
Result: Universal connectivity!The inter-network looks like one large switch, User interface is sub-network independent
Shivkumar KalyanaramanRensselaer Polytechnic Institute
97
Inter-Networks: Networks of Networks
Internetworking involves two fundamental problems: heterogeneity and scale
Concepts: Translation, overlays, address & name resolution, fragmentation: to handle heterogeneityHierarchical addressing, routing, naming, address allocation, congestion control: to handle scaling
Two broad approaches: circuit-switched and packet-switched
Shivkumar KalyanaramanRensselaer Polytechnic Institute
98
Scalable Forwarding, Structured Addresses
Address has structure which aids the forwarding process.Address assignment is done such that nodes which can be reached without resorting to L3 forwarding have the same prefix (network ID)A simple comparison of network ID of destination and current network (broadcast domain) identifies whether the destination is “directly” connected
I.e. Reachable through L2 forwarding onlyWithin L3 forwarding, further structure can aid hierarchical organization of routing domains (because routing algorithms have other scalability issues)
Network ID Host ID
Demarcator
Shivkumar KalyanaramanRensselaer Polytechnic Institute
99
Flat vs Structured AddressesFlat addresses: no structure in them to facilitate scalable routing
Eg: IEEE 802 LAN addressesHierarchical addresses:
Network part (prefix) and host partHelps identify direct or indirectly connected nodes
Shivkumar KalyanaramanRensselaer Polytechnic Institute
100
Internet Routing DriversTechnology and economic aspects:
Internet built out of cheap, unreliable components as an overlay on top of leased telephone infrastructure for WAN transport.
Cheaper components => fail more often => topology changes often => needs dynamic routing
Components (including end-systems) had computation capabilities.
Distributed algorithms can be implemented Cheap overlaid inter-networks => several entities could afford to leverage their existing (heterogeneous) LANs and leased lines to build inter-networks.
Led to multiple administrative “clouds” which needed to inter-connect for global communication.
Shivkumar KalyanaramanRensselaer Polytechnic Institute
101
Internet Routing Model2 key features:
Dynamic routingIntra- and Inter-AS routing, AS = locus of admin control
Internet organized as “autonomous systems” (AS).AS is internally connected
Interior Gateway Protocols (IGPs) within AS. Eg: RIP, OSPF, HELLO
Exterior Gateway Protocols (EGPs) for AS to AS routing. Eg: EGP, BGP-4
Shivkumar KalyanaramanRensselaer Polytechnic Institute
102
Intra-AS and Inter-AS routing
inter-AS, intra-ASrouting in
gateway A.c
network layerlink layer
physical layer
a
b
b
aaC
A
Bd
Gateways:•perform inter-ASrouting amongst themselves•perform intra-ASrouters with other routers in their AS
A.cA.a
C.bB.a
cb
c
Shivkumar KalyanaramanRensselaer Polytechnic Institute
103
Intra-AS and Inter-AS routing: Example
Host h2
a
b
b
aaC
A
Bd c
A.aA.c
C.bB.a
cb
Hosth1
Intra-AS routingwithin AS A
Inter-ASrouting
between A and B
Intra-AS routingwithin AS B
Shivkumar KalyanaramanRensselaer Polytechnic Institute
104
Requirements for Intra-AS RoutingShould scale for the size of an AS.
Low end: 10s of routers (small enterprise)High end: 1000s of routers (large ISP)
Different requirements on routing convergence after topology changes
Low end: can tolerate some connectivity disruptionsHigh end: fast convergence essential to business (making money on transport)
Operational/Admin/Management (OAM) ComplexityLow end: simple, self-configuringHigh end: Self-configuring, but operator hooks for control
Traffic engineering capabilities: high end only
Shivkumar KalyanaramanRensselaer Polytechnic Institute
105
Requirements for Inter-AS RoutingShould scale for the size of the global Internet.
Focus on reachability, not optimalityUse address aggregation techniques to minimize core routing table sizes and associated control trafficAt the same time, it should allow flexibility in topological structure (eg: don’t restrict to trees etc)
Allow policy-based routing between autonomous systemsPolicy refers to arbitrary preference among a menu of available options (based upon options’ attributes)In the case of routing, options include advertised AS-level routes to address prefixesFully distributed routing (as opposed to a signaled approach) is the only possibility.Extensible to meet the demands for newer policies.
Shivkumar KalyanaramanRensselaer Polytechnic Institute
106
λiµλ
µi
If information about λ i , λ and µ is known in a central location where control of λ i or µ can be effected with zero time delays,
the congestion problem is solved!Unfortunately, we have incomplete info, require a distributed solution with time-varying time-delays
The Congestion Problem
λ1
λn
CapacityDemand
•Problem: demand outstrips available capacity
Shivkumar KalyanaramanRensselaer Polytechnic Institute
107
Congestion: A Close-up View
knee – point after which throughput increases very slowlydelay increases fast
cliff – point after whichthroughput starts to decrease very fast to zero(congestion collapse)delay approaches infinity
Note (in an M/M/1 queue)delay = 1/(1 – utilization)
Load
Load
Thro
ughp
utD
elay
knee cliff
congestioncollapse
packetloss
Shivkumar KalyanaramanRensselaer Polytechnic Institute
108
Congestion Control vs. Congestion Avoidance
Congestion control goalstay left of cliff
Congestion avoidance goalstay left of knee
Right of cliff: Congestion collapse
Load
Thro
ughp
ut
knee cliff
congestioncollapse
Shivkumar KalyanaramanRensselaer Polytechnic Institute
109
Goals of Congestion ControlTo guarantee stable operation of packet networks
Sub-goal: avoid congestion collapse
To keep networks working in an efficient status Eg: high throughput, low loss, low delay, and high utilization
To provide fair allocations of network bandwidth among competing flows in steady state
For some value of “fair” ☺109
Shivkumar KalyanaramanRensselaer Polytechnic Institute
110
CC Techniques: Self-clockingPrPb
Ar
Ab
ReceiverSender
As
Implications of ack-clocking:
More batching of acks => bursty traffic
Less batching leads to a large fraction of Internet traffic being just acks (overhead)
Shivkumar KalyanaramanRensselaer Polytechnic Institute
111
CC Techniques: Additive Increase/Multiplicative Decrease (AIMD) Policy
x0
x1
x2
Efficiency Line
Fairness Line
User 1’s Allocation x1
User 2’s Allocation
x2
Assumption: decrease policy must (at minimum) reverse the load increase over-and-above efficiency line
Implication: decrease factor should be conservatively set to account for any congestion detection lags etc
Shivkumar KalyanaramanRensselaer 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
Shivkumar KalyanaramanRensselaer Polytechnic Institute
113
Fundamental QoS Problems
B
Scheduling DisciplineFIFO
B
In a FIFO service discipline, the performance assigned to one flow is convoluted with the arrivals of packets from all other flows!
Cant get QoS with a “free-for-all”Need to use new scheduling disciplines which provide “isolation” of performance from arrival rates of background traffic
Shivkumar KalyanaramanRensselaer Polytechnic Institute
114
Fundamental QoS ProblemsConservation Law (Kleinrock): Σρ(i)Wq(i) = K
Irrespective of scheduling discipline chosen:
Average backlog (delay) is constantAverage bandwidth is constant
Zero-sum game => need to “set-aside” resources for premium services
Shivkumar KalyanaramanRensselaer Polytechnic Institute
115
QoS Big Picture: Control/Data Planes
Internetwork or WANWorkstationRouter
Router
RouterWorkstation
Control Plane: Signaling + Admission Control orSLA (Contracting) + Provisioning/Traffic Engineering
Data Plane: Traffic conditioning (shaping, policing, markingetc) + Traffic Classification + Scheduling, Buffer management
Shivkumar KalyanaramanRensselaer Polytechnic Institute
116
Internet RegulationFCC has largely had a hands-off policyEarly development of internet in part was influenced by high cost of telecom links
Packet switching developed as better multiplexing technologyCommon-carriage regulation has affected Inet:
Eg: modems were like fax machine for the common carrierUse of basic service (eg: telephony) to provide enhanced service (eg: internet access) => not subject to FCC or state jurisdiction
Led to community bulletin-boards, ISPs, value-added networks (frame-relay?)…Home-to-ISP treated as local call (even if crossed state-boundaries)
ILECs prohibited from offering inter-LATA servicesDSL viewed as basic service => must unbundle DSL to allow 3rd
parties to offer internet access over ILEC DSL
Shivkumar KalyanaramanRensselaer Polytechnic Institute
117
Summary: List of Internet Problems
Basics: Direct/indirect connectivity, topologiesLink layer issues:
Framing, Error control, Flow control Multiple access & Ethernet:
Cabling, Pkt format, Switching, bridging vs routingInternetworking problems: Naming, addressing, Resolution, fragmentation, congestion control, traffic management, Reliability, Network Management
Shivkumar KalyanaramanRensselaer Polytechnic Institute
118
Additional Reading
Internet Design Philosophy:Saltzer, Reed, Clark: "End-to-End arguments in System Design" Clark: "The Design Philosophy of the DARPA Internet Protocols": RFC 2775: Internet Transparency: In HTML
top related