Lecture 3, 1Spring 2003, COM1337/3501CCN: Direct Link Networks Direct Link Networks Textbook:...

Spring 2003, COM1337/3501 CCN: Direct Link Networks Lecture 3, 1

Direct Link Networks

Textbook: Computer Networks: A Systems Approach, L. Peterson, B. Davie, Morgan Kaufmann

Chapter 2.


Direct Links: Outline

• Physical Layer– Link technologies– Encoding

• Link Layer– Framing– Error Detection– Reliable Transmission (ARQ protocols)– Medium Access Control:

• Existing protocols: Ethernet, Token Rings, Wireless


Link Technologies• Cables:

– Cat 5 twisted pair, 10-100Mbps, 100m– Thin-net coax, 10-100Mbps, 200m– Thick-net coax, 10-100Mbps, 500m– Fiber, 100Mbps-2.4Gbps, 2-40km

• Leased Lines:– Copper based: T1 (1.544Mbps), T3 (44.736Mbps)– Optical fiber: STS-1 (51.84Mbps), STS-N (N*51.84Mbps)

• Last-Mile Links:– POTS (56Kbps), ISDN (2*64Kbps)– xDSL: ADSL (16-640Kbps, 1.554-8.448Mbps), VDSL (12.96Mbps-55.2Mbps)– CATV: 40Mbps downstream, 20Mbps upstream

• Wireless Links: Cellular, Satellite, Wireless Local Loop


Encoding

• Signals propagate over a physical medium– modulate electromagnetic waves

– e.g., vary voltage

• Encode binary data onto signals– e.g., 0 as low signal and 1 as high signal

– known as Non-Return to zero (NRZ)

Bits

NRZ

0 0 1 0 1 1 1 1 0 1 0 0 0 0 1 0


Problem: Consecutive 1s or 0s

• Low signal (0) may be interpreted as no signal• High signal (1) leads to baseline wander

– Receiver compares rx_signal to avg_signal

• Clock drift: Unable to recover clock


Alternative Encodings• Non-return to Zero Inverted (NRZI)

– make a transition from current signal to encode a one; stay at current signal to encode a zero

– solves the problem of consecutive ones

• Manchester– transmit XOR of the NRZ encoded data and the clock– only 50% efficient.


Encodings (cont)

Bits

NRZ

Clock

Manchester

NRZI

0 0 1 0 1 1 1 1 0 1 0 0 0 0 1 0


Encodings (cont)• 4B/5B

– every 4 bits of data encoded in a 5-bit code– 5-bit codes selected to have no more than one leading 0 and

no more than two trailing 0s – thus, never get more than three consecutive 0s– resulting 5-bit codes are transmitted using NRZI – achieves 80% efficiency

4b data 5b code 4b data 5b code0000 11110 1000 100100001 01001 1001 100110010 10100 1010 101100011 01010 1011 101110100 01011 1100 110100101 01110 1101 110110110 10010 1110 111000111 10011 1111 11101


Framing• The physical layer provides a mean to transmit a sequence

of bits

• How can one determine the beginning/end of a frame?

• Solutions:– Character-based framing (use special control characters)

– Bit-oriented framing with flags

– Length counts

– Clock based


Character Based Framing• BISYNC: BInary SYNchronous Communication

– SYN: Synchronous idle, SOH: Start of Header, STX: Start of text, ETX: End of text

• Problem 1: if control characters appear within the header, or CRC.– These are known locations, one can skip control characters in these fields

• Problem 2: if CTRL characters appear in the packet.– Use a Data Link Escape (DLE) character before ETX when it appears

within the packet

– If DLE appears within the packet replace it with DLE DLE.

SYN SYN SOH Header Body ETX CRC SYNSTX


Bit Based Framing• Sentinel-based (Bit oriented)

– delineate frame with special pattern: 01111110

– e.g., HDLC, SDLC

– problem: special pattern appears in the payload

– solution: bit stuffing

• sender: insert 0 after five consecutive 1s

• receiver: delete 0 that follows five consecutive 1s

• Disadvantage: potentially increases the length by 20%

Header Body

8 16 16 8

CRCBeginningsequence

Endingsequence


Counter Based Framing

– include payload length in header– e.g., DDCMP

– problem: count field corrupted– solution: catch when CRC fails

SY

N

Header Body

8 8 4214 168

SY

N

Cla

ss CRCCount


Clock Based Framing• SONET: Synchronous Optical Network • Each frame is 125 micro-seconds long: 810 bytes

for STS-1• STS-n (STS-1 = 51.84 Mbps)Overhead Payload

90 columns

9 rows




• Link Layer– Framing– Error Detection– Reliable Transmission (ARQ protocols)– Medium Access Control (MAC):



Error Detection

• Bits can get corrupted on transmissions• How to detect, and if possible, correct them• Error detection techniques:

– Parity bit

– Internet checksum

– Cyclic redundancy check (CRC)


Parity Bit

• Simplest of all techniques• Single parity bit

– Add a bit that takes the XOR of all the data bits

– What kind of errors can it detect?

• Two dimensional parity:– Arrange the data in a 2D matrix

– Take the parity along each row and row column

– Send the data with the parity bits

– What kind of errors can it detect?


Internet Checksum Algorithm• View message as a sequence of 16-bit integers; sum using

16-bit ones-complement arithmetic; take ones-complement of the result.

u_shortcksum(u_short *buf, int count){ register u_long sum = 0; while (count--) { sum += *buf++; if (sum & 0xFFFF0000) { /* carry occurred, so wrap around */ sum &= 0xFFFF; sum++; } } return ~(sum & 0xFFFF);}


Cyclic Redundancy Check

• Add k bits of redundant data to an n-bit message– want k << n

– e.g., k = 32 and n = 12,000 (1500 bytes)

• Represent n-bit message as n-1 degree polynomial– e.g., MSG=10011010 as M(x) = x7 + x4 + x3 + x1

• Let k be the degree of some divisor polynomial– e.g., C(x) = x3 + x2 + 1


CRC (cont)• Transmit polynomial P(x) that is evenly divisible

by C(x) – shift left k bits, i.e., M(x)xk

– subtract remainder of M(x)xk / C(x) from M(x)xk

• Receiver polynomial P(x) + E(x)– E(x) = 0 implies no errors

• Divide (P(x) + E(x)) by C(x); remainder zero if:– E(x) was zero (no error), or– E(x) is exactly divisible by C(x)


Selecting C(x)• All single-bit errors, as long as the xk and x0 terms have

non-zero coefficients.• Any odd number of errors, as long as C(x) contains the

factor (x + 1)• Any ‘burst’ error (i.e., sequence of consecutive error bits)

for which the length of the burst is less than k bits.

• Most burst errors of larger than k bits can also be detected

• See Table 2.6 on page 102 for common C(x)


Reliable Transmission: ARQ

• Automatic Repeat reQuest (ARQ)• Underlying physical channel:

– Each transmitted frame may be delayed by an arbitrary and variable time

– Some frames might be lost and may never arrive– Assume that error detection works correctly– Frames that arrive are assumed to do so in order, with or without errors

(this assumption is not always necessary, true for direct links)

• Correctness of ARQ protocol: each packet is released to the network layer once and only once, without error

• Efficiency: Use of link bandwidth (effective throughput)


Acknowledgements & TimeoutsSender Receiver

Tim

eout

Tim

e

Sender Receiver

Tim

eout

Tim

eout

Sender Receiver

Tim

eout

Tim

eout

Sender Receiver

Tim

eout

Tim

eout

(a) (c)

(b) (d)


Stop-and-Wait

• Sender ensures that each frame is received correctly before sending the next frame.

• How to distinguish between successive packets?– Use sequence numbers

• But sequence numbers may grow out of bound?– A 1-bit sequence number suffices for Stop-and-Wait!

Sender Receiver


Problems with Stop-and-Wait

• Keeping the pipe full• Example

– 1.5Mbps link x 45ms RTT = 67.5Kb (8KB)

– 1KB frames implies 1/8th link utilization

• How to keep the pipe full?– Send more frames without waiting for acks


Sliding Window• Allow multiple outstanding (un-ACKed) frames• Upper bound on un-ACKed frames, called window

Sender Receiver

Tim

e

……


Sliding Window: Sender

• Assign sequence number to each frame (SeqNum)• Maintain three state variables:

– send window size (SWS)– last acknowledgment received (LAR)– last frame sent (LFS)

• Maintain invariant: LFS - LAR <= SWS

• Advance LAR when ACK arrives • Buffer up to SWS frames

SWS

LAR LFS

… …


Sliding Window: Receiver• Maintain three state variables

– receive window size (RWS)– largest frame acceptable (LFA)– last frame received in order (LFR)

• Maintain invariant: LFA - LFR <= RWS

• Frame SeqNum arrives:– if LFR < SeqNum < = LFA accept; update LFR, if necessary– if SeqNum < = LFR or SeqNum > LFA discarded

• Send cumulative ACKs • Buffer up to RWS packets

LFR LFA

RWS


Buffer Sizes• SWS set to bandwidth-delay product estimate• For RWS, two common settings

– RWS = 1: receiver will not buffer any frames that arrive out of order

– RWS = SWS

• Does not make sense to set RWS > SWS


Sequence Number Space• SeqNum field is finite; sequence numbers wrap around• Sequence number space must be larger then number of

outstanding frames• If RWS=SWS, SWS <= MaxSeqNum-1 is not sufficient

– suppose 3-bit SeqNum field (0..7)– SWS=RWS=7– sender transmit frames 0..6– arrive successfully, but ACKs lost– sender retransmits 0..6– receiver expecting 7, 0..5, but receives second incarnation of 0..5

• SWS < (MaxSeqNum+1)/2 is correct rule• Intuitively, SeqNum “slides” between two halves of sequence

number space


Concurrent Logical Channels• Multiplex 8 logical channels over a single link• Run stop-and-wait on each logical channel• Maintain three state bits per channel

– channel busy– current sequence number out– next sequence number in

• Header: 3-bit channel num, 1-bit sequence num– 4-bits total

• Separates reliability from order• Data link layer protocol used in ARPANET


Medium Access Control ProtocolsMultiple Access ProtocolsMultiple Access Protocols

ContentionContention

Dynamic Dynamic ResolutionResolution

StaticStaticResolutionResolution

Conflict-freeConflict-free

Dynamic Dynamic AllocationAllocation

Static Static AllocationAllocation

TokenTokenPassingPassing

ReservationReservationIDIDProbabilisticProbabilisticProbabilisticProbabilisticTime ofTime ofarrivalarrival


MAC Protocols: Evaluation

• Throughput• Delay• Buffering• Stability

We also generally assume that:– channel is errorless

– feedback is available


Carrier Sense Protocols• Use the fact that in some networks you can sense the medium to

check whether it is currently free– 1-persistent CSMA– non-persistent CSMA– p-persistent protocol– CSMA with collision Detection (CSMA/CD)

• 1-persistent CSMA– when a station has a packet:

• it waits until the medium is free to transmit the packet• if a collision occurs, the station waits a random amount of time

– first transmission results in a collision if several stations are waiting for the channel


Carrier Sense Protocols (Cont’d)• non-persistent CSMA

– when a station has a packet:• if the medium is free, transmit the packet• otherwise wait for a random period of time and repeat the algorithm

– higher delays, but better performance than pure ALOHA

• p-persistent protocol– when a station has a packet wait until the medium is free:

• transmit the packet with probability p• wait for next slot with probability 1-p

– better throughput than other schemes but higher delay

• CSMA with Collision Detection (CSMA/CD)– stations abort their transmission when they detect a collision– e.g., Ethernet, IEEE802.3


Ethernet• History: evolution from Aloha, CSMA, CSMA/CD (by Xerox

PARC) => Ethernet, => IEEE802.3 (Digital, Intel, Xerox)– There are slight differences between Ethernet and 802.3 (e.g., 802.3 length

field is used for packet type in Ethernet, various transmission speeds for 802.3 from 1 to 10Mbps)

• Physical layer (10Mbps Ethernet): – Manchester encoding (bit synchronous, no-dc component)– Cabling: maximum 500 meters with up to 4 repeaters (max 2500m)

10Base5 Thick coax 500 m 100 nodes Good for backbones

10Base2 Thin coax 200 m 30 Cheapest system

10Base-T Twisted pair 100 m 1024 Easy maintenance

10Base-F Fiber optics 2000 m 1024 Between buildings


Frame Format (IEEE802.3)

• Preamble : 7x10101010… (allows the receiver’s clock to synchronize)

• SF: 10101011• 10Mbps has only 6 bytes addresses:

– Unicast: unique per adaptor (ranges are allocated to manufacturers)– Broadcast: FF:FF:FF:FF:FF:FF– Multicast: first address bit = 1– Internet Multicast: 01:00:5e:00:00:00 -to- 01:00:5e:7f:ff:ff

• Pad: minimum frame length of 64 bytes


Ethernet Algorithm• Receiver: accepts frames with a correct CRC

• Sender: CSMA/CD 1-persistent algorithm– If the adaptor has a frame and the line is idle: transmit, otherwise wait

until idle line then transmit

– If a collision occurs:• When detected a 32-bit jamming sequence is sent

• Binary exponential backoff: select a random number [0, 2i-1] and waits for that many slots before transmitting

• After ten collisions the randomization interval is frozen to max 1023

• After 16 collisions the controller throws away the frame

• What is the reason for having a minimal frame length? (Hint RTT: 51.2s)


Ethernet Performance

• Assume that retransmissions occur with probability p, k stations ready to transmit:– Probability that a station acquires the channel: A=kp(1-p)k-1

– Maximum: when p=1/k, k-> A->1/e– Probability that a contention interval has exactly j slots is: A(1-A)j-1

– Mean number of slots per contention is: 1/A -> e = 2.718…– Slot duration: 2 = 51.2s– Channel efficiency: P/(P+2e), where P is transmission time for a packet


Ethernet Performance


Ethernet Capture Effect

• A and B have a large queue of packets• There exists a situation where B will keep increasing

its backoff interval (and finally dropping its packet) while A is transmitting its packets

• One of the reasons why frame is dropped after 16 collisions


Token Passing MAC• Token Bus (IEEE802.4):

– broadcast bus

– logical ring

– token: special control frame

– only the token holder station can transmit frames

– 0, 2, 4, 6: traffic priority classes

• Token Ring (initiated by IBM => IEEE802.5 => FDDI):– token regenerated/modified at each node

– stations have two modes:• listen (forwards bits with delay 1)

• transmit (seizes the first token by transforming into the start of frame)


Token Ring


Token Ring (cont)• Idea

– Frames flow in one direction: upstream to downstream – special bit pattern (token) rotates around ring– must capture token before transmitting– release token after done transmitting

• immediate release• delayed release

– remove your frame when it comes back around– stations get round-robin service

• Frame Format

Control

8 8 8 24

CRCStart offrame

End offrame

Destaddr

Body

4848

Srcaddr

Status

32


Fiber Distributed Data Interface

• Evolution of IEEE802.5• Designed for fiber (100Mbps) but also supports coax

and twisted pair• Architecture: dual ring

– Tolerates one broken link or one station failure

• Stations buffer at least 9 bits and at most 80 bits• Uses 4B/5B encoding• Specific Timed-Token Algorithm


Token Times

• Token Holding Time (THT)– upper limit on how long a station can hold the token

• Token Rotation Time (TRT)– how long it takes the token to traverse the ring.– TRT <= ActiveNodes x THT + RingLatency

• Target Token Rotation Time (TTRT)– agreed-upon upper bound on TRT


Token Maintenance• Lost Token

– no token when initializing ring– bit error corrupts token pattern– node holding token crashes

• Generating a Token (and agreeing on TTRT)– execute when join ring or suspect a failure– send a claim frame that includes the node’s TTRT bid– when receive claim frame, update the bid and forward– if your claim frame makes it all the way around the ring:

• your bid was the lowest• everyone knows TTRT• you insert new token


Maintenance (cont)

• Monitoring for a Valid Token– should periodically see valid transmission (frame or token)– maximum gap = ring latency + max frame < = 2.5ms– set timer at 2.5ms and send claim frame if it fires


Ad Hoc Wireless Networks

• Physical transmission: spread-spectrum radio and diffused infrared

• Key issue: collision avoidance, as for the Ethernet– More complex than the Ethernet because the nodes are not

directly connected with one another

– Multi-hop network

• Potential problems:– Hidden terminal problem

– Exposed terminal problem


802.11: Multiple Access with Collision Avoidance (MACA 1990)

• MACA is designed for ad-hoc wireless networks• When a station S1 has a packet to transmit to station S2

– S1 senses the channel. If the channel is busy defers the transmission until idle

– if channel is idle S1 sends a special packet called Request-To-Send (RTS) to S2

– (if the RTS is correctly received by S2) S2 sends a Clear-To-Send (CTS), CTS includes the frame length

– (if the CTS is correctly received by S1) S1 starts the data transmission

• Stations which sense:– RTS: defer transmission until after CTS– CTS: defer transmission until the transmission of data completes

• If a station does not receive CTS in response to its RTS, it invokes an exponential backoff

Lecture 3, 1Spring 2003, COM1337/3501CCN: Direct Link Networks Direct Link Networks Textbook:...

Documents

Transcript of Lecture 3, 1Spring 2003, COM1337/3501CCN: Direct Link Networks Direct Link Networks Textbook:...