csci 183/183w/232direct link networks1 direct link networks encoding and framing error...

Csci 183/183W/232 Direct Link networks 1

Direct Link Networks• Encoding and Framing• Error Detection/Correction• Reliable Transmission will be covered

later (with TCP)• Media Access Control • (Wired) LAN Technologies• Readings

– Chapter 2 except Section 2.5, which is delayed to the discusson on Transport layer.


Encoding

• Signals propagate over a physical medium– modulate electromagnetic waves (on amplitude,

frequency, and phase): the process of encoding source data onto a carrier signal with frequency f_c

– e.g., vary voltage (amplitude)

• 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


Problems: Consecutive 1s or 0s

• Low signal (0) may be interpreted as no signal

• High signal (1) leads to baseline wander– Receiver keeps an average of the signal it has seen

so far and uses this average to distinguish between low and high signals

• Unable to recover clock– Clocks at the sender and receiver use signal

transition to synchronize each other


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)

• 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


Encodings (cont)

Bits

NRZ

Clock

Manchester

NRZI

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


Framing• Break sequence of bits into a frame –

why? determine where a frame starts and ends

• Typically implemented by network adaptor

Frames

BitsAdaptor Adaptor Node BNode A


Approaches• Sentinel-based

– delineate frame with special pattern: 01111110– e.g., HDLC, SDLC, PPP

– 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• What if the receiver sees a 1?

Header Body

8 16 16 8

CRCBeginningsequence

Endingsequence


Approaches (cont)• Couter-based

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

– problem: count field corrupted– solution: catch when CRC fails– How many frames are affected when counter or

sentinel is corrupted?

SY

N

Header Body

8 8 4214 168

SY

N

Cla

ss CRCCount


Approaches (cont)

• Clock-based– each frame is 125us long– e.g., SONET: Synchronous Optical Network

Overhead Payload

90 columns

9 rows

Special bit pattern


Handling Errors

• Data can be corrupted during transmission– Bit values changed

• Frame includes additional information– Set by sender– Checked by receiver

• Error-detection vs error-correction– Both need redundant information– Detection: error exists or not.– Correction: repair if there was an error

• Statistical guarantee


Error Detecting and Correcting Codes

• How many check/redundancy bits?– To detect single-bit error

• 1 bit: even/odd parity

– To correct a single-bit error in m-bit message• Need a minimum of r bits such that• (m + r + 1) 2r

• Example: 3-bit message needs 3-bit redundancy

• Correction or detection+retransmission?


Error Detection Techniques

• Checksum– Treat data as sequence of integers– Compute and send arithmetic sum– Detects some multiple bit errors not all– Internet Checksum

• Cyclic Redundancy Check– Mathematical function of data– More complex to compute– Handles more errors


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)


CRC Example

Generator 11011111100110011010000 Message1101

10011101

10001101

10111101

11001101

10001101

101 Remainder


Choice of C(x)

• Want to ensure C(x) doesn’t divide E(x)• We can detect

– All single-bit errors if• C(x) has at least 2 terms

– All double-bit errors if • C(x) doesn’t divide xj + 1

– X15 + x14 + 1 doesn’t divide xj + 1 for any j below 32768

– Any odd number of errors if• C(x) contains the factor x+1

– Any “burst” error of length less than or equal to k bits

– Most burst errors of length greater than k bits


Error Detection Summary

• To detect data corruption– Sender adds additional information to packet– Receiver checks

• Techniques– Parity bit– Checksum– Cyclic Redundancy Check


Hamming Error Correcting Code

• Achieve the theoretical lower limit of the number of redundant bits

• Correcting single-bit errors• Method:

– Put information bits at positions that are not equal to powers of two– Compute the check bits at positions that are power of two

• A check bit at position 2^j, where j=0, 1, …, checks an information bit at position i if and only if the binary representation of i contains a 1 at position j.

• The value of an “even” check bit equals the xor of all checked information bits

• At the receiver side, all checked bits and the check bit are xored and a zero indicates error free; otherwise, the sum of the positions of the check bits with a non-zero value indicates the bit that is in error

• Example– Information bits: 1001000– Coded bits: 00110010000


Error Recovery

• Reliable delivery over unreliable channel– How to recover from corrupted/lost packets

• Error detection and retransmission– With acknowledgements and timeouts– Also called Automatic Repeat Request (ARQ)– Retransmission incurs round trip delay

• Error correcting codes– Also called Forward Error Correction (FEC)– No sender retransmission required


Summary

• Encoding • Framing• Error correction/detection codes

– Parity/Checksum/CRC


Shared Media andLocal Area Networks

• Shared Media Access Problem– Media access control (MAC) and LAN– MAC addresses and network adaptors (NICs)– MAC protocols: random access vs. controlled access

• Ethernet• Token Ring and FDDI (read yourself)• 802.11 Wireless LAN (WiFi and WiMAX)

(next lecture)


Multiple Access Links and LANs

Two types of “links”:• point-to-point, e.g.,

– PPP for dial-up access, or over optical fibers

• broadcast (shared wire or medium), e.g.– traditional Ethernet– 802.11 wireless LAN


Typical LAN Structure

RAM

RAMROM

Ethernet Processor

• Transmission Medium

• Network Interface Card (NIC)

• Unique MAC “physical” address


Adaptors Communicating

• link layer implemented in “adaptor” (aka NIC), with “transceiver” in it– Ethernet card, dial-up

modem, 802.11 wireless card• sending side:

– encapsulates packet in frame– adds error checking bits, flow

control, reliable data transmission, etc.

• receiving side– looks for errors, flow

control, reliable data transmission, etc

– extracts packet, passes to receiving node

• data link & physical layers are closely coupled!

sendingnode

frame

rcvingnode

network packet

frame

adapter adapter

link layer protocol


MAC (Physical) Addresses• Addressing needed in shared media

– MAC (media access control) or physical addresses – To identify source and destination interfaces and get

frames delivered from one interface to another physically-connected interface (i.e., on same physical local area network!)

• 48 bit MAC address (for most LANs)– fixed for each adaptor, burned in the adapter ROM– MAC address allocation administered by IEEE

• 1st bit: 0 unicast, 1 multicast.• all 1’s : broadcast

• MAC flat address -> portability – can move LAN card from one LAN to another


MAC (Physical, or LAN) Addresses

MAC addressing operations on a LAN:• each adaptor on the LAN “sees” all frames• accept a frame only if dest. (unicast) MAC

address matches its own MAC address• accept all broadcast (MAC= all 1’s) frames• accept all frames if set in “promiscuous” mode• can configure to accept certain multicast

addresses (first bit = 1)


MAC Sub-layer

Data linklayer

802.3CSMA-CD

802.5Token Ring

802.2 Logical link control

Physicallayer

MAC

LLC

802.11Wireless

LAN

Network layer Network layer

Physicallayer

OSIIEEE 802

Various physical layers

OtherLANs


Broadcast Links: Multiple Access

Single shared communication channel• Only one can send successfully at a time• Two or more simultaneous transmissions

– interference!• How to share a broadcast channel

– media access control uses same shared media

• Humans use multi-access protocols all the time


Random Access• Stations contend for channels• Overlapping transmissions (collisions)

can occur– Carrier sensing?– Collision detection?

• Protocols– Aloha– Slotted Aloha– Carrier Sense Multiple Access: Ethernet


Controlled Access

• Stations reserve or are allocated channel– No collisions– Allocation: static or dynamic

• Protocols– Static channel allocation

• Time division multiple access– Demand adaptive channel allocation

• Reservation protocols• Token passing (token bus, token ring)


Taxonomy of MAC Protocols

WiFi (802.11)


Pure (unslotted) Aloha

• Simpler, no synchronization• Just send: no waiting for beginning of slot


Slotted Aloha• Time is divided into equal size slots• Nodes transmit at the beginning of a slot• If collision, retransmit later

Success (S), Collision (C), Empty (E) slots


Performance of Aloha Protocols

G = offered load = Np0.5 1.0 1.5 2.0

0.1

0.2

0.3

0.4

Pure Aloha

Slotted Aloha

S =

th

rou

gh

pu

t =

“g

ood

pu

t”

(su

ccess

rate

)


Carrier Sense Multiple Access

• Aloha is inefficient (and rude)– Doesn’t listen before talking

• CSMA: Listen before transmit– If channel idle, transmit entire packet– If busy, defer transmission

• How long should we wait?– Human analogy: don’t interrupt others

• Can carrier sense avoid collisions completely?


Persistent and Non-persistent CSMA

• Non-persistent– If idle, transmit– If busy, wait random amount of time– If collision, wait random amount of time

• p-persistent– If idle, transmit with probability p– If busy, wait till it becomes idle– If collision, wait random amount of time


CSMA/CD

CSMA with collision detection (CD)• Listen while talking• Stop transmitting when collision detected

– Compare transmitted and received signals– Save time and bandwidth– Improvement over persistent and nonpersistent

protocols

• Human analogy– Polite conversationalist

• Worst case time to detect a collision?


CollisionsA B

A B

A B

A B


Worst Case Collision Detection Time


CSMA/CD: Illustration


Ethernet Overview• History

– developed by Xerox PARC in mid-1970s– roots in Aloha packet-radio network– standardized by Xerox, DEC, and Intel in 1978– similar to IEEE 802.3 standard

• CSMA/CD– carrier sense– multiple access– collision detection

• Frame Format

Destaddr

64 48 32

CRCPreamble Srcaddr

Type Body

1648


Ethernet“Dominant” LAN technology: • cheap $20 for 100Mbs!• first widely used LAN technology• Simpler, cheaper than token ring LANs and ATM• Kept up with speed race: 10, 100, 1000 Mbps

Metcalfe’s Ethernetsketch


Ethernet Frame FormatSending adapter encapsulates IP datagram (or other

network layer protocol packet) in Ethernet frame DIX frame format

IEEE 802.3 format

Destaddr

8 bytes 6 4

CRCPreamble Srcaddr Type Data

26 0-1500

Destaddr

8 bytes 6 4

CRCPreamble Srcaddr Length Data

26 0-1500

• Ethernet has a maximum frame size: data portion <=1500 bytes• It imposes a minimum frame size: 64 bytes (excluding preamble) If data portion <46 bytes, pad with “junk” to make it 46 bytes Q: Why minimum frame size in Ethernet?


Fields in Ethernet Frame Format• Preamble:

– 7 bytes with pattern 10101010 followed by one byte with pattern 10101011 (SoF: start-of-frame)

– used to synchronize receiver, sender clock rates, and identify beginning of a frame

• Addresses: 6 bytes– if adapter receives frame with matching destination address,

or with broadcast address (eg ARP packet), it passes data in frame to network layer protocol (specified by TYPE field)

– otherwise, adapter discards frame

• Type: indicates the higher layer protocol, mostly IP but others may be supported such as Novell IPX and AppleTalk– 802.3: Length gives data size; “protocol type” included in

data

• CRC: checked at receiver, if error is detected, the frame is simply dropped


IEEE 802.3 MAC: EthernetMAC Protocol:• CSMA/CD• Truncated binary exponential backoff

– for retransmission n: 0 < r < 2k * ASlotTime, where k=min(n,10)

– give up after 16 retransmissions

• Slot Time is the critical system parameter– upper bound on time to detect collision– upper bound on time to acquire channel– upper bound on length of frame segment generated

by collision– quantum for retransmission scheduling– max{round-trip propagation, MAC jam time}


IEEE 802.3 Parameters

• 1 bit time = time to transmit one bit– 10 Mbps 1 bit time = 0.1 s

• Maximum network diameter 2.5km– Maximum 4 repeaters

• “Collision Domain” – Distance within which collision can occur and be detected – IEEE 802.3 specifies:

worst case collision detection time: 51.2 s

• Slot time– 51.2 s = 512 bits = 64 bytes

• Why minimum frame size?– 51.2 s => minimum # of bits can be transited at

10Mpbs is 512 bits => 64 bytes is required for collision detection


Ethernet MAC Protocol: Basic Ideas

1-persistent CSMA/CD• Carrier sense: station listens to channel first

– Listen before talking

• If idle, station may initiate transmission– Talk if quiet

• Collision detection: continuously monitor channel– Listen while talking

• If collision, stop transmission– One talker at a time

• Exponential binary back-off algorithm


Ethernet CSMA/CD Alg. Flow Chart


Ethernet CSMA/CD Algorithm

1. Adaptor gets datagram and creates frame

2. If adapter senses channel idle, it starts to transmit frame. If it senses channel busy, waits until channel idle and then transmits

3. If adapter transmits entire frame without detecting another transmission, the adapter is done with frame ! Signal to network layer “transmit OK”

4. If adapter detects another transmission while transmitting, aborts and sends jam signal

5. After aborting, adapter enters exponential backoff: after the mth collision, adapter chooses a K at random from {0,1,2,…,2m-1}. Adapter waits K*512 bit times and returns to Step 2

6. Quit after 16 attempts, signal to network layer “transmit error”


Ethernet’s CSMA/CD (more)

Jam Signal: make sure all other transmitters are aware of collision; 32 bits;

Bit time: .1 microsec for 10 Mbps Ethernet ;for K=1023, wait time is about 50 msec

Exponential Backoff: • Goal: adapt retransmission

attempts to estimated current load– heavy load: random wait

will be longer

• first collision: choose K from {0,1}; delay is K x 512 bit transmission times

• after second collision: choose K from {0,1,2,3}…

• after ten collisions, choose K from {0,1,2,3,4,…,1023}


CSMA/CD EfficiencyRelevant parameters

– cable length, signal speed, frame size, bandwidth

• tprop = max prop between 2 nodes in LAN• ttrans = time to transmit max-size frame

• Efficiency goes to 1 as tprop goes to 0• Goes to 1 as ttrans goes to infinity• Much better than ALOHA, but still decentralized, simple, and cheap

transprop tt /51

1efficiency


IEEE 802.3 Physical Layer

(a) transceivers(b)

10base5 10base2 10baseT 10baseFX

Medium Thick coax Thin coaxTwisted

pairOptical fiber

Max. Segment Length

500 m 200 m 100 m 2 km

Topology Bus Bus StarPoint-to-point link

IEEE 802.3 10 Mbps medium alternatives

Thick Coax: Stiff, hard to work with T connectors flaky

Hubs & Switches!


Ethernet Technologies: 10Base2• 10: 10Mbps; 2: under 200 meters max cable length• thin coaxial cable in a bus topology

• repeaters used to connect up to multiple segments• repeater repeats bits it hears on one interface to its other interfaces: physical layer device only!• has become a legacy technology


10BaseT • 10 Mbps rate• T stands for Twisted Pair• Nodes connect to a hub: “star topology”; 100 m max distance between nodes and hub

• Hubs are essentially physical-layer repeaters:– bits coming in one link go out all other links– no frame buffering– no CSMA/CD at hub: adapters detect collisions– provides net management functionality

hub

nodes


Ethernet Hubs & Switches

(a)

Single collision domain(b)

High-Speed backplane or interconnection fabric

Twisted Pair CheapEasy to work withReliableStar-topology CSMA-CD

Twisted Pair CheapBridging increases scalabilitySeparate collision domainsFull duplex operation


Evolution of EthernetFrom early 80’s 10Base Ethernet to 90’s

100Base (Fast) Ethernet to today’s Gigabit Ethernet to 10 Gigabit

Ethernet, ……

IEEE 802.3 Original Parameters• transmission Rate: 10 Mbps• Min Frame: 512 bits = 64 bytes• slot time: 512 bits/10 Mbps = 51.2 sec

– 51.2 sec x 2x105 km/sec =10.24 km– 5.12 km round trip distance

• max Length: 2500 meters + 4 repeaters

• For compatibility, desire to maintain same frame format!– Each x10 increase in bit rate, must be accompanied by x10

decrease in distance ?!


100Base T (Fast) Ethernet: Issues• 1 bit time = time to transmit one bit

– 100 Mbps 1 bit time = 0.01 s• If we keep the same “collision domain”, i.e., worst case collision detection time kept at 51.2 s

Q: What will be the minimum frame size?– 51.2 s => minimum # of bits can be transited at

100Mpbs is 5120 bits => 640 bytes is required for collision detection

– This requires change of frame format and protocol!

• Or we can keep the same minimum frame size, but reduce “collision domain” or network diameter!• slot time from 51.2 s to 5.12 s!• maximum network diameter 100 m!


Fast (100Mbps) Ethernet

100baseT4 100baseT 100baseFX

MediumTwisted pair category 3

UTP 4 pairsTwisted pair category 5

UTP two pairs

Optical fiber multimode

Two strands

Max. Segment Length

100 m 100 m 2 km

Topology Star Star Star

IEEE 802.3 100 Mbps Ethernet medium alternatives

To preserve compatibility with 10 Mbps Ethernet:• Same frame format, same interfaces, same protocols• Hub topology only with twisted pair & fiber• Bus topology & coaxial cable abandoned• Category 3 twisted pair (ordinary telephone grade) requires 4

pairs• Category 5 twisted pair requires 2 pairs (most popular)• Most prevalent LAN today


Gigabit Ethernet

Gigabit Ethernet Physical Layer Specification(IEEE 802.3 1 Gigabit Ethernet medium alternatives)

1000baseSX 1000baseLX 1000baseCX 1000baseT

MediumOptical fiber multimode

Two strands

Optical fiber single modeTwo strands

Shielded copper cable

Twisted pair category 5

UTP

Max. Segment Length

550 m 5 km 25 m 100 m

Topology Star Star Star Star


Gigabit Ethernet• use standard Ethernet frame format• allows for point-to-point links and shared broadcast

channels• in shared (half-duplex hub) mode, CSMA/CD is used

- slot time increases to 512 bytes- small frames need to be extended to 512 B- carrier extension:– frame bursting: allow stations to transmit burst of

short frames• Commonly used today: Gigabit switches!

– Full-Duplex at 1 Gbps for point-to-point links– Frame structure preserved but CSMA-CD essentially

abandoned– Extensive deployment in backbone of enterprise data

networks and in server farms


Carrier Extension

RRRRRRRRRRRRRFrame

Carrier Extension

512 bytes

• For 10BaseT : 2.5 km max; slot time = 64 bytes• For 1000BaseT: 200 m max; slot time = 512 bytes• Carrier Extension : continue transmitting control

characters [R] to fill collision interval.• This permits minimum 64-byte frame to be

handled.• Control characters discarded at destination.• For small frames net throughput is only slightly

better than Fast Ethernet.


Frame Bursting

512 bytes

ExtensionFrame Frame Frame Frame

Frame burst

• Source sends out burst of frames without relinquishing control of the network.

• Uses Ethernet Interframe gap filled with extension bits (96 bits)

• Maximum frame burst is 8192 bytes• Three times more throughput for small frames.


10 Gigabit Ethernet IEEE 802.3 10 Gbps Ethernet medium alternatives

10GbaseSR 10GBaseLR 10GbaseEW 10GbaseLX4

Medium

Two optical fibersMultimode at 850 nm

64B66B code

Two optical fibers

Single-mode at 1310 nm

64B66B

Two optical fibers

Single-mode at 1550 nmSONET compatibility

Two optical fibers multimode/single-mode with four wavelengths at 1310 nm band8B10B code

Max. Segment Length

300 m 10 km 40 km 300 m – 10 km

• Frame structure preserved• CSMA-CD protocol officially abandoned• LAN PHY for local network applications• WAN PHY for wide area interconnection using SONET OC-192c • Extensive deployment in metro networks anticipated


Server

100 Mbps links

10 Mbps links

ServerServer

Server

100 Mbps links

10 Mbps links

Server

100 Mbps links

10 Mbps links

Server

Gigabit Ethernet links

Gigabit Ethernet links

Server farm

Department A Department B Department C

Hub Hub Hub

Ethernet switch

Ethernet switch

Ethernet switch

Switch/router Switch/router

Typical Ethernet Deployment


Ethernet Summary

• 1-persistent CSMA/CD• 51.2 s to seize the channel• Collision not possible after 51.2 s• Minimum frame size of 64 bytes• Binary exponential backoff• Works better under light load• Delivery time non-deterministic


Summary (continued)

• Address– Unique number assigned to station– Put in frame header– Recognized by hardware

• Address forms– Unicast– Broadcast– Multicast


Summary (continued)

• Type information– Describes data in frame– Set by sender– Examined by receiver

• Frame format– Header contains address and type information– Payload contains data being sent


Summary (continued)• LAN technologies

– Ethernet (bus)– Token Ring– FDDI (ring)– Wireless 802.11

• Wiring and topology– Logical topology and Physical topology (wiring)– Hub allows

• Star-shaped bus• Star-shaped ring

csci 183/183w/232direct link networks1 direct link networks encoding and framing error...

Documents