Data Link Layer

Goals:

13: Link Layer, Multiple Access Protocols

Last Modified: 4/9/2003 2:32:56 PM

r understand principles

Overview:

behind data link layer services:

m

r link layer services

r error detection, correction r multiple access protocols and r link layer addressing, ARP
m m m m m

m m

sharing a broadcast channel: multiple access link layer addressing error detection, correction

LANs

r specific link layer technologies: Ethernet hubs, bridges, switches IEEE 802.11 LANs PPP ATM

r instantiation and

implementation of various link layer technologies

5: DataLink Layer

5 a-1

5: DataLink Layer

5 a-2

Link Layer: setting the context

r two physically connected devices: m host-router, router-router, host-host r unit of data: frame

Link Layer
r r

Node-to-node connectivity Point-to-point or multiple access

m m

Multiple access requires addressing Both require rules for sharing the links

M Ht M Hn Ht M Hl Hn Ht M

application transport network link physical

r
data link protocol phys. link adapter card
5: DataLink Layer 5 a-3

Examples:
m m

network link physical

Hl Hn Ht M frame

Point-to-point (single wire, e.g. PPP, SLIP) Broadcast (shared wire or medium; e.g, Ethernet or wireless) m Switched (e.g., switched Ethernet, ATM etc)

5: DataLink Layer

5 a-4

Communication Technologies

Data Model?
Packet Mode bursty discrete transmissions r Circuit Mode continuous traffic
r

Wired LANs, Wireless LANs (RF or light), Cellular Telephones, Satellites, Packet Radio, Wired Telephone, Voice

5: DataLink Layer

5 a-5

5: DataLink Layer

5 a-6

Basics of Link Layer

r r

Multiple Access
Multiple Access - fundamental to communication r Two or more communicators use a shared medium to share information r Multiple Access Protocol - Rule for sharing medium to facilitate communication?
r
m

Multiple Access Protocols Error Detection/Correction

Claim: humans use multiple access protocols all the time

5: DataLink Layer 5 a-8

Can simultaneous transmissions cause interference?

5: DataLink Layer

5 a-7

Multiple Access protocols

r Algorithm that determines how stations share channel,

MAC Protocols: a taxonomy

Three broad classes: r Channel Partitioning
m m

i.e., determine when station can transmit r Note: communication about channel sharing must use channel itself! (or be agreed upon ahead of time) r what to look for in multiple access protocols: synchronous or asynchronous m information needed about other stations
m

divide channel into smaller pieces (time slots, frequency) allocate piece to node for exclusive use

r Random Access
m m

allow collisions recover from collisions

robustness (e.g., to channel errors) m performance

m

r Polling Style m tightly coordinate shared access to avoid collisions

Goal: efficient, fair, simple, decentralized

5: DataLink Layer 5 a-9 5: DataLink Layer 5 a-10

Channel Partitioning : TDMA

TDMA: time division multiple access
r access to channel in "rounds" r each station gets fixed length slot (length = pkt

Channel Partitioning : FDMA

FDMA: frequency division multiple access
r channel spectrum divided into frequency bands m each station assigned fixed frequency band r unused transmission time in frequency bands go idle r example: 6-station LAN, 1,3,4 have pkt, frequency

trans time) in each round

r unused slots go idle r example: 6-station LAN, 1,3,4 have pkt, slots 2,5,6

bands 2,5,6 idle

5: DataLink Layer 5 a-11

frequency bands

idle

time

5: DataLink Layer 5 a-12

Channel Partitioning: CDMA

CDMA (Code Division Multiple Access)
r unique code assigned to each user; ie, code set partitioning r used mostly in wireless broadcast channels (cellular,

TDMA vs FDMA
r In TDMA, each station gets the whole channel spectrum

some of the time

r In FDMA, each station gets part of the channel

r all users share same frequency, but each user has own

satellite,etc)

spectrum all of the time

chipping sequence (ie, code) to encode data

r In CDMA, each station is assigned a code that

r encoded signal = (original data) X (chipping sequence)

r decoding: inner-product of encoded signal and chipping r allows multiple users to coexist and transmit

sequence

determines what portions of the channel spectrum they use and for how long to avoid collision with others when and in what way!
m

r All require lots of coordination about who speaks What if didnt want to coordinate things so tightly?

simultaneously with minimal interference (if codes are orthogonal)

5: DataLink Layer 5 a-13

5: DataLink Layer 5 a-14

Random Access protocols

r

Random Access Protocols

r When node has packet to send m transmit at full channel data rate R. m no a priori coordination among nodes r two or more transmitting nodes -> collision, r random access MAC protocol specifies: m how to detect collisions m how to recover from collisions (e.g., via delayed retransmissions) r Examples of random access MAC protocols: m slotted ALOHA m ALOHA m CSMA and CSMA/CD (Ethernet) m Remember Ethernet grew out of technology for broadcast in Hawaiian Islands?
5: DataLink Layer 5 a-15 5: DataLink Layer 5 a-16

Random access protocols are alternative to tight coordination

When want to transmit, transmit and hope for the best m If bad things happen, protocol says how to recover
m

Random Access: Slotted Aloha

r time is divided into equal size slots (= pkt trans. time) r node with new arriving pkt: transmit at beginning of

Slotted Aloha efficiency

Q: what is max fraction slots successful?
A: Suppose N stations have packets to send m each transmits in slot with probability p m prob. successful transmission S is:
by single node:

next slot

r if collision: retransmit pkt in future slots with

probability p, until successful.

S= (prob it sends) * (prob all others do not) = p (1-p)(N-1)

by any of N nodes S = Prob (only one transmits) = N p (1-p)(N-1) = 1/e = .37 as N -> infty Success (S), Collision (C), Empty (E) slots
5: DataLink Layer 5 a-17

choosing optimum p as n -> infty ...

At best: channel use for useful transmissions 37% of time!

5: DataLink Layer 5 a-18

Random Access: Pure (unslotted) ALOHA

r unslotted Aloha: simpler, no synchronization r pkt needs transmission: m send without awaiting for beginning of slot r collision probability increases: m pkt sent at t 0 collide with other pkts sent in [t0-1, t 0+1]

Pure Aloha (cont.)

P(success by given node) = P(node transmits) . P(no other node transmits in [p0-1,p0] . P(no other node transmits in [p0-1,p0] = p . (1-p) . (1-p) P(success by any of N nodes) = N p . (1-p) . (1-p)
S = throughput = goodput (success rate)
0.4 0.3

choosing optimum p as n -> infty ...

= 1/(2e) = .18
Slotted Aloha

0.2 0.1

protocol constrains effective channel throughput!

Pure Aloha
0.5

5: DataLink Layer 5 a-19

G = offered load = Np

1.0

1.5

2.0 5: DataLink Layer 5 a-20

CSMA: Carrier Sense Multiple Access

CSMA: listen before transmit: r If channel sensed idle: transmit entire pkt
r If channel sensed busy, defer transmission

CSMA collisions
collisions can occur:
propagation delay means two nodes may not year hear each others transmission

spatial layout of nodes along ethernet

Persistent CSMA: retry immediately with probability p when channel becomes idle (may cause instability) m Non-persistent CSMA: retry after random interval
m

collision:

entire packet transmission time wasted

r human analogy: dont interrupt others!

note:

role of distance and propagation delay in determining collision prob.

5: DataLink Layer 5 a-21 5: DataLink Layer 5 a-22

CSMA/CD (Collision Detection)

CSMA/CD: carrier sensing, deferral as in CSMA
m m m

CSMA/CD collision detection

collisions detected within short time colliding transmissions aborted, reducing channel wastage persistent or non-persistent retransmission

r collision detection: m easy in wired LANs: measure signal strengths, compare transmitted, received signals m difficult in wireless LANs: receiver shut off while transmitting r human analogy: if start talking at same time some one

else does dont just continue talking

5: DataLink Layer 5 a-23

5: DataLink Layer 5 a-24

Compromise? Polling Style MAC protocols

channel partitioning MAC protocols: m share channel efficiently at high load
m

Polling style MAC protocols

Polling:
r master node

Token passing:
r control token passed from

inefficient at low load: delay in channel access, 1/N bandwidth allocated even if only 1 active node!

invites slave nodes to transmit in turn Clear to Send msgs

one node to next sequentially.

r Request to Send, r concerns: m polling overhead m latency m single point of failure (master)

r token message r concerns: m token overhead m latency m single point of failure (token)

Random access MAC protocols m efficient at low load: single node can fully utilize channel m high load: collision overhead Polling style protocols (taking turns) look for best of both worlds!
5: DataLink Layer 5 a-25

5: DataLink Layer 5 a-26

Reservation-based protocols
Distributed Polling:
r time divided into slots
m

Summary of MAC protocols

r

What do you do with a shared media?

m m

r begins with N short reservation slots

Channel Partitioning, by time, frequency or code

Time Division,Code Division, Frequency Division

reservation slot time equal to channel end-end propagation delay m station with message to send posts reservation m reservation seen by all stations r after reservation slots, message transmissions ordered by known priority

Random access

ALOHA, S -ALOHA, CSMA, CSMA/CD carrier sensing: easy in some technologies (wire), hard in others (wireless) CSMA/CD used in Ethernet polling from a central cite, token passing

Polling Style

5: DataLink Layer 5 a-27

5: DataLink Layer 5 a-28

Basics of Link Layer

Multiple Access Protocols r Error Detection/Correction
r

Error Detection
EDC= Error Detection and Correction bits (redundancy) D = Data protected by error checking, may include header fields Error detection not 100% reliable! protocol may miss some errors, but rarely larger EDC field yields better detection and correction

5: DataLink Layer 5 a-29

5: DataLink Layer 5 a-30

Smart Redundancy
In general, more bits of redundancy the stronger the error detection/correction abilities but smart redundancy r What if transmitted another copy of the same thing?
r
m

Recall: Internet checksum

We saw this a bunch of times in upper layers is this a good choice for the link layer? Sender:
r treat segment contents as sequence of 16-bit integers r checksum: addition (1s complement sum) of segment

How many bits till not detected? Ability to correct?

r sender puts checksum value into UDP checksum field

contents

Can we do better than that with less space?

5: DataLink Layer 5 a-31

Receiver: r compute checksum of received segment r check if computed checksum equals checksum field value

5: DataLink Layer 5 a-32

Intelligent choice for link layer?

r

Example: Parity
Single Bit vs Two Dimensional Two Dimensional Bit Parity : Detect and correct single bit errors Bit Parity: Example of using Want even number of 1s in each dimension redundant bits intelligently for increased error detection/correction capability!

Tailored to type and frequency of errors expected in the specific technology being used
Some error m Some error
m

technologies (like fiber) have very low rates technologies (like wireless) have high rates

Single Bit Parity:

Detect single bit errors

How to we tailor the number of bits to use and *how* we use them to get the desired effect??
5: DataLink Layer 5 a-33

5: DataLink Layer 5 a-34

Beyond parity?
How can we generalize this example of single vs double bit parity? r Is there a theory of using redundant bits efficiently based on the types of errors we expect to find? r Cyclic Redundancy Checks (CRC) views both the data and the redundant bits as binary polynomials and ensures that they satisfy a certain mathematical relationship
r
5: DataLink Layer 5 a-35

Checksumming: Cyclic Redundancy Check

r view data bits, D, as a binary number or binary polynomial m 101011= X^5+X^3+X^1+X^0 = X^5+X^3+X+1. r choose r+1 bit pattern/polynomial (generator), G

r goal: choose r CRC bits, R, such that m <D, R> = D* 2r XOR R (shift D over place R in the end) m <D,R> exactly divisible by G (modulo 2) m receiver knows G, divides <D,R> by G. If non -zero remainder: error detected! m can detect all burst errors less than r+1 bits r widely used in practice (ATM, HDCL)

5: DataLink Layer 5 a-36

CRC Example
Want: D.2r XOR R = nG equivalently: if we divide D.2r by G, want reminder R

Common CRC Polynomials (G)

r CRC-12 used for transmission of streams of 6-bit

characters and generates 12-bit FCS

m

CRC-12: X^12+X^11+X^3+X^2+X+1

r Both CRC-16 and CCRC-CCITT are used for 8 bit

transmission streams and both result in 16 bit FCS. Considered to give adequate protection for most applications.
m m

R = remainder[

D.2r ] G
5: DataLink Layer 5 a-37

r CRC-32 gives extra generates 32 bit FCS. Used by

CRC-16: X^16+X^15+X^2+1 (USA) CRC-CCITT: X^16+X^12+X^5+1 (Europe)

the local network standards committee (IEEE802) and in some DOD applications.
m

CRC-32: X^32+X^26+X^23+X^22+X^16+X^12+X^11+X^10+X^8+X ^7+X^5+X^4+X^2+X+1 5: DataLink Layer

5 a-38