1

CSCD 330Network ProgrammingSpring 2014

Lecture 17Link Layer

Hardware and Protocols

Reading: Chapter 5 in text

Some slides provided courtesy of J.F Kurose and K.W. Ross, All Rights Reserved, copyright 1996-2007

Who is this?

Robert (Bob) Metcalf – Inventor of Ethernet

Bob Metcalfe Helped build early Internet while still an

undergraduateInvented Ethernet—while at Xerox Palo Alto

Research CenterFounded 3Com Corp.Served as publisher at IDG's Infoworld

Publishing Co.Wrote three booksSince January 2001 has been a venture capitalist

with Polaris Ventures

Nice article on him … if you want to read ithttp://electronicdesign.com/Articles/Index.cfm?

AD=1&ArticleID=2855

2

3

Introduction to Link LayerSo far,

Application Layer - applications interface to people

Messages get delivered Transport Layer – application to application

Segments get delivered Network Layer – destination to destination

Datagrams get delivered Link Layer – node to node delivery

Frames are delivered

Link Layer Overview

• Main question at this layer• How do I deliver packets or frames

between two nodes in the network?

• Historical reasons look at other link layer schemes• But, Ethernet is currently dominant

technology 802.11 Wireless another protocol Covered in a different chapter, in a different course and at a future time in a galaxy far away ... 4

5

Link Layer

• 5.1 Introduction and services

• 5.1a NIC Cards• 5.2 Error detection and

correction • 5.3Multiple access

protocols• 5.4 Link-layer

Addressing• 5.5 Ethernet

6

Link Layer - IntroductionLink Layer Terms• Both Hosts and routers• Called nodes

• Communication channels connect adjacent nodes along communication path, links• Wired links• Wireless links

• Layer-2 packet is a frame encapsulates datagramData-link layer responsible for

transferring a frame from one node to adjacent node over a link

Link Layer Services

• The services offered by Link Layer are ....• Framing• Link Access• Reliable Delivery• Flow Control• Error Detection and Correction• Half dupleax and full duplex

7

8

Link Layer Services

• Framing• Encapsulate datagram into frame, add

header, trailer• MAC (Media Access Control) addresses used

in frame headers to identify source, destination• Example: 00:13:02:BA:43:56

• Reliable Delivery Between Adjacent Nodes• Wireless links: higher error rates,• Does more for reliability than wired• Does compute errors

• Wired relies on upper layers to handle it• Channel Access• Share channel if shared medium• Point to point if not shared

9

Reliable DeliveryQuestion on ReliabilitySo, why would you need both link-level and transport layer reliability?• Because dealing with reliability at two levels!• Transport layer reliability • For end-to-end delivery between two processes

• Link layer reliability • For delivery between two nodes connected by a single link

10

Link Layer Services Continued• Flow control• Pacing between adjacent sending and

receiving nodes• Error detection• Errors caused by signal attenuation,

noise• Receiver detects presence of errors: • Signals sender for retransmission or drops frame

• Error correction• Receiver identifies and corrects bit

error(s) without resorting to retransmission

• Half-duplex and full-duplex• With half duplex, nodes at both ends of

link can transmit, but not at same time

Link Layer Implementation• Where is the link layer

implemented?• Hardware or software?• Separate card or on chip?

• Mostly Implemented• Hardware• Network Interface Card (NIC)

11

Link Layer Implementation

• Alternative to actual NIC card ...• Built in to the motherboard

13

Where is the link layer implemented?• In each host• Link layer implemented

in “Network adaptor”

Network Interface Card (NIC)• Ethernet card, PCMCI

card, 802.11 card• Implements link and

physical layers

• Attaches to host’s system buses

• Combination of hardware, software, firmware

controller

physical

transmission

cpu memory

host bus (e.g., PCI)

network adaptercard

host schematic

applicationtransportnetwork

link

linkphysical

14

Network Interface Cards (NIC) Basics

• A Network Interface Card (NIC) • Establishes link between computer and

network, and manages link

• A NIC performs two crucial tasks:1. Establishes and manages computer’s

network connection2. Encodes and translates digital data into

signals for outgoing messages, and Translates signals into digital computer

data for incoming messages

15

Parallel vs. Serial Transmission

• NICs also manage transformation of network data’s form• Computer bus has series of parallel data lines

(Parallel transmission)• Signals traversing network media consist of a

linear sequence of bits of data (serial transmission)

• NIC takes outgoing transmission in parallel form and recasts them into their serial equivalents For incoming messages, process reverses• Converting between serial and parallel

transmission , NIC has memory, to hold data temporarily

16

Parallel vs. Serial Transmission

• Bus width - number of parallel lines in a computer bus• Industry Standard Architecture (ISA) bus – 16-bit

obsolete• Peripheral Component Interconnect (PCI) bus –

32-bit, 64-bit , faster

Bus width

17

Network Interface Card

• Heart of card is a special purpose processor chip• Does framing, link access, flow control and

error detection• Encryption too, for wireless• Examples• Intel 8254x controller implements

Ethernet,• Atheros AR5006 controller handles 802.11

protocol• Trend ... Integrate adapters into

motherboard, not separate cards anymore• Particularly true for laptops and phones

devices

18

Link Layer

• 5.1 Introduction and services

• 5.2 Error detection and correction - Later

• 5.3Multiple access protocols

• 5.4 Link-layer Addressing

• 5.5 Ethernet

Error Detection and CorrectionCauses of Transmission Errors• Signal noise, attenuation or distortion• Both on wired and wireless mediums• See next slide ...

Definitions

• What is Signal Attenuation?• Loss of signal strength measured in decibels (dB)

dec•i•bel - Defined a unit used to express differences in power, in

acoustics or electronics: equal to ten times the common logarithm of the ratio of two signals.

• Occurs for several reasons• Range - both wireless and wired transmissions

gradually dissipate in strength over longer distance• Interference - on wireless networks, radio

interference or physical obstructions like walls dampen communication signals

• Wire size - on wired networks, thinner wires suffer from higher (more) attenuation than thicker wires

Errors at Link Layer

Two Ideas for Errors at Link Layer

Error detection

– Parity checks, cyclic redundancy codes, checksum

Error correction

– Send redundant information with frame

– When receiving frame incorrectly, receiver makes “educated guess” about the original frame

Error Detection

• Error Detection• Lets … Receiver• Determine whether received message has

been corrupted• Checksum simplest method• Checksum, function of message, gets

appended to message• Receiver uses same function to

calculate checksum of received message and compare it with appended checksum

DataLink Layer 23

Example of ChecksumInternet Checksum Goal: Detect “errors” (e.g., flipped bits) in

transmitted packet (note: used at transport layer only)

Sender: Treat segment contents as sequence of 16-bit

integers Checksum: Addition (1’s complement sum) of

segment contentsOnly done on IP packet header

Receiver: Compute checksum of received segment Check if computed checksum equals checksum

field value: NO - error detected YES - no error detected

24

Error Detection

• Parity Checking – Even Parity • If you get odd number of bits including parity bit– > A bit changed in transmission• Single parity bit is an error

detecting code,• Not an error correcting code • No way to determine which bit is corrupted

• Two-dimensional parity checks parity of both column and row, pinpoints error and allows for correction

25

Parity Checking Simplest TechniqueSingle Bit ParityDetect single bit errors

Two Dimensional Bit ParityDetect and correct single bit errors

Simple detection

For Even Parity

Number of 1’s in data + parity bit must be an Even number

Can locate and correct errors

1

9 1’s Add a 1

26

Parity Checking

• Comment • Is a limit to parity schemes• A parity bit only guaranteed to detect

an odd number of bit errors One, three, five, and so forth • If even number of bits (two, four, six

and so on) are flipped, the parity bit appears to be correct, even though the data is corrupt• This is for even parity• Odd parity will have the reverse problem

27

Checksumming Cyclic Redundancy Check (CRC)• Want algorithm detects some number of

changed bits• Ideally ... implemented in hardware• CRC Scheme• CRC algorithms treat message as an

enormous binary number,• Divide it by another fixed binary

number,• Make remainder from this division the

checksum• Widely used in practice (Ethernet, 802.11

WiFi, ATM)• It has better powers of detection than bit

parity or checksums … uses entire number, handle more bit errors – nice example belowhttp://www.repairfaq.org/filipg/LINK/F_crc_v3.html

28

Link Layer

• 5.1 Introduction and services

• 5.2 Error detection and correction

• 5.3 Multiple access protocols

• 5.4 Link-layer Addressing

• 5.5 Ethernet

29

Access Links and Protocols

Two types of “links”• Point-to-point

• PPP for dial-up access• Point-to-Point link between Ethernet switch and

host

• Broadcast (shared wire or medium)• Old-fashioned Ethernet• 802.11 wireless LAN• Satellite

shared wire (e.g., cabled non-switch Ethernet)

shared RF (e.g., 802.11 WiFi)

shared RF(satellite)

humans at acocktail party

(shared air, acoustical)

Communication Between Nodes

• Point to Point Links• These are dedicated channels between a

node and switch or modem• No need to share the link

• Broadcast Links• Multiple sending and receiving nodes all

connected to single broadcast channel• Central problem at link layer• How to “share” the link between nodes• Must coordinate access between

competing nodes• Known as Multiple Access Problem

Multiple Access Problem

• Question is• How do you fairly share the medium at the

Link level?• Everyone is connected to same network• All want to send their data all the time.• How do we fairly share the medium?

32

Multiple Access Protocols• Idea is .... can send and receive in both

directions single shared broadcast channel • Two or more simultaneous transmissions by

nodes means interference • Collision if node receives two or more

signals at the same time … gibberish, must retransmit, BAD!

Multiple Access Protocol• Distributed algorithm, determines how nodes

share channel, i.e., determine when node can transmit

• Communication about channel sharing must use channel itself! • No out-of-band channel for coordination

33

Ideal Multiple Access Protocol

Rules - Broadcast Channel, Rate = R bps1. When one node wants to transmit, can send

at full rate R2. When M nodes want to transmit, each can

send at average rate R/M ... Fair3. Fully decentralized• No special node to coordinate

transmissions• No synchronization of clocks, slots

4. Simple

In a perfect world ...

Examine Following Strategies

• Look at each of the strategies and see how close each meets the IDEAL Multiple Access Channel properties ...

35

Multiple Access Protocols: A TaxonomyThree broad classes1.Channel Partitioning• Divide channel into smaller pieces• Time slots, frequency, code division

• Allocate piece to node for exclusive use

2. Random Access• Channel not divided, allow collisions• Recover from collisions

3. Taking turns• Nodes take turns, but nodes with more to send can take longer turns

Channel Partitioning

37

Channel Partitioning Protocols: TDMATDMA: Time Division Multiple Access • Access to channel in "rounds" • Each station gets fixed length slot

Length = packet transit time, in each round• Whats the problem?• Unused slots wasted• Example: 6-station LAN, 1,3,4 have pkt, slots 2,5,6 idle 1 3 4 1 3 4

6-slotframe

38

TDMA: Time Division Multiple Access

• Analogy• Take turns talking• Fix each person a time slot• All get a turn even when they have

nothing to say

• Our class – 5 slot limit

Marty, Laura Kevin Calvin Chris

Talks Silent Talks Silent Talks

Wasted two slots where others could have had useful things to say …

39

Channel Partitioning Protocols: FDMA

FDMA: frequency division multiple access

• Channel spectrum divided into frequency bands

• Each station assigned fixed frequency band

• Whats the problem? Some frequency bands are idle, limited

to your band• Example: 6-station LAN, 1,3,4 have pkt,

frequency bands 2,5,6 idle fre

qu

en

cy

ban

ds

time

FDM cable

Random Access Protocols

41

Random Access Protocols• When node has packet to send• Transmit at full channel data rate R• No a priori coordination among nodes

• Two or more transmitting nodes - collision

• Random access Medium Access Control protocol specifies• Key Problems• How to detect collisions• How to recover from collisions

• Examples• Slotted ALOHA• ALOHA• CSMA, CSMA/CD, CSMA/CA

Aloha Protocol

• Simple protocol developed by Norman Abramson and colleagues at University of Hawaii, in 1968• Used low-cost amateur radio-like systems to

create a computer network linking distributed campuses of the University

• Protocol allows every system to send a frame if its ready to send

• Based on a shared medium network• This was basis for Ethernet !!!!

Pure ALOHA

In pure ALOHA, frames are transmitted at completely arbitrary times.

44

Slotted ALOHAImprovement on Aloha

Assumptions• All frames same size• Time divided into equal size slots• Time to transmit 1 frame

• Nodes transmit only at slot beginning • Nodes are now synchronized• If 2 or more nodes transmit in slot, all

nodes detect collision

Slotted ALOHA

Operation• When node obtains fresh frame,

transmits in next slot• If no collision, node can send new

frame in next slot• If collision, node retransmits frame in

each subsequent slot with probability p until success• Probability is key and is between 0 and

1

46

Slotted ALOHA

• Probability is key • Probability p means each node

independently tosses a coin, heads or tails• P – says how likely it is node gets a head• Node transmits under this event

• (1-p) – is the probability of getting tail on a coin• Means node skips a slot and tosses again in next slot

Slotted Aloha C – collisionE – emptyS - Success

48

Slotted ALOHA

Pros• Single active node

can continuously transmit at full rate of channel

• Highly decentralized: Slots in nodes need to be in sync

• Simple

Cons• Collisions, wasting

slots• Idle slots• Nodes may be able

to detect collision in less than time to transmit packet

• Clock synchronization

C – collisionE – emptyS - Success

Takes 9 slots

49

Slotted Aloha efficiency

• Suppose: N nodes with many frames to send, each transmits in slot with probability p

• Probability that given node has success in a slot =

p(1-p)N-1

• Prob that any node has a success = Np(1-p)N-1

• Max efficiency: find p* that maximizes, successful nodesNp(1-p)N-1

• For many nodes, take limit of Np*(1-p*)N-1 as N goes to infinity, gives:

Max efficiency = 1/e = .37

Whats the Efficiency of Aloha?Efficiency : long-run fraction of successful slots (many nodes, all with many frames to send)

At best: channelused for useful transmissions 37%of time!

!

Slotted ALOHA

Throughput versus offered traffic for ALOHA systems.

51

CSMA (Carrier Sense Multiple Access)

• Max throughput achievable by slotted ALOHA is 0.368.

• CSMA improved throughput compared to Aloha protocols.

• Big Improvement ....

• Listen before you speak!• Listens to the channel before

transmitting a packet (avoid avoidable collisions)

52

Assumptions with CSMA Networks

1. Constant length packets2. No errors, except those caused by collisions3. Each host can sense the transmissions of all

other hosts4. The propagation delay is small compared to

the transmission time

53

CSMA (Carrier Sense Multiple Access)

CSMA: Listen before transmit:

If channel sensed idle• Transmit entire frameIf channel sensed busy• Defer transmission • Human analogy: Don’t interrupt

others!

54

CSMA collisions

Collisions can still occur:propagation delay means two nodes may not heareach other’s transmission

Collision:entire packet transmission time wasted

spatial layout of nodes

55

CSMA/CD (Collision Detection)CSMA/CD: Carrier sensing, deferral as

in CSMA• Collisions detected within short time• Colliding transmissions aborted, reduces

channel wastage • Collision Detection• Easy in wired LANs• Measure signal strengths, compare

transmitted and received signals• Difficult in wireless LANs: received signal

strength overwhelmed by local transmission strength

56

CSMA/CD collision detection

57

CSMA/CD

• Sense the channel• If idle, transmit immediately• If busy, wait until the channel becomes idle

• Collision detection• Abort a transmission immediately if a collision

is detected• Try again later after waiting a random amount

of time

• Will revisit this later when we describe Ethernet ...

58

Taking Turns

59

“Taking Turns” MAC protocols

Polling: • Master node

“invites” slave nodes to transmit in turn

• Typically used with “dumb” slave devices

• Bluetooth uses polling

• Concerns:• Polling overhead • Latency• Single point of

failure (master)

master

slaves

poll

data

data

Token Passing Protocols

• Token Ring protocol was developed by IBM mid-1980s

• Access method involves token-passing• In Token Ring, computers are connected so that

signal travels around the network from one computer to another in a logical ring

• A single electronic token moves around ring from one computer to the next

• If computer does not have information to transmit, it simply passes the token on to the next workstation

• If computer wishes to transmit and receives an empty token, it attaches data to the token

Token Passing Protocols

• Token Ring protocol requires a star-wired ring using twisted pair or fiber optic cable

• It operates, transmission speeds of 4 Mbps or 16 Mbps

• Due to the increasing popularity of Ethernet,• Use of Token Ring in decreased dramatically

http://en.wikipedia.org/wiki/IBM_token_ring

62

“Taking Turns” MAC protocols

Token passing:• Control token

passed from one node to next sequentially.

• Token message• Concerns:• Token

overhead • Latency• Single point of

failure (token)

T

data

(nothingto send)

T

63

Summary of Protocols

• Channel partitioning, by time, frequency or code• Time Division, Frequency Division

• Random access (dynamic), • ALOHA, S-ALOHA, CSMA, CSMA/CD• Carrier sensing: easy in some

technologies (wire), hard in others (wireless)

• CSMA/CD used in Ethernet• CSMA/CA used in 802.11

• Taking turns• Polling from central site, token passing• Bluetooth, FDDI, IBM Token Ring

64

End