COMPUTER NETWORKS Data link layer protocolsnetgroup.polito.it/courses/Didattica/CN/layer2_3.pdfPag....

Pag. 1

COMPUTER NETWORKS – Data link layer protocols

COMPUTER NETWORKS – Data link layer protocols - 1 Copyright Gruppo Reti – Politecnico di Torino

Data link layer (layer 2)

Gruppo Reti TLC

[email protected]

http://www.telematica.polito.it/


Copyright

• Quest’opera è protetta dalla licenza Creative Commons

NoDerivs-NonCommercial. Per vedere una copia di questa

licenza, consultare

http://creativecommons.org/licenses/nd-nc/1.0/

oppure inviare una lettera a: Creative Commons, 559

Nathan Abbott Way, Stanford, California 94305, USA.

• This work is licensed under the Creative Commons

NoDerivs-NonCommercial License. To view a copy of this

license, visit:

http://creativecommons.org/licenses/nd-nc/1.0/

or send a letter to Creative Commons, 559 Nathan Abbott

Way, Stanford, California 94305, USA.


Data link layer functions

• Frame delineation

– Explicit delimiters (flag) before and after packet transmission

– Lenght indicator

– Fixed length

– Silence between packets

• Multiplexing (of higher layer protocols)

• Addressing

• Error detecion

• Window protocol

– Flow control

– Sequence and error control through retransmissione

• Multiple access protocols for shared media

Pag. 2



Data link layer protocols

• Most derived from the SDLC (Syncronous Data

Link Control) defined in the IBM netowrks

architecture

• Ansi standardized SDLC as ADCCP (Advanced

Data Communication Control Procedure) and ISO

as HDLC (High-level Data Link Control)

• CCITT (later ITU-T) derived LAP (Link Access

Procedure) and LAPB (Link Access Procedure

Balanced)

• In Local Area Networks (LANs) a MAC sublayer

solves the multiple access problem



• Several protocols belong to the HDLC family

– LAP-D (Link Access Procedure D-Channel)

– LAP-F (Link Access Procedure to Frame Mode

Bearer Service)

– LLC 802.2 (Logical Link Control)

– PPP (Point-to-Point Protocol)

– LAPDm (LAP for the mobile D channel)



• Starting from HDLC…

• ….. we move to

– LAP-B, X.25 and ISDN B channel

– LLC, for LANs

– PPP, for point to point connections • Used at home

– LAP-F, used in Frame Relay

– ATM, not derived from HDLC, used in B-ISDN

Pag. 3



Example of public network

• DTE: Data Terminal Equipment

• DCE: Data Circuit-terminating Equipment

NETWORK

DCE

DCE

DCE

DCE

DTE

DTE

DTE

DTE

Datsa link layer

protocols


• PDU format:

• Bit oriented protocol, with bit-stuffing to

ensure data transparency (flag 01111110

must not appear in other fields)

• Address field is used in multi-point

configuration (master/slave)

• Control field differentiates the PDUs

01111110 address control data CRC 01111110

8 8 8/16 >=0 16 8

Data format (many layer 2 protocols)


HDLC: Master/Slave Configuration

• Protocol used to manage the communication among master and slaves

• PDUs sent from the master are named command, PDUs send from slaves are named response

master

slave

Pag. 4



HDLC: operational modes

• Normal Response Mode (NRM)

– Suited for point-to-point of unbalanced multi-

point.

– One primary station (Master) and several slaves

– The Master sets the P/F bit to 1 to enable slave

transmission

– The slave sets the P/F bit to 1 in the last PDU


HDLC: operational modes

• Asynchronous Response Mode (ARM)

– Unbalanced configuration. The slave can send

data without waiting to be polled by the master.

• Asynchronous Balanced Mode (ABM)

– Balanced point-to-point configurations. The P/F

bit requires immediate response.

– Most widely used in data link layer protocols


HDLC: three types of PDUs

• Control field

– Differentiates three types of PDUs

• P/F = poll/final bit

0 N(S) P/F N(R)

1 0 S S P/F N(R)

1 1 M M P/F M M M

Information

Supervision

Unnumbered

Pag. 5



HDLC: three types of PDUs

• Information

– Data sent after connection opening

• Supervision

– ACKs (positive and negative)

• Unnumbered

– Link management

– Data sent in connectionless mode


HDLC: two numbering schemes

• Normal numbering (modulo 8) and extended

numbering (modulo 128)

• Control field equal to either 1 o 2 byte

0 P/F

1 0 S S x x x x P/F N(R)

1 1 M M P/F M M M

Information

Supervision

Unnumbered

N(R) N(S)


HDLC vs LAP-B

• LAP-B (ISDN B channel) uses only a subset

of PDUs defined by HDLC

• We briefly describe the PDUs used by LAP-B

Pag. 6



LAP-B: information PDU

• Data transfer

• N(S) and N(R) fields needed by the window

protocol to provide error and sequence

control

– N(S) = transmitted PDU sequence number

– N(R) = acknowledge number, refers to the

expected PDU at the receiver

0 N(S) P/F N(R)


LAP-B: supervision PDUs (S)

• ACK transfer

• RR (Receiver Ready - C/R)

– Positive ACK

• RNR (Receiver Not Ready - C/R)

– Positive ACK and flow control signal sent from the

receiver which is unavailable (ON-OFF flow control)

• REJ (Reject - C/R)

– Request for retransmission of all PDU starting from N(R)

1 0 S S P/F N(R)


LAP-B: unnumbered PDUs (U)

• Mainly used to control connection

management

• 5 M bits. A limited number of PDUs is used

by LAP-B among the 32 available PDUs

• Command PDUs:

– SABM(E) (Set Asynchronous Balanced Mode),

used to (re)open the connection

• E = Extended numbering scheme

– DISC (Disconnect): the connection is aborted

1 1 M M P/F M M M

Pag. 7



LAP-B: unnumbered PDUs (U)

• Response PDUs

– UA (Unnumbered Acknowledgment):

• ACK for PDUs opening the connection or to answer to

DISC commands

– DM (Disconnect Mode)

• Connection was not set up correctly

– FRMR (FRaMe Reject)

• Answer to the reception of a correct but unknown

PDU

• 24 additional bits to explain the reason why the PDU

was rejected


format command response code in control field

1 2 3 4 5 6 7 8

Information I (Information) 0 N(S) P N(R)

Supervision RR (Receiver Ready) RR (Receiver Ready) 1 0 0 0 P/F N(R)

RNR (Rec. Not Ready) RNR (Rec. Not Ready) 1 0 1 0 P/F N(R)

REJ (Reject) REJ (Reject) 1 0 0 1 P/F N(R)

Unnumbered SABM (Set Asynchr. Balanced Mode) 1 1 1 1 P 1 0 0

DISC (Disconnect) 1 1 0 0 P 0 1 0

DM (Disconnect Mode) 1 1 1 1 F 0 0 0

UA (Unnumbered Acknowledgement) 1 1 0 0 F 1 1 0

FRMR (Frame Reject) 1 1 1 0 F 0 0 1

LAPB: command and response PDUs


LAPB: Poll/Final Bit

• In command PDUs

– the P/F bit is used to poll stations (i.e. to require an answer) when

set to 1

• In response PDUs

– the P/F bit is used to answer (final) to command PDUs with a P/F bit

set to 1

• Poll bit set to 1 by DTE (or DCE) is a request to answer

immediately for DCE (or DTE).

• Final bit set to 1 specifies the answer to the poll request

• DTE and DCE cannot send a command PDU with the P bit

set to 1 unless a response PDU with the F bit set to 1 was

received

Pag. 8



LAPB addresses

• DTE has address 00000011 (3)

• DCE has address 00000001 (1)

• 3 is the value of the address field in

– command PDUs DCE DTE and in

– response PDUs DTE DCE

• 1 is the value of the address field in

– command PDUs DTE DCE and in

– response PDUs DCE DTE

• The address field permits to distinguish command

PDUs from response PDUs and to understand

whether the P/F bit is a poll or a final


Layered architecture:

X.25 and ISDN B channel

L3

L2

L1

LAP-B

Physical

LAP-B

Physical

Switching node User terminal

Error control

User

terminal

X.25

packet layer

Physical

LAP-B

X.25

packet layer

Physical

LAP-B

X.25 packet layer


Q V Q V

V,UA,F

Q,I00

V,I10

V,I00,P

V,I21 Q,I10,P

V,RR2,F

V,I32

V,I42

V,I52

V,I62

V,I72,P

V,RR3

V,RR4

Q,I25

Q,I36

V,I03

V,I14

V,I24 Q,I47

V,RR2,F

V,RR5,P Q,I52,P

V,RR3,F Q,RNR6,F

V,RNR6,P

V,RR3,F

V,DISC,P

V,UA,F

V,SABM,P

modo non

connesso

modo non

connesso

Q,RR2,F

LAPB: example of data transfer

Pag. 9



Q V

V,I10

V,I00

V,I20

V,I30

V,I40

V,I50

V,I60 V,RR7

trasferim.

dati

tim

eout

V,RR0,P V,RR7,F

V,I70

LAPB: error recovery via timeout


Q V

Q,I00

V,I10

V,I00

V,I21

Q,I10

V,I32

V,I23

Q,I21

V,REJ2 Q,I32

Q,I42

trasferim.

dati

V,I34

V,I45 Q,I53

LAPB: error recovery via REJ


Q V

Q,I10

V,I11

V,I00,P

V,I22

V,I21,F

V,I33,P

V,I45

Q,I31

Q,I41

Q,I51

trasferim.

dati

V,I15,P

V,I26 V,RR2,F

Q,I00

V,RR1,F

LAPB: error recovery via P/F bit

Pag. 10



Q

V,RR7

tim

eout

V,REJ7

V

V,I10

V,I00

V,I20

V,I30

V,I40

V,I50

V,I60

trasferim.

dati

V,I00

V,I70

V,I10

LAPB: error recovery via

time-out and REJ


Data link layer in LANs

• IEEE 802.3 data link layer in LANs has two

sub layers

– LLC: Logical Link Control

– MAC: Medium Access Control

LLC

MAC

physical

network 3

2

1


IEEE 802.2 Logical Link Control

• Derived from HDLC

• Standard ISO 8802/2.

• Main characteristics

– Byte oriented protocol

– Does not use flag for packet delineation (done

by the MAC)

– No erroro control (done by the MAC)

– PDU have source and destination addresses

used for higher layer protocol multiplexing

– Variable packet lenghth

Pag. 11



LLC PDU

• One unused bit in the source address is

exploited to distinguish between command

and response.

• Control field

– 8 bit in unnumbered PDUs

– 16 bit otherwise

• Maximun packet size depends on the MAC

DSAP SSAP control information

address address 8 bit 8 bit 8 o 16 bit 8 x M bit


LLC PDU

• LLC addresses identify the higher layer

protocols transported bt the LLC protocol

• Address field has two bit, unused for

addressing, reserved for control purposes

• All 0s address identifies the LLC itself

• Up to 63 addresses available


SNAP PDU

• Extended addressing scheme, named SNAP

(SubNetwork Access Protocol)

• Used for connectionless transfer

• Identifies IP protocol among others

0AAH 0AAH 03H 000000H 0800H (3)-PDU

DSAP SSAP controllo informazione

OUI protocol type (3) (2)

Pag. 12



Point to Point Protocol (PPP)

RFC 1661 • PPP (Point to Point Protocol) is mainly used

to connect customers to Internet ISP over

telephone or ADSL lines

• Also available over ISDN connections and

even over SONET/SDH

• Point-to-point connection is easier to manage

with respect to braodcast or radio channel

• PPP is simple (likely the simplest data link

layer protocol)


PPP functionalities

• PDU delineation

• Content transparency

• Error detection (but no correction)

• Higher layer protocols multiplexing

• Monitoring of link activity

• Negotiation of layer three (IP) addresses

– Useful for ISPs


Functions not available in PPP

• Error correction

• Flow control

• Sequence control

– Point to point link

• Multipoint link management

Pag. 13



PPP PDU

• Flag for “framing” or PDU delineation

• Address

– Not used and kept for compatibility with HDLC

• Control

– Not used and kept for compatibility with HDLC

• Protocol

– Higher layer protocol multiplexing


Byte Stuffing

• Data transparency is obtained by defining a mechanism to

send, in the data portion of the packet , the byte with value

01111110

• The sender inserts an escape byte 01111101 before any

01111110 byte or 01111101 byte

• The first byte equal to 01111101 is discarded by the

receiver


PPP Link Control Protocol

• PPP-LCP protocol opens (and close) the PPP

connection

– Options are negotiated

– Max frame length; authentication protocol; skip address

and control fields

• DEAD is the initiaI state

• Once the PPP connection has been established

and the authentication process was successful, for

the IP protocol

– the IP Control Protocol (RFC1332) is used to set up IP

parameters (addresses, datagram compression, etc)

Pag. 14



PPP Data Control Protocol


Frame Relay

• Standard to create packet networks based on virtual circuits

(normally permanent VCs) on a wide area

• The standard was originally proposed within the ISDN

framework

• Today used

– to create VPNs (Virtual Private Networks) for companies

– to interconnect LANs

– to build logical topologies to interconnect Internet routers for ISP

• Bit rate ranging from 64 kb/s to 2 Mb/s

• Variable size packets (well suited to data traffic)

– Maximum size 4096 byte

• http://www.frforum.com


Frame Relay

• Operates on Permanent Virtual Circuit

(although signaling protocols to deal with

SVC are defined)

Frame Relay

network

DTE

DTE DTE

DTE

DCE

DCE

DCE

DCE

Virtual

Circuits

Pag. 15



LAPF

• Frame Relay defines the LAPF protocol (Link Access Procedure to Frame mode bearer services)

• LAPF is divided in two parts:

– DL-CORE (reccomendation I.233) • Used in all network nodes

– DL-CONTROL • Optionally used only by end users (today, mainly IP

routers)

• In most applications, it is not used


Core and edge approach

L>=

3

L2

L1

DL-CORE

Physical

DL-CORE

Physical

Frame Relay

switching node

User terminal

Error control

User terminal

Higher layer

protocols

DL-CONTROL

Physical

DL-CORE

Higher layer

protocols

DL-CONTROL

Physical

DL-CORE


LAPF packet

• Packet delineation through flag and

bitstuffing to guarantee data transparency

Flag Address Control CRC Flag Information

DL-CORE

DL-CONTROL (like HDLC with extended numbering)

Pag. 16



LAPF packet

• ADDRESS field

contains

– the DLCI (Data Link

Connection

Identifier),the virtual

circuit identifier

– some additional bits for

congestion control and

traffic policing

DL-CORE

DL-CONTROL

ADDR (LSB)

FLAG

FLAG

ADDR (MSB)

information

CRC (LSB)

CRC (MSB)

control


ADDRESS field • DLCI: Data Link

Connection

Identifier

• FECN/BECN:

forward/backward

explicit congestion

notification

• DE: discard

eligibility

• C/R:

command/response

• D/C: DLCI or DL-

CORE

• EA: extension bit

default format (2 byte)

format 3 byte

format 4 byte

EA

0

EA

1

EA

0

EA

0

EA

1

EA

0

EA

0

EA

0

EA

1

lower DLCI or DL-CORE control

C/R

C/R

C/R

DE

DE

DE

BECN FECN

FECN

FECN

D/C

D/C

upper DLCI

upper DLCI

upper DLCI

DLCI

DLCI

DLCI

lower DLCI

lower DLCI or DL-CORE control

BECN

BECN


B-ISDN

• Private and public networks

• Integrated network

– Support all type of services, with different transmission

speeds and quality of service requirements over the

same network infrastructure

• Standardized by ITU-T and ATM Forum

• ISDN (re)evolution

• Exploit ATM as a transport, multiplexing and

switching technique

Pag. 17



Asynchronous Transfer Mode:

ATM • Defined as a data link layer protocol

• Wide area network exploiting packet

switching networks with virtual circuit service

• Characteristics

– High bit rate (starting from 622 Mb/s)

– Low latencies for voice services

– Fixed size data units (cells)

• 53 byte (48 byte of data)

• http://www.atmforum.com


B-ISDN: reference model

Management

plane

Control plane User plane

Higher layers Higher layer

AAL (ATM Adaptation Layer)

ATM layer

Physical layer

Pla

ne

ma

na

ge

me

nt

La

ye

r ma

na

ge

me

nt


Core and edge approach

in the user plane

L>=3

L2

L1

ATM

Physical

ATM

Physical

ATM switching node User

terminal

Error detection only on-demand

for some AAL

User

terminal

Higher layers

protocols

AAL

Physical

ATM

Higher layers

protocols

AAL

Physical

ATM

Pag. 18



ATM: which OSI layer?

• Original idea

– End-to-end transport

– ATM from desktop to

desktop

• Reality:

– IP router interconnection

– IP over ATM

– ATM is yet another data

link technology


ATM cell format

• Header (5 bytes) + payload (48 bytes)

• Fixed size cell

– To ease the switching task at high speed (synchronous

switching)

• Small cell size

– Reduced latency (can be obtained by increasing

transmission speed)

– Small packetization delay for interactive voice services

– Segmentation overhead

• Slightly different format at network edge and core


GFC

VCI

VPI

VCI CLP

HEC

PT

DATA

VPI VCI

VCI

VCI CLP

HEC

PT

DATA

VPI VCI

VPI

UNI CELL NNI CELL

ATM cell format

5 BYTE

48 BYTE

Pag. 19



ATM cell format

• ATM cell header (5 bytes = 40bit)

– GFC (4 bit): Generic Flow Control

– VPI (8-12 bit): Virtual Path Identifier

– VCI (16 bit): Virtual Circuit Identifier

– PT (3 bit): Payload Type

– CLP (1 bit): Cell Loss Priority

– HEC (8 bit): Header Error Code


ATM cell format

• GFC - Generic Flow Control

– Only at the UNI interface.

– The network issues information to user on the

number of cells that can be sent

– Two control algorithms:

• ON-OFF

• Credit based


ATM cell fromat

• VPI - Virtual Path Identifier

– Variable length:

• 8 bit at the UNI (256 VP’s)

• 12 bit at the NNI (4096 VP’s)

– Some VPIs are reserved to network

management functions and to signalling

Pag. 20



ATM cell format

• VCI: Virtual Circuit Identifier

– Identifies a single virtual circuit within a given

VPI.

– 65536 VC’s are available for each VP.

– Example: link at 2,4 Gb/s, 1 VP, all VCs with the

same capacity 36Kb/s for each VC.


ATM cell format

• PT - Payload Type

– Classifies the payload information type.

– It contains an identifier named Payload Type

Identifier (PTI).

– Among the eight possible codes,

• four are reserved to network functions

• four to user function


PT MEANING

0 0 0

User cell EFCI No congestion

AAL 5 indication=1 1 0 0

User cell EFCI Congestion


User cell EFCI Congestion


PT field (Payload Type)

User cell EFCI No congestion

AAL 5 indication=0

Pag. 21



PT SIGNIFICATO

0 0 1

OAM cell

(Operation and Maintenance) 1 0 1

RM cell

(Resource Management) 0 1 1

Not used

Reserved for future use 1 1 1

Campo PT (Payload Type)

OAM cell

(Operation and Maintenance)


ATM cell format

• CLP - Cell Loss Priority

– Two priority levels at the ATM layer (within each

VC)

– In ATM switches, it permits to selectively discard

cells in case of buffer congestion

– CLP=0 indicates a high priority cell


ATM cell format

• HEC - Header Error Code

– It permits to check the correctness of the ATM

cell header only

– No error detection on paylod!

– Single errors are corrected

– Two errors are detected

• SEC/DED: Single errore correction/ Double Error

Detection

Pag. 22



B-ISDN: reference model

Management

plane

Control plane User plane

Higher layers Higher layer

AAL (ATM Adaptation Layer)

ATM layer

Physical layer

Pla

ne

ma

na

ge

me

nt

La

ye

r ma

na

ge

me

nt


AAL: ATM Adaptation Layer

• Integrates ATM transport to offer service to users

• Servide dependent layer

• Examples of AAL functions:

– Transmission errors detection and managment

– Segmentation and reassembly

– Cell loss management

– Flow control

– Synchronization


AAL: ATM Adaptation Layer

• It defines four classes of service (service

classes)

– Through three main parametrs:

• Source transmission speed

• Type of connection (connection

oriented/connectionless)

• Temporal relation between end user

Pag. 23



AAL: 4 service classes

• A: CBR traffic, constant but rate, connection

oriented, synchronism required AAL 1

• B: VBR traffic, connection oriented,

synchronism required AAL 2

• C: VBR traffic, connection oriented,

synchronism not required AAL 3/4

• D: VBR traffic, connectionless, synchronism

not required AAL 5


Class A Class B Class C Class D

Synchronism required between source and dest

Speed

Connection

type

AAL type

Possible

applications

required not required

costant

(CBR)

variable

(VBR)

Connection oriented connectionl

ess

AAL 1 AAL 2 AAL 3/4 - 5

voice 64kbit/s

video CBR

video/audio

VBR data data

AAL service classes


AAL layer: architecture

• The AAL layer is subdivided into two sub-

layers

– convergence sublayer (CS):

• Service and ATM traffic convergence

• Multiplexing

• Error detection

• Synchronism recovery

– segmentation and reassembly (SAR):

• Segmentation in transmission, reassembly in

reception of CS PDUs

Pag. 24



AAL

SSCS

CPCS CS

SAR

AAL architecture

• CS convergence sublayer

• SAR segmentation and reassembly

• SSCS service specific CS

• CPCS common part CS

• Some sub-layers can be empty


AAL format

AAL 1

AAL 2

AAL 3

AAL 4

AAL 5

ATM Cell Header SN IT SAR - SDU LI CRC

44 byte

ATM Cell Header ST SN RES SAR - SDU LI CRC

44 byte

ATM Cell Header ST SN MID SAR - SDU LI CRC

44 byte

ATM Cell Header SN SNP SAR - SDU

47 byte

ATM Cell Header

48 byte


AAL 3/4

CPI B Tag

BA size

AAL payload pad AL Lenght

SAR header

SAR trailer

SAR header

SAR trailer

SAR header

SAR trailer

2 byte 44 byte 2 byte SAR - PDU

1B 1B 2B 0-3B 1B 2B 2B

ST SN MID LI CRC

2 4 10 bit 6 bit 10 bit

ST=EOM

ST=COM

ST=BOM

E Tag CPCS PDU

SAR PDU

Pag. 25



AAL 5

SAR

Layer

PDU

CS

Layer

PDU

End of segment = 1

CS Layer Payload

Le

ng

ht

CR

C -

32

48 bytes

SAR

payload

48 bytes

SAR

payload

48 bytes

SAR

payload

1- 65535B 0-47B 2B 2B 4B


AAL 5

• No CS layer

• SAR layer exploit all 48 byte payload

• Last cell created by the segmentation

process has the third bit in the PT field of the

ATM header set to 1

– Layer separation principle violated!

• Error control over the full CS-PDU


AAL 5

• Advantages

– simplicity

– efficiency

– Improved reliability (CRC - 32)

• Disadvantages

– Use the third bit of the PT field in the ATM

header

• Breaks layer separation principle

– Loss of the cell with the PT bit set =1 implies that

two full CS-PDUs are lost

Pag. 26



Layer 2 protocols: a comparison

Protocol Packet delineation Layer 3 protocol

multiplexing

Error detection Error correction

(window

protocol)

LAPB Delimiter Higher layer YES YES

LAPF core +

LAPF control

Delimiter Via virtual

circuits

YES in

LAPF core

Optional in

LAPF control

(edge)

ATM (core)+

AAL (edge)

Done at the

physical layer

Via virtual

circuits

YES in

AAL (edge)

NO

PPP Delimiter YES YES NO

LLC Done in MAC

IEEE 802.3

YES Optional Optional

Ethernet MAC Silence YES YES NO

In gray info either not presented in this class or later descriebd

COMPUTER NETWORKS Data link layer protocolsnetgroup.polito.it/courses/Didattica/CN/layer2_3.pdfPag....

Documents

Transcript of COMPUTER NETWORKS Data link layer protocolsnetgroup.polito.it/courses/Didattica/CN/layer2_3.pdfPag....