coe 341: data & computer communications dr. marwan abu-amara

COE 341: Data & Computer CommunicationsDr. Marwan Abu-Amara

Chapter 7:

Data Link Control Protocols

COE 341 – Dr. Marwan Abu-Amara 2

Contents

1. Error Controla. Stop-and-Wait ARQ

b. Go-Back-N ARQ

c. Selective-Reject ARQ

2. Flow Controla. Stop-and-Wait flow control

b. Sliding-Window flow control

3. High-Level Data Link (HDLC)


What is Data Link Control The logic or procedures used to convert the raw

stream of bits provided by the physical layer into a “reliable” connection

Requirements and Objectives: Frame synchronization Error control Flow control Addressing Multiplexing data and control on connection Link management


Error & Flow Control

We now study two main functions of the data-link layer Error control Flow control

Why Error & Flow Control together? We usually lump Error Control and Flow Control

discussions together because the data-link protocols used for error control are also used for flow control


Error Control Definition Error Control is the second type of methods used to

“handle” errors in frames First type is FEC

Error Control methods are those that use retransmission in the case an error occurs in a frame This process is called Automatic Repeat Request (ARQ)

In general, a retransmission is needed if a: Frame is erroneous Frame is lost (i.e. does not arrive or arrives too late)

There are several reasons for this to happen (e.g. existence of noise, network congestion…)


Flow Control Definition We define Flow Control as: “The set of

procedures used to restrict the amount of data that a TX can send before waiting for an acknowledgement from the RX” (per Forouzan) This is to avoid overwhelming the RX by the flow

of data from TX RX does not ‘absorb’ the received data instantly It buffers (temporarily stores) the data it receives to do

some processing before sending it upward to higher layers

Without flow control, the RX buffer may overflow and data gets lost


Flow Control

Assume: No frames lost No frames arrive in error Frames arrive in the same order they were sent,

following a variable propagation delay


A Model of Frame Transmission

Loss

ErrorOnly one frame traveling on the linkat any given time


Error & Flow Control Protocols There are three main types of error control

protocols (ARQ-based): Stop-and-Wait ARQ Sliding Window ARQ

Go-back-N ARQ Selective-Reject ARQ

There are two main types of flow control protocols: Stop-and-Wait Sliding Window


Categories of Error Control


Categories of Flow Control


Stop-and-Wait ARQ

Stop-and-Wait ARQ is a simple protocol: TX keeps a copy of the last frame sent After receiving the frame, RX sends back an ACK After receiving this ACK, TX sends another frame

and so on… Both Data & ACK frames are alternately

numbered with “0” or “1” Data Frame “0” is acknowledged by ACK “1” Data Frame “1” is acknowledged by ACK “0”


Stop-and-Wait ARQ: Normal Operation


Stop-and-Wait ARQ: Abnormal Operation Stop-and-Wait ARQ deals with encountered

anomalies as follows: Lost or Damaged Frames: RX discards them silently

i.e. without sending Negative-ACK (NACK) back to TX RX keeps its current value for R (R var. defined on prev.

slide) Lost or Damaged ACK:

TX discards damaged ACK TX keeps a timer, after sending a frame, within which ACK

must be received. Otherwise, ACK is considered lost. In both situations, Lost and Damaged ACK, the TX sends

the frame again


Stop-and-Wait ARQ: Lost Frame


Stop-and-Wait ARQ: Lost ACK


Stop-and-Wait ARQ: Delayed ACK Another abnormality: TX receiving a late ACK

Timer has already expired The ACK was considered lost

Frame is re-sent again This frame will be duplicate at RX and discarded Also its ACK will be discarded when received back at TX

Then the late ACK arrives Now TX can send the next frame


Stop-and-Wait ARQ: Delayed ACK

Lost


Stop-and-Wait ARQ Piggybacking Piggybacking is a method

that combines the data and the ACK in one frame

It is useful in bidirectional communications

Stations can send their data along with ACK to data previously received

Piggybacking is faster and saves bandwidth


Stop-and-Wait ARQ Drawback In Stop-and-Wait ARQ, the line is not efficiently

utilized because only one frame is sent at a time Its ACK must be awaited for During this waiting period no frames are sent

Inefficiency gets even worse when TX speed is high

TX quickly sends frame then sits idle Propagation distance is high

It takes longer for the frame and its ACK to reach destination Both cases leave the TX waiting idle for longer times


Stop-and-Wait Efficiency Let:

Tf : Time to transmit a frame = Lf / R TACK : Time to transmit an ACK = LACK / R TProp : Propagation time = (link distance) / V TProc : Processing time

We define the frame total time as (see next slide): Ttotal = Tf + 2 TProp + TProc + TACK

The utilization of the link is defined as: U = (Tf)/(Ttotal)


Stop-and-Wait Efficiency It is common to ignore TProc and TACK

because they are negligible compared to other times

In which case the utilization becomes: U = (Tf)/(Tf +2Tprop) = 1/(1 + 2a) Where: a = Tprop/Tf a is called the “length of the link in bits”: length

of medium in bits compared to frame length (Lf) Notice that for very small a (i.e. when 1st

transmitted bit reaches RX, source will still be transmitting), U 100%

Notice that for very big a (i.e. frame transmission is completed before 1st bit reaches destination), U 0%

Efficient for links where a << 1 (long frames on a short link, i.e. fills link)

TX RX

Tf

Tprop

Tprop

Tproc

TACK


Stop and Wait Efficiency: Example Compare the efficiency of stop-and-wait error control for two

links using the parameter ‘a’: Frame size, L = 1000 characters of 8 bits each, = 8000 bits

200-m optical fiber link Data rate, R = 1 Gbps Typical wave velocity, V = 2 x 108 m/s

Frame TX time, tf = L/R = 8000/(1x109) = 8 s

Propagation time, tprop = d/V

= 200/(2x108) = 1 s

End of first frame reaches RX after 8+1 = 9 s from start

ACK takes 1 s more to reach TX, which sends 2nd frame after 10 s

Utilization = 8/10 = 80% (=1/(2a+1))

8

9

2002 10 0.125

80001 10

prop

f

dT VaLTR

Satellite link between 2 ground stations d = 2 x 36,000 km, Data rate, R = 1 Mbps Typical wave velocity, V = 3 x 108 m/s

Frame TX time, tf = L/R = 8000/(1x106) = 8 ms

Propagation time, tprop = d/V

= 2x36x106/(3x108) = 240 ms

End of first frame reaches RX after 8+240 = 248 s from start

ACK takes 240 ms more to reach TX, which sends 2nd frame after 488 ms

Utilization = 8/488 = 1.6% (=1/(2a+1))

6

8

6

2 36 103 10 30

80001 10

prop

f

dT VaLTR


Sliding Window Protocol Stop-and-Wait can be very inefficient when a > 1 Protocol:

Assumes full duplex line Source A and Destination B have buffers each of size W frames For m-bit sequence numbers:

Frames are numbered: 0, 1, 2, …, 2m-1, 0, 1, … (modulo 2m)

ACKs (RRs) are numbered: 0, 1, 2, …, 2m-1, 0, 1, … (modulo 2m)

A is allowed to transmit up to W frames without waiting for an ACK B can receive up to W consecutive frames ACK J (or RR J), where 0 J 2m-1, sent by B, means B has

received frames up to frame J-1 and is ready to receive frame J B can also send RNR J: B has received all frames up to J-1 and is

not ready to receive any more Window size W, can be less or equal to 2m-1


Sliding Window Protocol Example of Sliding-Window-Protocol: m = 3 bits, W = 7

Observations:• A may Tx W = 7

frames (F0, F1, …, F6)• After F0, F1, & F2 are

Tx-ed, window is shrunk (i.e. can not transmit except F3, F4, …, F6)

• When B sends RR3, A knows F0, F1 & F2 have been received and B is ready to receive F3

• Window is advanced to cover 7 frames (starting with F3 up to F1)

• A sends F3, F4, F5, & F6

• B responds with RR4 when F3 is received – A advances the window by one position to include F2

W

W

W

W W = distance between first unacknowledged frame and last frame that can be sent


Go-back-N ARQ Go-back-N ARQ improves the efficiency of the line by

sending up to W frames before worrying about ACK This is called Pipelining (several tasks are started before the

1st is finished) For this, the frames must be sequentially numbered Sequence number must be included in the header of

the frame If m bits are reserved for sequence number then the

sequence numbers range from 0 to 2m – 1 Example:

m = 3 bits, then seq. numbers range is 0-7 and could be used as: 0, 1, 2, 3, 4, 5, 6, 7, 0, 1, 2, 3, 4, 5, 6, 7, 0, 1,…


Go-back-N ARQ Sliding Window At the TX side, to hold the outstanding frames until

they are individually ACK, we use a window The window size W cannot exceed 2m – 1 W is fixed in this protocol but may be variable in others (e.g.

TCP) Each time a proper ACK is received for a frame the

window slides past this frame, hence the name Sliding Window The acknowledged frames can be purged out of the TX

memory As an example, next slide shows that frame 0 and 1

have been acknowledged So the sliding window slides just past them in (b)


TX Sliding Window

In (b), frame 0 and frame 1 were properly acknowledged


RX Sliding Window At the RX side, the size of the window is

always 1 The RX window is centered on the next

expected frame number If any other frame arrives (i.e. out of sequence

arrival), it is immediately discarded If the right frame arrives, the window slides

past it to zoom on the next expected frame


RX Sliding Window


Go-back-N ARQ Control Variables TX keeps track of three variables

S: Sequence number of the recently sent frame SF: Sequence number of the first frame in the window

SL: Sequence number of the last frame in the window

The frame size W is related to these variables by: W = SL– SF + 1

On the other hand the RX has only one variable: R: Expected frame sequence number


Go-back-N ARQ: TX & RX Control Variables


Timers & Acknowledgments TX sets a timer for each frame sent RX has no timers RX sends an ACK if a frame arrives with no errors and in order If RX receives damaged or out of sequence frame, it silently

discards them until it receives the expected frame The silence of the RX causes the TX timer to expire This in turn causes the TX to go back and send all frames starting

from the non-acknowledged frame This is why it is called Go-back-N ARQ

Example: If TX has already sent frame-6 but the timer for frame-3 expires

without receiving an ACK for it, the TX goes back and send frames 3, 4, 5, 6


Go-back-N ARQ Normal Operation


Go-back-N ARQ Lost Frame Operation

Here, ACK2 acknowledges frames 0 & 1 at the same time


Quiz: Window Size Condition for Go-back-N ARQ Earlier, we mentioned

that the size of the window must be W (2m – 1) i.e. W < 2m where “m” is the number of bits reserved (in the frame overhead) for the sequence number

By comparing the figures across explain the need for this condition


Sliding Window Protocol

Animation for Sliding Window protocol Sliding Window Protocol Simulation (http://www.cs.stir.

ac.uk/~kjt/software/comms/jasper/SWP3.html)


Two neighboring nodes (A and B) use a sliding-window protocol with a 3-bit sequence number. As the ARQ mechanism, go-back-N is used with a window size of 4. Assuming A is transmitting and B is receiving, show the window positions for the following succession of events:

a) Before A sends any frames

b) After A sends frame 0, 1, 2 and B acknowledges 0, 1 and the ACKs are received by A

c) After A sends frames 3, 4, and 5 and B acknowledges 4 and the ACK is received by A

Example: Problem 7-9


Example: Problem 7-9 - Solutiona)

0 1 2 3 4 5 6 7 0• • • • • •

b)

0 1 2 3 4 5 6 7 0• • • • • •

c)

0 1 2 3 4 5 6 7 0• • • • • •

W

W

W


Sliding Window Protocol - PiggybackingWhen using sliding window protocol in full

duplex connections: Node A maintains its own transmit window Node B maintains its own transmit window A frame contains: data field + ACK field There is a sequence number for the data field, and a

sequence number for the ACK field


Selective-Repeat ARQ Go-back-N ARQ simplifies the job of the RX

RX keeps track of one control variable R only Out-of-sequence frames are simply dropped, not buffered

However, this protocol is inefficient for noisy links where the probability of damaged frames is high This means many retransmissions will occur which uses up

the bandwidth of the link For noisy links, a more efficient mechanism is used in

which only the damaged frame is retransmitted (not all N frames) It is called Selective-Repeat ARQ


Selective-Repeat ARQ TX & RX Windows Window configuration for the TX is the same as for

Go-back-N ARQ However, the window size should be W (½ 2m) i.e. W

2(m-1) RX window must have the same size as the TX

This is a major difference with Go-back-N ARQ where window W = 1 for the RX

RX window specifies a range of a acceptable frame sequence numbers

Also, in selective-repeat ARQ the RX uses negative ACK (NACK) to report the sequence of a damaged frame to the TX


Selective-Repeat ARQ: TX & RX Control Variables

Here, S = SF


Selective-Repeat ARQ Operation: Lost Frame Example

Here, ACK2 acknowledges frames 0 & 1 at the same time

0


Quiz: Window Size Condition for Selective-Repeat ARQ Earlier, we mentioned

that the size of the window must be W 2m – 1 where “m” is the number of bits reserved (in the frame overhead) for the sequence number

By comparing the figures across explain the need for this condition


Sliding Window Protocol - Efficiency Refer to Appendix A When window size is W (for error free), link utilization,

U, is given by

where a = Tprop/Tf (i.e. length of link in bits) Sliding window protocol can achieve 100% utilization if

W (2a + 1) The smaller the W needed the better! (Why?)

To get high value for U, small value for a is needed as well !!

1 (2 1)

(2 1)2 1

W aU W

W aa


Sliding Window Efficiency: Example Compare the efficiency of Sliding Window flow control for two

links using the parameter ‘a’: Fame size, L = 1000 characters of 8 bits each, = 8000 bits

200-m optical fiber link Data rate, R = 1 Gbps Typical wave velocity, V = 2 x 108 m/s Frame TX time, tf = L/R = 8 s

Propagation time, tprop = d/V = 1 s

a = tprop / tf = 0.125 100 % link utilization is achieved with

window size W:

W (2 a+1) (2 x 0.125 +1) 1.25

i.e. W = 2 (A window of just 2 frames)

Satellite link between 2 ground stations d = 2 x 36,000 km, Data rate, R = 1 Mbps Typical wave velocity, V = 3 x 108 m/s Frame TX time, tf = L/R = 8 ms

Propagation time, tprop = d/V = 240 ms

a = tprop / tf = 30 100 % link utilization is achieved with

window size W:

W (2 a+1) (2 x 30 +1) 61

W = 61, k = 6 bit (Large window and buffer sizes)

For k = 3 bits, W = 7:

Utilization U = W/(2a+1) = 7/(61) = 11.5% > 1.6% for Stop and wait.


Stop-and-Wait & Sliding Window as Flow Control Protocols

Problem: Qualitatively explain how the Stop-and-Wait and

the Sliding Window protocols could be used for flow control

Hint: Think about the use of ACK and when the RX could

send it back to the TX


High-Level Data Link Control Protocol (HDLC) One of the most important data link control protocols

and it is the basis for many data link control protocols. The job of the HDLC layer (Link Layer, Layer 2) is to

ensure that data passed up to the next layer has been received exactly as transmitted (i.e error free, without loss and in the correct order)

Another important job is flow control, which ensures that data is transmitted only as fast as the receiver can receive it.

There are two distinct HDLC implementations HDLC NRM (see SDLC), and HDLC Link Access Procedure Balanced (LAPB)

Usually when referring to HDLC people mean LAPB or some variation


High-Level Data Link Control Protocol (HDLC) To satisfy a variety of applications, HDLC defines

three types of stations, two link configurations, and three types of data transfer modes of operation.

Station types: Primary Station (PS):

Responsible for controlling the operation of the link Frames issued by the PS are called commands

Secondary Station (SS): Frames issued by the SS are called responses – operates under the control of a primary station

Combined Station: issues commands and responses Link configurations:

Unbalanced: one primary plus one or more secondary Balanced: two combined stations


High-Level Data Link Control Protocol (HDLC) Transfer modes:

Normal Response Mode (NRM) – used in unbalanced config.; secondary may only tx data in response to a command from primary

Asynchronous Response Mode (ARM) – used in unbalanced config.; Secondary may initiate data tx without explicit permission; primary still retains line control (initialization, error recovery, …)

Asynchronous Balanced Mode (ABM) – used in balanced config.; either combined station may tx data without receiving permission from other station

Animation for HDLC


HDLC – Applications NRM:

Point-multipoint (multi-drop line): one computer (primary) polls multiple terminals (secondary stations) number of terminals are connected to a host computer

ARM: rarely used ABM: most widely used (no polling involved)

Full duplex point-to-point


HDLC NRM


HDLC ABM


HDLC – Frame Structure


HDLC – Frame Format Flag:

Size: 1 Byte Special pattern 0 1 1 1 1 1 1 0 used as frame begin/end and synch. Used in Header and trailer

Address: Size: 1 Byte (or more for larger networks) If primary station created the frame, the address is that of the destination

secondary station If secondary station created the frame, the address is that of the source

secondary station Networks not using “primary/secondary” (e.g. Ethernet) use 2-Byte address

(source/destination) Control:

Size: 1 or 2 Bytes Used for error & flow control

Information: Size: Varies from network to network. Always fixed within a network Contains user data from Network layer or Network Management information

FCS: Size: 2 or 4 Bytes Implements ITU-T CRC for error detection


HDLC – Frame Structure – Flag Field

Flag Field: unique pattern 01111110 Used for synchronization To prevent this pattern form occurring in data bit stuffing

procedure is used Tx-er inserts a 0 after each 5 1s Rx-er, after detecting flag, monitors incoming bits – when a pattern of

5 1s appears; the 6th/7th bit are checked: If 0, it is deleted If 10, this is a flag If 11, this is an ABORT

• Pitfalls of bit stuffing: one bit errors can split one frame into two or merge two frames into one


HDLC Data Transparency If the data field of an HDLC frame contains a

pattern identical to the flag pattern (01111110), the RX will interpret it as end-of-frame flag Next bits will be considered part of the next frame This is called lack of Transparency

Bit Stuffing is the process of adding one extra 0 whenever there are five consecutive 1s in the data so that the receiver does not mistake the data for a flag


HDLC Bit Stuffing & Removal


HDLC Bit Stuffing Problems: With bit stuffing, Single-bit errors could

split a frame into 2 or merge two frames into 1.

01111100(Bit stuffed)

(Bit stuffed)

01111110

FrameSplitting

FrameMerging

01111110

01111100


HDLC Frame Structure – Address Field

Address field identifies the secondary station that transmitted or is to receive frame

Not used (but included for uniformity) for point-to-point links

Extendable – by prior arrangementAddress = 11111111 (single octet) used by the primary

to broadcast to all secondary stations

Extended Address Field


HDLC Frame Structure – Control Field

Three types of frames: I, S, and U Information frame (I): carry user data (upper layers) – flow and error control info

is piggybacked on these frames as well Supervisory frame (S): carry flow and error control info when piggybacking is not

used Unnumbered frame (U): provide supplementary link control

First 2 bits of the control field determine the type of frame, the remaining are organized as shown in figure 7.7.c and d (next slide)

Poll/Final (P/F) bit: In command frames (P): used to solicit response from peer entity In response frames (F): indicate response is the result of soliciting command


HDLC Frame Structure – Control Field

Extension of sequence number 7-bit sequence numbers rather than 3-bit ones

“Set-mode” command extends control field to 16 bit for S and I frames

Unnumbered frames always use 8-bit


HDLC Frame Structure – Information/FCS Fields Information field:

Present ONLY in I-frames and some U-frames Contains integer number of octets Length is variable – up to some system defined maximum

FCS field: CRC error detecting code Calculated from ALL remaining bits in frame (excluding flag

field) Normally 16 bits (CRC-CCITT polynomial = X16+X12+X5+1), or 32-bit optional FCS using CRC-32


HDLC Frames HDLC uses Synchronous transmission HDLC defines three types of frames

Information frames (I-frames) Used to transport user data and user control information

relating to the user data (e.g. piggybacking) Supervisory frames (S-frames)

Used to transport control information only Unnumbered frames (U-frames)

Used in the link management


HDLC Generic Frame Format

Indicates that the first transmitted field is the Header flag, then the address, then the control…

Header Trailer


HDLC I, S & U Frames Format


HDLC I-Frames I-frames are designed to carry user data from

Network-layer They can include flow & error control information

(piggybacking) The bits of the I-frame Control field are:

First bit is always “0”. It means it is an I-frame Next 3-bits, called N(S), define the sequence of the frame

So possible sequences are 0,1,2,3,4,5,6,7 Next bit is called P/F (Poll or Final)

When a primary station polls other stations, it sets this bit to 1 When a secondary station responds to a poll, it sets this bit to 1

Next 3-bits, called N(R), define the value of the ACK when piggybacking is used


HDLC I-Frames


HDLC S-Frames


HDLC U-Frames


U-frame control command and responseCommand/response Meaning

SNRMSNRM Set normal response mode

SNRMESNRME Set normal response mode (extended)

SABMSABM Set asynchronous balanced mode

SABMESABME Set asynchronous balanced mode (extended)

UPUP Unnumbered poll

UIUI Unnumbered information

UAUA Unnumbered acknowledgment

RDRD Request disconnect

DISCDISC Disconnect

DMDM Disconnect mode

RIMRIM Request information mode

SIMSIM Set initialization mode

RSETRSET Reset

XIDXID Exchange ID

FRMRFRMR Frame reject


HDLC – Operation HDLC operation consists of the exchange of I-

frames, S-frames, and U-frames between two stations.

Table 7.1 list types of Control/Response functions for various frame types

The operations of HDLC involve three phases: Initialization (by either side): U-Frames

Both agree on various options Data Transfer (by the two sides): I- and S-Frames

Exchange of user data and control info for flow and error control

Disconnect (by either side): U-Frames Signaling termination of operation


HDLC – Operation

Initialization:

(Link Setup & Disconnect) Issue SABME command and starts timer B responds with UA (or DM if request is

rejected) A receives UA and initializes its variables To disconnect: issue DISC command Followed by UA

SABME: Set Asynchronous balanced/extended mode;7-bit sequence


HDLC – Operation Data Transfer:

Once initialization is complete a logical path is established.

Both sides start to send I-frames (Full-duplex exchange) starting with seq. number 0

N(S), N(R) both are seq. number to support flow and error control

N(R) is the ACK for the I-frame received; it enables the HDLC module to indicate which number I-frame it expect to receive next.

RR is used (S-frame) when there is no reverse user data (I-frame) traffic


HDLC – Operation Disconnect:

Either side can issue a DISC frame to request disconnect

The remote entity MUST accept the request by sending UA.

Any outstanding I-frame will be lost.

Busy condition: Notice the use of P/F bit. When A is unable to keep up with

the speed of the transmitter “B” or buffer is full.

A sends RNR, to halt the transmission of B

To check the readiness of A, periodically B sends RR frame with P set.

Once the condition of being busy is cleared A responds with F=1


HDLC – Operation Reject RecoveryReject Recovery:

I-frame 4 was lost B receives I-frame 5 (out of

order) – responds with REJ 4 A resend I-frame 4 and all

subsequent frames (Go-back-N) Timeout RecoveryTimeout Recovery::

A sends I-frame 3 – but it is lost Timer expires before

acknowledgement arrives A polls Node B B responds indicating it is still

waiting for frame 3 – B set the F bit because this a response to A’s solicitation


HDLC Example: Piggybacking w/o Errors Next slide shows an exchange using piggybacking when

there is no error Station A begins the exchange of information with an I-

frame numbered 0 followed by another I-frame numbered 1 Station B piggybacks its acknowledgment of both frames

onto an I-frame of its own Station B’s first I-frame is also numbered 0 [N(S) field] and

contains a 2 in its N(R) field, acknowledging the receipt of A’s frames 1 and 0 and indicating that it expects frame 2 to arrive next

Station B transmits its second and third I-frames (numbered 1 and 2) before accepting further frames from station A Its N(R) information, therefore, has not changed: B frames 1 and 2

indicate that station B is still expecting A’s frame 2 to arrive next


HDLC Example: Piggybacking w/o Errors


HDLC Example: Piggybacking with Errors Suppose frame 1 sent from station B to station

A has an error Station A informs station B to resend frames 1

and 2 (the system is using the Go-Back-N mechanism)

Station A sends a reject supervisory frame to announce the error in frame 1

Next slide shows the exchange


HDLC Example: Piggybacking with Errors

coe 341: data & computer communications dr. marwan abu-amara

Documents