Course: Architectures and Computer Networks Chapter 2

Chapter 2: Data Link Protocols

To enable communication between machines identified by their addresses, it is necessary to define a

a large number of rules regarding the structuring of dialogue, the format of messages
transmission, their logical sequence, information encoding, transmission rhythm,
etc. The set of rules, comparable to rules of spelling and grammar
defining the sentence structure of a language, is called connection protocol
data or communication protocol. A program (communication software), installed
on the equipment that must communicate remotely, it executes. In order to ensure a
maximum interoperability between different devices, the standardization instances have
worked on the definition of communication protocols at an international level.

Definition: A protocol is a set of rules and data formats.

to ensure proper data exchange between two devices or
two programs. A data link protocol aims to make the circuit reliable
of data.

While the data circuit transmits binary elements, the link protocol
data works on blocks of binary elements called frames. The frame is therefore
the unit of data it manages. It carries user data and also contains,
order information, necessary for the protocol to ensure the smooth running of the
dialogue (some frameworks, the supervisory frameworks, are in fact limited to only
order information). A frame contains different fields.
Each of them is a block of binary elements whose meaning and interpretation
are specified in the protocol definition.
The protocol must also define the rules of dialogue and specify how to
correct the detected errors. Finally, we must be able to detect equipment failures or
complete disconnections to alert the user of the unavailability of the service.

Remark
Defining a data link protocol involves specifying: the format of the frames
exchanged, the conditions for delimiting the frameworks (start and end) and their validity, the position
and the meaning of the different fields of a frame, the error detection technique
used, the rules of dialogue (supervision of the connection) and the procedures to be followed afterwards
detection of errors or failure of the connection.

DATA FORMATTING
In theory, the start and end frame delimiters are independent of the
transmission technique used. In practice, some processes use specific features
online coding to delimit the frames. The most common solutions are the
delimitation by a special binary sequence or the explicit indication of the length of the
frame.

1.1 Delimitation by a specific sequence of binary elements

Frames having any number of bits, a specific sequence, called

banner (or flag), is used to indicate both the beginning and the end of frames. In general, it is
compose the byte 01111110. A transparency mechanism is necessary to avoid

ISTAG, LIG
1
Cours : Architectures et réseaux informatiques Chapter 2

find this sequence within a frame: in the broadcast, we insert into the body of the
bring me a binary element 0 after encountering five consecutive binary elements of value
1.

In reception, the binary element with value 0 must be removed after encountering five.
consecutive binary elements of value 1. Such a mechanism (bit stuffing) prohibits
the emission of more than five binary elements of value 1 in the body of the frame, since
this configuration is reserved for its delimitation. This method allows for the transmission of
frames of any length without particular constraints.

Example: Let's take the following useful data: 0110 1111 1110 1001. Preceded and followed
of flags, they will actually be issued in the form: 01111110 0110 1111 10110 1001
01111110. In this sequence, the flags are underlined and the inserted bit for transparency
is in bold and underlined.

1.2 Delimitation by transmitting the length of the data field

Another delimitation method involves indicating, in a specific field, the

number of useful bytes contained in the frame. After a start frame sequence, one or
several bytes indicate its length (the location of this field is fixed in relation to the
start of frame), which is usually expressed in bytes or in number of words (of 16 or 32
bits, for example).
This process leads to a limitation of the frame size: thus, if the length, expressed
in bytes, is coded on a byte, we are limited to frames of 256 bytes. In doing so, we avoid
the transparency issues since the receiver does not interpret the data in any way
received as delimiters. The length of the frame can be adjusted to a fixed length
by adding binary padding elements. The field then specifies the size of the data.
tools transported in the frame.

[Link] OF THE VALIDITY OF THE TRANSMITTED INFORMATION

Error control involves checking the validity of the transmitted data. If we
admit that the transmission service is not reliable, one must protect oneself against potential
errors, so detect them and then correct them. To do this, we add to the transmitted information a
redundancy, that is to say control information calculated by a specified algorithm
in the protocol from the data block. Upon receipt, the same algorithm is executed
to check if the redundancy is consistent. If so, it is considered that there is no
transmission error and the received information is processed; otherwise, we are sure that
the information is invalid and it is ignored.

Error correction is done either through a new attempt to

transmission, either by exploiting the wealth of redundant information, which locates and
corrects the detected errors. When using simple error detection, the receiver
has no way to locate the detected error: it can be located in both the
redundancy field than in the data field. In any case, the frame is
considered invalid and ignored by the receiver. It is still possible that errors of
transmission appears and that, unfortunately, the coherence remains true. We then find ourselves
in the presence of residual errors. In this case, the error control mechanism could not
detect that several transmission errors have mutually compensated each other. The rate

ISTAG, LIG
2
Course: Computer Architectures and Networks Chapter 2

Residual errors must be as low as possible, but they can never be zero in a
real data link.

2.1 Validity control: protection at the code level

Protection at the code level consists of organizing an internal redundancy within it.
: parmi toutes les combinaisons possibles, certaines sont retenues comme valides. Ce type de
protection is possible when data transmission is done character by character (we introduce a
redundancy for each transmitted character). For example, a bit is added to each character.
of parity said vertical parity or VRC (Vertical Redundancy Check)1, calculated as follows:
for each character, we take the modulo 2 sum of its bits. If the number of 1 bits is even,
we add 0 at the end of the character, and if the number of 1 bits is odd, we add 1.
Le contrôle de validité par VRC est fréquemment utilisé sur les liaisons asynchrones.
For example, for the character M encoded by 1001101, the parity bit is 0. It is transmitted in
this order 10110010 (the 7 data bits starting from the least significant then the bit of
Parity). The general disadvantage of parity checks is that errors are not detected.
doubles.

2.2 Validity Control: frame-level protection

Protection at the frame level involves adding redundancy to each frame.

depending on the set of binary elements that constitute it. Several techniques are
foreseeable, but we will only examine longitudinal parity and control here
polynomial, which are the most well-known and widely used methods.

Longitudinal parity control or LRC (Longitudinal Redundancy Check): For

improve error detection in transmissions using parity checks, we
often associates longitudinal parity and vertical parity (VRC + LRC). To do this, we add to the
end of the frame, a code word called longitudinal parity or LRC, consisting of the sum
modulo 2 of all bits of the same rank.

Example: Let the sequence of characters L, 2, M to be transmitted, encoded in CCITT no 5 by

the hexadecimal values 4C, 32, and 4D. In even parity, the parity bits (in bold in the
Each character is worth 1, 1, and 0 respectively. The parity character
longitudinal is calculated as follows:

11 0 0 1 1 0 0 character L + VRC parity;

10 1 1 0 0 1 0 character 2 + VRC parity;
01 0 0 1 1 0 1 character M + VRC parity;
00 1 1 0 0 1 1 LRC character to add at the end of the data block as a character of
control.

The sequence of emitted binary elements is therefore 0011 0011 0100 1101 1011 0010 1100
1100, if we transmit the characters one after the other, starting with the weights
weaknesses of each character.

Polynomial control: Polynomial control, commonly referred to by abuse of

Cyclic redundancy check or CRC is widely used in
modern protocols because it allows detecting errors on multiple bits. We
let's just describe the process here without making it a theory.

ISTAG, LIG
3
Course: Architectures and Computer Networks Chapter 2

In polynomial control, the frame to be transmitted is considered as a group of

bits to which we correspond a polynomial P(x), such that the coefficient of degree i corresponds
to the value of the bit. Calculation algorithms are done modulo 2 on polynomials [by
example, (x7+ x3+(x3+ x)= x7We choose a polynomial G(x) of degree r, called
generator polynomial, characteristic of the control. At transmission, P(x) is multiplied by xrand on
divide the polynomial obtained by G(x). The remainder denoted R(x), obtained by Euclidean division, is
strictly lower degree than r. It is added at the end of the frame as a control code. Thus
:xr*P(x)=G(x)*Q(x)+R(x).(1)
We transmit the polynomial T(x), made from P(x) and the remainder R(x) and defined by
equation (2):
T(x)=xr*P(x)+R(x).(2)
According to equations (1) and (2) and taking into account that the calculations are performed modulo
2, this polynomial satisfies:
T(x) = G(x)*Q(x).
It is therefore divisible by G(x).
We have seen that the data circuit can modify the information. Let E(x) be the polynomial
associated with the errors introduced by the circuit. The received data has the associated polynomial
S(x), defined by: S(x) = T(x) + E(x). Upon reception, we divide S(x) by G(x) and obtain a
resteR1(x) that verifies the following equation:
S(x) = G(x) * Q1(x) + R1(x).
If SiR1(x) is zero, it is assumed that E(x) is zero and that the information received corresponds to that.
issued. If R1(x) is not zero, the polynomial E(x) is also not zero: the data circuit has
introduce one or more errors and the received information should be ignored.
Example:
The information 1000001110000100 is associated with P(x) = x15+ x9+ x8+ x7+ x2.
Let the generator polynomial of degree 12:
G(x) = x12+ x11+ x3+ x2+ x + 1.
The division of dex12*P(x)parG(x) gives:
R(x) = x11+ x9 + x8+ x7+ x6+ x4+ 1.
We transmit:
1000001110000100101111010001
x12*P(x) R(x)
At the reception, we check that the remainder of the division by G(x) is zero.

3. Features of a link protocol

3.1 FLOW CONTROL

Let us suppose that A is a computer and B is a slow printer, equipped with a capacity
limited memory, requiring him to remember all the information sent by A
as long as they are not printed. If the pace of sending information is significantly
above its printing speed, there is quickly saturation of memory and loss
information by B. A control mechanism for the pace of sending must be put in place.
information to the receiver, called flow control
To perform flow control, we introduce two supervision frames, RR
(Receiver Ready) and RNR (Receiver Not Ready). These frames carry no
useful information and are only used for managing the dialogue. They are generated and utilized
by the binding protocol and are invisible to the user. The mechanism is as follows: to
For each frame reception, equipment B sends an RR frame if it is ready to accept.

ISTAG, LIG
4
Course: Computer Architectures and Networks Chapter 2

other templates or an RNR template if he no longer wants to receive new ones. In this last
cas, Benvoie RR as soon as he is ready to accept new frames.

3.2 MANAGEMENT OF RECEIPTS

Let's now assume that the circuit is not completely reliable and introduces some
errors. To the flow control mechanism described above, a process must be added.
of the acknowledgment of the information frames received.
In data link protocols, a strategy is rather used.
of positive acquittal, using the previous supervision frameworks (RR and RNR). The
the operational strategy of B becomes:
If B receives a correct frame, the equipment sends an acknowledgment (RR frame or
RNR), according to its memory state to ensure flow control.
If it receives an erroneous frame, it does not memorize it and does not send anything back, as if it
had received nothing.

3.3 FRAME NUMBERING OF INFORMATION

The protocol numbers each frame of information. The numbering is placed in

the header, just like the type of the frame. Two frames with different numbers are
considered as carrying distinct data units. The receiver protocol
exploit the header to check if the frame is correct and in sequence. If so,
The information contained in the data field is delivered to the user.
N(S) [S for send] is the variable that gives the frame number. This
variable, coded on a few bits, is taken modulo M, an integer that can take the values:
2 (the minimum to distinguish between two successive different frames), 8 or 128. The introduction
of numbers in the frames requires counters at each station. Furthermore, it is necessary
initiate the dialogue so that the two stations agree on the initial values
of meters.

3.4 NOTION OF WINDOW

In order to increase the effectiveness of the dialogue, we introduce the notion of anticipation, which is to...
say the possibility of sending multiple frames in succession, without having received the acknowledgment of
previous frames. Thus, a supervision frame no longer acknowledges a single frame but a
a set of frames that follow one another without error. The number of successive frames that can be
sending without receipt of acknowledgment is limited by a noted value, called window.

3.5PROTOCOLGO-BACK-N

In the Go-back-N strategy, a supervisory frame called REJ

(Reject) requests the retransmission of frames starting from the erroneous frame.

3.6 PIGGY-BACKING

In bidirectional exchanges, each device sends frames.

numbered information and acknowledges the frames I that he has received. There are therefore two kinds of
sent frames: the I frames and the acknowledgment frames (RR, RNR or REJ according to the state of the
reception). Such a mechanism is not very efficient; it is improved by using the I frames.

ISTAG, LIG
5
Course: Computer Architectures and Networks Chapter 2

to convey both the information to be sent and the acknowledgments of the received frames
previously.
Each frame I has two numbers: a number N(S) [the sequence number of the
frame I], and a number N(R) acknowledging the frames issued in the opposite direction. This mechanism
is called piggy-backing. Finally, when a station has no frame I to transmit, it can
always use RR frames to acknowledge the traffic it receives.

A data link protocol can provide several services depending on the quality of the
transmission
Service without receipt, connection, or flow control when one wishes to use it.
a very simple protocol where the data circuit is of excellent quality.
Service with acknowledgment but without connection or flow control that allows
to improve the reliability of the connection a little but does not guarantee the non-duplication of the
messages.
Service with acknowledgment, connection, and flow control that includes numbering
frames and acknowledgments. This service, the only one to offer a real guarantee of reliability,
is also the most complex to implement. Several strategies can be distinguished in the
acknowledgement management: Stop-and-Wait (use of a look-ahead window)
equal to 1), the Go-back-Net the Selective Reject. The Stop-and-Wait is not very effective, the
Go-back-N is the most commonly used; Selective Reject does not bring any obvious gain.
performances in the majority of cases.

4. Description of the HDLC (High Level Data Link Control) protocol

HDLC is the protocol standardized by the ITU (International Telecommunications Union).

which describes a full-duplex transmission operating on a point-to-point link;
the transmission is synchronous and bit-oriented. This protocol implements the mechanism of
transparency that makes it completely independent of the coding of the transported data. HDLC
can transport information using variable-length codes. Its most common variant
Known is of the Go-back-N type with a flow control mechanism. It operates in mode
balanced or symmetrical, meaning that both stations have the same prerogatives and
can possibly operate in a half-duplex mode.

4.1 STRUCTURE OF A HDLC FRAME

The frame is the unique structure of any length that carries all the
Flags mark the beginning and the end; a single flag marks the end of a
frame and the beginning of the next one when two frames are emitted consecutively. The
The table below describes the different fields of the frame, in their order of appearance:
The address field spans one byte and identifies one end of the link.
The 'fieldControl' describes the type of the frame: it spans 1 byte (over 2 bytes in the
extended mode.
The field Information is optional. It contains any number of elements.
binaries representing the user's data.
The FCS (Frame Control Sequence) field is the frame control sequence,
obtained by a polynomial check whose generating polynomial is x16 + x12 + x5
1. This generator polynomial is the one recommended by ITU recommendation V41.

ISTAG, LIG
6
Course: Computer Architectures and Networks Chapter 2

We start by emitting the low-weight bits (from bit 1 to bit 8 of each field).
The transmission of binary elements is continuous; in the absence of specific emission, the
equipment emits flag sequences to maintain synchronization between the
two ends of the data link.

4.2 DIFFERENT TYPES OF HDLC FRAMES

There are three types of frames identified by the Control field: the frames
Information or frames allow the transmission of user data. The
supervision frames or S frames allow for acknowledgment and flow control; they
do not carry data, just like unnumbered frames or U frames
These are used to control the link: initialization, release,
notification of unrecoverable errors... Only a frame I can transmit data;
it is numbered by the variable N(S) and also contains the acknowledgment of the frames
received in reverse order (piggy-backing process), thanks to the number N(R).

The value 0 of bit 1 is specific to frame I. The value of the P/F bit depends on the status of
the equipment (primary or secondary) and the nature of the framework (request or response).

Supervision frames (tramesS)

The frames release the frames and indicate the availability status of the stations (suitability
or not to receive new frames I). Containing a number N(R), they are used for control
of errors and flow control. The three supervision frames are:
The RR frame indicates that the equipment is ready to receive new frames.
numéroN(R) gives the number of the next expected frame. It means that all the
Frames with number(s) strictly less than N(R) have been received. A device
can also send RR frames to indicate its status or to request the status of the
station located at the other end.
The RNR frame acknowledges the received frames and furthermore indicates that the equipment is not
not able to receive new frames. The number N(R) is the same
meaning only in the framework.
The REJ framework is used to request the immediate cessation of current broadcasts and a
retransmission from the frame important the number indicated in N(R).

ISTAG, LIG
7
Course: Computer Architectures and Networks Chapter 2

Unnumbered frames (framesU)

Frames are used for the additional control functions of the link. Let's mention
the main ones:
SABM (Set Asynchronous Balanced Mode) to initialize operation in mode
balanced.
DISC (DISConnect) to logically break the connection between the two stations.
UA (Unnumbered Acknowledgement) to acknowledge orders like SABM
or DISC.
FRMR (FRaMe Reject) to reject an invalid order (correct from the perspective
from the detection of errors but inconsistent with the state of the dialogue).
DM (Disconnect Mode) to indicate the disconnection status of a station. It
is used, in particular, to respond negatively to a request for initialization by
SABM.

5. Specific case of the PPP (Point to Point Protocol)

The PPP protocol is the point-to-point link protocol used on the Internet. It uses the
subscriber's telephone lines to access the network (the connection typically concerns a
personal computer and the Internet service provider). This is a very
simplified HDLC that does not include -except for options- neither flow control nor mechanism of
review on errors. The format of a PPP frame is as follows:

The 8 bits of the Address field are set to 1 (the connection being point-to-point, a single value)
just the address is enough).
The Control field has the same meaning as in HDLC. The PPP Data field
starts with two bytes (the protocol field), which identify the level protocol
superior to which the frame is intended; it ends with a FCS field whose calculation method
is identical to that of an HDLC frame.

ISTAG, LIG
8
Course: Architectures and Computer Networks Chapter 2

The only frame carrying data on a reliable link is a UI type U frame.

This frame contains a field of information but is not
numbered (because there is no flow control). The absence of an error recovery mechanism
does not mean that the circuit is reliable: the FCS field is used to validate the received frames.
PPP also includes a set of sub-protocols chosen at the opening of the
data link to secure exchanges: LCP (Line Control Protocol), PAP (PPP
Authentication Protocol, CHAP (Challenge Authentication Protocol) and NCP (Network)
Control Protocol).
At the initiation of a transfer, PPP negotiates the parameters of the exchange through the protocol
LCP; PAP allows the exchange - in clear - of passwords before the data transfer.
If a secure exchange is desired, CHAP can be used, which performs encryption.
throughout the communication by means of a prior exchange of public keys and keys
secrets.
Finally, the NCP protocol is used to negotiate connection parameters (the options of
transfers chosen by each endpoint of the connection, independently of one another) and the
network level settings.

Directed Work Chapter 2:

Exercise 1:

Let the sequence of binary data located in the information field of a frame
HDLC : 011110111110011111100011.
1. What is the actual suite provided for the transmission support (for this data
only) ?
What happens if the twelfth bit of the actually transmitted sequence was misrecognized?
of the receiver?

Exercise 2:

Write the sequence of bits actually transmitted for an SABM frame sent by a device
from address A (03 in hexadecimal) to a device with address B (01 in hexadecimal).
1. Bit number 5 (bitP) is set to 1. It will be assumed that the FCS of this frame is worth in binary.
1101011111111011.
2. By which frame does the equipment B respond?

Exercise 3:

Calculate the CRC and the LRC of the message HELLO using even parity, knowing that H is coded by
0001001Epar1010001,Lpar0011001etOpar1111001Specify the order of transmission of the message.
built.

Exercise 4:

We want to transmit the sequence of 16 bits of data: 2BE3 (in hexadecimal), the first bit
transmission corresponding to the least significant bit of the number 2. Error protection is
made by vertical parity (of even parity) and longitudinal.
1. Give the bit sequence of the four characters and the complete binary sequence transmitted to
receiver for this data block.

ISTAG, LIG
9
Course: Computer Architectures and Networks Chapter 2

2. Assuming that, due to a transmission error, on the 19ebit of the found sequence
To the question, modify it, calculate the value of the remainder found by the receiver.

Exercise 5:

We want to check the data block consisting of the two bytes encoded with even parity.
00110011 and 11110011.
What is the LRC corresponding to this block of data?
2. Represent the LRC in polynomial form.
3. We want to verify this block of data by a polynomial check of polynomial
generators8+ 1. Give the value of the polynomial LRC(x). What do you observe?

Exercise 6:

Let the sequence of binary elements be 0 1 1 1 0 1 0 0 0 0 1 0 1 0 1.

1. Calculate the error control block for this data, assuming that we use a
code polynomial of the generator polynomial5+ x3+ 1.
2. We receive the following block: 0 0 0 1 0 1 0 0 0 0 1 0 1 0 1 0 0 0 1 0. The error checking
using the same generator polynomial, what decision is made by the receiver
regarding this block?

Exercise 7:

We consider the binary data sequence: 010110110000111101010100, made up of two

8-bit characters followed by a block of error control calculated using a polynomial code
of the generator polynomial81. Does the receiver detect transmission errors in the
received suite? Why?

Exercise 8:

The first bits to be transmitted in a data frame managed by the HDLC protocol
are as follows: 10101110. The generator polynomial used is the normalized polynomial by the
recommendationV41.
1. Find the remainder of the polynomial division of the message by this polynomial.
2. Assuming that the transmission of the 8 bits was done without error, represent it.
sequence of operations performed by the receiver from the first bit of data to
the reception of the last control bit.

Exercise 9:

A device A dialogue with a device B according to the LAP-B protocol, via a link
satellite. The satellite re-emits the received signals without any processing and is located 200 km away.
of altitude. All delays due to processing are negligible. The equipment communicates
at 9,600 bit/s and limit the size of the data field to 64 bytes.
1. Any equipment receiving a correct information frame processes it immediately.
if he has no information to convey. Assuming that there is none
transmission error, determine what the minimum window size is
anticipation for effective transmission, when Aest is the only equipment
emit data.

ISTAG, LIG
10
Course: Computer Architectures and Networks Chapter 2

2. Same question with a satellite now located at an altitude of 36,000 km.

Exercise 10:

In a data link protocol, it is assumed that each transmitter can use at

maximum Maxseq + 1 different sequence numbers, counted from 0 to Maxseq. Explain
Why the emission window size must remain less than Maxseq. Highlight it.
a case of ambiguity.

Exercise 11:

1. A simultaneous bidirectional data exchange is considered between two stations A and

B, managed by the LAP-B protocol (a subset of HDLC using simple rejection of
errors). The link is already initialized and no timer is armed. The stationA
5 information frames to send to station B. It has 10 to emit towards A.
It is assumed that all frames are of the same length. The two stations
start their transmission at very similar moments to one another.
Diagram of exchanges indicating the number of frames sent and received with the
classic nomenclature: I, N(S), N(R) for a framework I bearing the number N(S) and
releasing the frames up to number N(R)–1. It is similarly assumed:
That there are no transmission errors on the frames of A.
That there is a transmission error on the sixth and seventh frames of B.
That the frames are correctly transmitted.
That the propagation time is equivalent to a quarter of the transmission duration of a
I.
What is the maximum size of the anticipation window.
That the acknowledgments of receipt (the transmission duration of which is equal to a quarter of the
duration of a frame I) are transmitted as soon as possible. They are in any case included in some
frames if there are any.
2. What would happen if the frames of A were ten times longer than those of B?

Exercise 12:

We consider two stations A and B using the LAP-B version of the HDLC protocol.
exercise, we assume that the data link is properly initialized and that the
Data transfer can start as soon as possible. The size of the anticipation window is 2.
the data exchange is full-duplex and both stations start simultaneously the
data transfer. It has only one frame of information to emit and its frame is without
error. 3 information frames to emit. We will consider several scenarios for
the data transfer from A:
1. In the first case, the first frame of Aest is incorrect the first time it is
issued but it is broadcast correctly. Provide the corresponding diagram for the
different frames exchanged between A and B.
2. In the second case, the second frame of Aest is incorrect the first time it is
issued but it is transmitted correctly. Provide the corresponding diagram for the
different frames exchanged between A and B.
In the last case, the last two frames of A are erroneous the first time.
They are issued but they are transmitted correctly. Provide the diagram.
corresponding to the different frames exchanged between A and B.

ISTAG, LIG
11
Course: Architectures and Computer Networks Chapter 2

We now consider that the two stations A and B are using the version of the protocol
HDLC requiring the retransmission of erroneous frames using selective reject SREJ.
4. Using the same working assumptions as in question 1, provide the diagram.
corresponding to the different frames exchanged between A and B.
5. Based on the same working assumptions as in question 2, provide the diagram.
corresponding to the different frames exchanged between A and B.
6. By using the same working assumptions as in question 3, provide the diagram.
corresponding to the different frames exchanged between A and B.

Exercise 13:

Let two stations transmitting bidirectionally simultaneously 4 consecutive I frames and using a
HDLC type procedure. It is assumed that the initialization of the data link has been
performed and that the propagation and acknowledgment time of the frames is negligible. We
suppose that the size of the window is sufficient not to block the processes
The numbered frames 0 and 2 are erroneous when both are emitted.
first times in both directions. They are properly reissued the next time.
1. Provide the diagram of the frames exchanged between the two stations with the mode of
simple rejection (REJ).
2. Provide the diagram of the frames exchanged between the two stations when they
use the selective rejection mode (SREJ instead of REJ).

Exercise 14:

A bank manages the data connections of its ATMs using the HDLC protocol.
with simple rejection8. Each automaton has a data link with the following characteristics
following:
binary rate: 14,400 bits/s;
anticipation window equal to 4.
The bank's server uses a data link configured as follows:
binary rate: 2.048 Mbit/s;
anticipation window equal to 7.

1. Can bank ATMs know that the login parameters are

different at both ends of the connection? Where does the adaptation between the services occur?
used at both ends?

The introduction of a customer's bank card initiates a banking operation, which takes place
as follows: after entering the client's data, the automated system sends a frame I describing
the operation requested from the bank server. The latter, after verifying the data
sent by the automated system, launches the customer account management application, then the application
create a receipt and send it to the automaton in a frame I. The automaton will be able to
then print the ticket to be provided to the customer (or display the operation result on the screen of
the automaton).

[Link] the frame exchange between the banking terminal and its network access point.
assuming that data transmission occurs with an error of
transmission in the first frame I sent in each direction (we are not concerned with
the management of bitP/F).

ISTAG, LIG
12
Course: Computer Architectures and Networks Chapter 2

3. This time, the error handling of transmission is done by selective rejection (SREJ) of the
frames (we no longer use the LAP-B protocol). By taking the same assumptions
as previously, provide the diagram describing the exchanges between the automaton and
its network access point.

ISTAG, LIG
13