Master Networks - ATMPLS 2008/2009

ATM

Exercise 1:

We consider an ATM network, composed of 4 interconnected nodes as shown in the figure.

below. On this network are connected 4 users A, B, C, and D. Nodes 1 and 4 are
VC/VP switches while nodes 2 and 3 are VP switches. The 2
Connections VC (A-C, B-D) are multiplexed on the same VP.

A nodes C

users 1 2 3 4 users

B D

1. Give an example of a switching table for nodes 1 and 2.

2. Let a Ping command be used by user A on machine C. The datagram

The obtained IPv4 is as follows (hexadecimal):

45 00 00 50 20 61 00 00 80 01 C5 64 C7 F5 B4 0A C7 F5 B1 09
08 00 00 1C 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F 10
11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F 20 21 22 23 24
25 26 27 28 29 2A 2B 2C 2D 2E 2F 30 31 32 33 34 35 36 37 38

a. To transmit this IP packet over the ATM link, what happens?

b. How many ATM cells are needed to transmit such a datagram?

c. Represent the complete format of the CPCS-PDU by a diagram.

d. What is the overall effectiveness of the transmission from the application's perspective?

If an ATM cell is lost, what happens: is it recovered on the link?

Exercise 2:
We consider the network whose topology is given in the following figure. It is a network
packet transfer.
Master Networks – ATMPLS 2008/2009

A
I am
B
j, n 3 k, o
1 l, p
2
4

1. It is assumed that this network is in connected mode. During a communication between A

and B, a path (virtual circuit) is open and passes through intermediate nodes 1, 3
And 2. Are these nodes routers?

2. It is considered that the network is of the ATM type.

a. How many virtual circuits can pass through node 3 (take into account the
figure above)?

b. What path will all the packets belonging to the same flow follow?

c. Would it be possible to consider that the references are not the same?
length in input and output of the same switch?

3. Suppose that A is browsing the Web and accesses servers

different at each packet broadcast.
a. What seems to you to be the best strategy (ATM switching or protocol
IP). Explain the reasons.

b. Suppose we want to improve the quality of service, what seems to you

be the best of both networks?

c. Supposons qu ’à un instant donné, 300 millions d ’utilisateurs soient connectés

on the Internet. If each client transmits at the speed of its interface, the
The Internet network would have to carry a traffic much higher than it can.
transporter. Therefore, there needs to be a flow control. What is the flow control of
Internet?

4. Let the virtual circuit between A and B consist of a succession of VCs (Virtual Circuits)
i, j, k, l and VP (Virtual Path) m, n, o, p. Suppose that node 3 is a switch.
which only commutes on the duct number (VP) and nodes 1 and 2 are
ATM switches.
Are there values i, j, k, l, m, m, o, p that are equal?

b. It is assumed that the transport from A to B concerns an analog video path,

digitized at 32 Mbit/s. The transport delay constraint for this type of
Analog data then digitized is 28 ms (as for speech). In
Master Networks – ATMPLS 2008/2009

assuming the transmission speed of signals over the media

If the speed is 250,000 km/s, give the maximum distance between A and B.

c. If we put ATM cells in a SONET frame, what is the bit rate?

useful knowing that SONET (Synchronous Optical Network) is a frame of
level 1 of 810 bytes every 125 µs. Each SONET frame contains 9
tranches, which, in turn, contains 3 bytes of supervision and 87 bytes of
data.

d. Suppose that A emits 2 types of applications to B. Propose a method

flow control based on a leaky bucket in the case where the 2 flows of
data follow two distinct virtual circuits? Is this mechanism sufficient?
if the 2 data streams are multiplexed on the same virtual circuit?

5. In order to go from one point to another in the network, it is necessary to cross a subnet.
nodes 1, 2, 3, 4 are IP routers. To go from A to B, you must cross 4 sub-
networks. If in order these sub-networks are of type Ethernet between A and 1, X.25
between 1 and 3, and between 1 and 4, ATM between 3 and 2, and between 4 and 2, and finally again
Ethernet between 2 and B.

B
3
1
2
4

a. Describe the architecture and protocol layers using a drawing

traversed to go from A to B.

b. We now assume that the 4 subnetworks are of the ATM type. What types
What nodes can we use at levels 1, 2, 3, and 4? Show that we can.
use existing ATM networks and transform them into MPLS networks.

Exercise 3:

We consider a communication network that uses ATM cell switching. To

carry the information from A to B, the virtual path that is opened passes through
two intermediate nodes C and D.
Master Networks - ATMPLS 2008/2009

B
A

D
C
E

1. Indicate how the virtual circuit is set up.

2. Provide the routing tables of the cells in nodes C and D.

3. If D is a VP switch, show how it performs its switching.

4. Indicate the structure of the cell headers on the different interfaces.

5. Indicate how flow control is carried out.

6. If the error rate on the communication lines is bad, how

Are the necessary corrections made to maintain transmission quality?

Exercise 4

We consider a switching network. Two stations A and B have established communication.

through this network and we note the number of switches in the network crossed by it
communication between A and B. The rate of all links is Dbit/s. The protocol of
the communication used is the same on all connections, it adds a header of bits to
each unit of data. We neglect the propagation times and the processing times in
the network switches as well as the acknowledgments. Station A must transfer a
file of size L bits at station B.

The network uses message switching and the file is transmitted in one go.
Message on each link. Give the expression Tfic1 of the transmission time.
file on this network.

2. The network uses packet switching and the file is split into packets.
containing P bits of data. Give the expression for the transmission time of
file on this network.

[Link] numérique : L= 64 000 octets ;H= 5 octets ;S= 2 commutateurs ;D=

64 kbit/s; for the packet size, we will take two values: Pa = 128 bytes and Pb =
Master Networks - ATMPLS 2008/2009

16 octets. Calculate and compare the values obtained for Tfic1 and Tfic2 (for the
packet switching, we will compare the two possible packet sizes.
4. What are the advantages and disadvantages of packet switching compared to
the message switching?

5. The connections have a binary error rate denoted by τ. Show that the
the probability that a frame of length l is received correctly is given by p = (1-
τ)lDerive the average number N of frame transmissions (assuming that the
the control protocol repeats the frame indefinitely, without anticipation, until
that it be correct).

6. Redo the digital application of question 3) taking into account the error rate
τ= 10-4Conclude. Are these techniques suitable for high speeds? Why?
What solutions exist for such environments?

Annex: format of IP and ICMP packets

The IP header is aligned to 32-bit words. Its length is therefore a multiple of 4 bytes.
by default, without options, the IP header is 20 bytes long:
4 bits 8 bits 16 bits
4 bits
IHL TOS Total length
Version
Identification Flags Fragment offset
TTL Protocol Header checksum
Source address
Destination address
Options
Padding

Data

"Version" indicates the header format. This field is used for version identification.
current protocol. The version described here (and currently used) is number 4;
IHL (IP Header Length) is the length of the IP header expressed in 32-bit words.
(at least 5, at most 15) ;
TOS (Type Of Service) defines the type of service to be applied to the packet based on
some parameters like transit time, security.
"Total Length" is the total length of the datagram, including header and data.
expressed in bytes;
"Identification" is a value provided by the issuer helping in the reassembly of
different fragments of the datagram. The only use of this field is therefore to allow
for a receiving entity to recognize the datagrams that belong to the same
initial datagrams that must therefore be subject to reassembly;
Master Networks – ATMPLS 2008/2009

"Flags" is used by fragmentation. It consists of two indicators: DF

(Don't Fragment) to prohibit fragmentation and MF (More Fragment) for
signifying fragments to follow.

"Fragment offset" indicates on 13 bits the relative position of the fragment in the
initial datagram, the displacement being given in units of 64 bits (only a
a complete datagram or a first fragment of a datagram can have this field set to
0) ;
TTL (Time To Live) represents an indication of the upper limit of time for
life of a datagram. This value ranges from 0 to 255;
"Protocol" indicates the protocol (higher level) used for the field of
datagram data:
Code Abbreviation Protocol Name Reference
(Dec)
1 ICMP Internet Control Message Protocol RFC792

"Header Checksum" is an error control area that only concerns the en-
header of the datagram;
"Source Address" is the IP address of the source of the datagram.
"Destination Address" is the destination IP address of the datagram;

Format of an ICMP message