You are on page 1of 7

MASTER IN TELECOMMUNICATIONS ENGINEERING

COMMUNICATION NETWORKS Final exam Duration: 2h 30 min


Provisional marks: 20/01/22 Date to see the test: 21/01/22; 10h-11h; Meet Allegation deadline: 23/01/22
Mobile phones must be switched off and out of reach
Professors: J. Casademont, X. Hesselbach
Programmable calculators cannot be used
Student's Name:
Questions must be answered in the blank space.
Question 1 (1 point): Regarding the study of the capacity of the standard 802.11 made with Markov chains, comment on the
conclusions that can be obtained from the following graphs where: n is the number of machines present in the BSS, IR stands
for Infrared, DS for Direct Sequence and FH for frequency hopping.

Question 2 (0,75 points): Explain the differences between the MAC algorithm of IEEE802.15.4 and the MAC algorithm of
IEEE802.11.

Question 3 (1 point): Explain the differences between IPv4 header and IPv6 header (scheme of headers in additional
information).

Question 4 (1 point): Explain the working procedure of the Dynamic Source Routing (DSR) algorithm for routing in mobile
adhoc networks.
MASTER IN TELECOMMUNICATIONS ENGINEERING

Problem 1 (3 points):

A network administrator has installed two access points (AP) running on the same radio channel in an industrial building. One of
the access points has two computers connected using 802.11b "short-format" mode and transmitting at 11 Mbps. The other
access point also has two computers connected but using 802.11g mode and transmitting at 36 Mbps.

As the industrial building has no walls or obstacles, the coverage area of the two formed BSS overlap completely. That is, all
devices see all other devices.
Switch
Internet
IEEE 802.3
1 Gbps

802.11b
11 Mbps 802.11g
Short Format 36 Mbps

a) (0.75 points) Explain the implications of this configuration by commenting especially on whether the two 802.11 networks can
function in this way.

In addition, the following data is available:


- Unicast packets have an average MAC payload field length of 500 bytes.
- Broadcast packets have an average MAC payload field length of 100 bytes.
- Consider that the 802.11 header does not use the "QoS Control" field.
- RTS / CTS is not used.
- Broadcast packets are sent at the minimum speed of the operating mode.
- Consider that there are no collisions or transmission errors.

b) (0.75 points) Suppose for a moment that the two BSS are configured to work in different channels and without overlap between
them. If computers only transmit unicast data to Internet and there is no other type of traffic, what would be the maximum
aggregated throughput of each of the two BSS?

Aggregated throughput is the sum of the throughput of all devices within each BSS. Because all stations in each of the
BSSs transmit the same type of packets, we can calculate the aggregated throughput by simply dividing the number of
bits of data in a packet by the total time spent transmitting that packet.

Packet transmission time is: Time = DIFS + BACKOFF + T_DATA + SIFS + T_ACK

Stations use backoffs because they transmit continuously.


The MAC header does not use address 4, or the HT Control field, or the QoS field. Therefore, it has a total length of 28
bytes.
The 8-byte LLC header is used.

Because the two BSSs are completely independent, BSS 802.11b will use 802.11b settings, and BSS 802.11g will use
802.11g settings.

BSS 802.11b:
Time11b = 50 µs + 310 µs + (96 + 536 * 8 / 11) µs + 10 µs + (96 + 14 * 8 / 11) µs = 962,0 µs

Throughput_aggregated_11b = 500 * 8 bits / 962,0 µs = 4,16 Mbps

BSS 802.11g:
Temps11g = 28 µs + 67,5 µs + (26 + é(22 + 536 * 8) / 144ù * 4) µs + 10 µs + (26 + é(22 + 14 * 8) / 144ù * 4) µs = 281,5
µs

Throughput_aggregated_11g = 500 * 8 bits / 281,5 µs = 14,21 Mbps


MASTER IN TELECOMMUNICATIONS ENGINEERING

c) (0.75 points) If we now consider that the two BSSs actually use the same radio channel, what is the maximum aggregated
throughput of each of the two BSS.

In this case, when a computer of one BSS is transmitting, the computers of the other BSS will not be able to transmit
because they will sense the channel occupied. Therefore, since the number of computers in each BSS is the same, we
can consider that, on average, alternate packets of each of the BSS will be transmitted:

Packet_BSS_11b - Packet _BSS_11g - Packet _BSS_11b - Packet _BSS_11g ...

(A more precise approach would have to contemplate that, as 802.11g has smaller DIFS and backoff, they would have
a higher access probability than 802.11b, and the transmission pattern would not be 50%-50%. Nevertheless, we do not
consider this aspect in the solution).

Then the aggregated throughput of a BSS will be: Thr_ag = Length_data / (Time11b + Time11g)

Since the average length of the data field of the packets sent in each BSS is the same, the aggregated throughput of the
two BSSs will also be the same and will be:

Throughput_aggregated_11g = Throughput_ aggregated _11b = 500 * 8 bits / (962,0 µs + 281,5 µs) = 3,22 Mbps

d) (0.75 points) Based on the situation of case b) where the radio channels are different, if now each computer transmits, in
addition to the unicast traffic, 100 packets / second of broadcast traffic at MAC level, what is the maximum aggregated throughput
of unicast traffic of each of the two BSS?

If we consider that broadcast packets must always be sent, we can perform the calculation of the unicast packet
throughput calculating the remaining time that each BSS has to transmit unicast packets.

We observe what happens in a period of, say, 1 second. We calculate how long it takes to send broadcast packets, and
the rest will be for unicast packets.

The number of broadcast packets per second in each BSS is 600. Each BSS has to upload 200 packets_broadcast /
second and has to download 400 packets_broadcast / second.

Once we have the time devoted to transmit unicast traffic during the 1 second period, we calculate how many unicast
packets we can send during it, and finally we look for the throughput multiplying by the data bits contained in each packet.

Broadcast packets do not use acknowledgement and are sent at the minimum speed of the mode used, which for 802.11b
short-format is 2 Mbps and for 802.11g is 6 Mbps.

Broadcast packets are sent twice over the radio channel.

BSS 802.11b:
Time11b_broad = 50 µs + 310 µs + (96 + 136 * 8 / 2) µs = 1000,0 µs
Time_left_for_unicast_11b = 1 sec - 600 packets_broadcast * 1000 µs / packets_broadcast = 400000 µs
Number of unicast packets sent in 1 second = 400000 µs / 962 µs = 415,8 packets
Unicast data bits sent in 1 second = Throughput aggregated BSS 802.11b = 415,8 packets * (500 * 8 bits/packet) =
1,66 Mbps

BSS 802.11g:
Time11g_broad = 28 µs + 67,5 µs + (26 + é(22 + 136 * 8) / 24ù * 4) µs = 309,5 µs
Time_left_for_unicast_11g = 1 sec - 600 packets _broadcast * 309,5 µs/ packets_broadcast = 814300 µs
Number of unicast packets sent in 1 second = 814300 µs / 281,5 µs = 2892,7 packets
Unicast data bits sent in 1 second = Throughput aggregated BSS 802.11g = 2892,7 packets * (500 * 8 bits packet) =
11,57 Mbps
MASTER IN TELECOMMUNICATIONS ENGINEERING

Problem 2 (1,25 points): An SDN controller is calculating the Fair Rate using Max-Min criteria for the greedy connections
established between each pair source-destination {Si, Di} of the following figure.

#1 – 10 Mbps D1, D3, #3– 15 Mbps #4– 12 Mbps D4, D5,


S1, S2 D2, S6
S5 D6

#2 – 12 Mbps
S3, S4

a) Indicate the type of information the controller needs to collect from the network to be able to calculate the fair rate.
b) Calculate the values justifying the procedure, and fill in the following table:

#Source 1 2 3 4 5 6
Fair Rate

c) Indicate the spare (unusable) bandwidth of the links.

#Link 1 2 3 4
Spare bandwidth

Solution:

a) The controller needs to get the following from the switches:


• the capacity of the links (about the interfaces of the switches)
• the path of each connection.
b)
#1: S1,2 = 10/2 = 5 Mbps
#2: S3,4 = 12/2 = 6 Mbps
#3: S2,4,5 = 15/3 = 5 Mbps
#4: S4,5,6 = 12/3 = 4 Mbps ß bottleneck

#3: S2 = [15 – (4+4)] / [3-2] = 7 Mbps à link #1 is bottlenecking S2.


#2: S3 = [12 – 4] / [2-1] = 8 Mbps

#Source 1 2 3 4 5 6
Fair Rate 5 5 8 4 4 4

c)
#Link 1 2 3 4
Spare bandwidth 0 0 2 0
MASTER IN TELECOMMUNICATIONS ENGINEERING

Problem 3 (1,25 points):

The network design done by an operator in a rural area to deploy GPON using fiber-to-the-home, is providing a guaranteed fixed
rate of 100 Mbps to each of the 1218 subscribers.
Technical features:
Each OLT can support up to 128 subscribers.
The DBA has been implemented using frames of 38912 bytes every 125 µs (upstream and downstream, both), supporting
128 fixed size GEM frames, assuming no headers.

a) Calculate the maximum number of subscribers supported actually by each OLT.


b) Using a single splitting stage from OLT to ONTs, calculate the number of splitters required to cover completely the
population.
c) Calculate the number of ports required in the splitters.
d) Calculate the splitter attenuation from OLT to each ONT.

Solution:

a) Single GEM frame rate = (38912 ·8 / 125 µs) / 128 GEM frames = 19.46 Mbit/s.
GEM frames consumption per subscriber = Ceiling (100 Mbps / 19.46 Mbit/s) = 6 frames.
Subscribers supported in each OLT = Int (128/6) = 21.
b) OLTs required = Ceiling (1218 / 21) = 58. à 58 splitters required.
c) All the splitters: 21 ports, all in use.
d) Attenuation = 10 log 21 = 13.22 dB.

Problem 4 (0,75 points): According to the modulation used in ADSL (G.992.1) (256 carriers, 4.3125 KHz each, 4000 bauds and
up to 15 bits/symbol):
a) Calculate the granularity (minimum step in bitrate).
b) Propose a solution to make the granularity thinner while keeping the total rate provided. Show a numerical example.

Solution:

a) Granularity = 1 bit / symbol x 4000 baud= 4 Kbps.


Max. bit rate per carrier = 4000 x 15 = 60 Kbps.

b) Reduce the baud rate and increasing the maximum bits/symbol. For instance:
Baud Rate = 2000; 30 bits / symbol. à 2000 x 30 = 60 Kbps. Now granularity = 2 Kbps.
MASTER IN TELECOMMUNICATIONS ENGINEERING

ADDITONAL INFORMATION:
IPv4 and IPv6 headers:

HEADERS LENGTH IN BYTES: IP = 20; TCP = 20; UDP = 8; RTP = 12

IEEE 802.11 Data:


IEEE
Std 802.11-2007 LOCAL AND METROPOLITAN AREA NETWORKS:SPECIFIC REQUIREMENTS

7.2.1.3 ACK frame format


Standard DIFS (µs) SIFS (µs) SLOT (µs) CWmin
IEEE CWmax
Std 802.11-2007 LOCAL AND METROPOLITAN AREA NETWORKS:SPECIFIC REQUIREMENT
802.11 50 10 20 The frame format
31 for the ACK frame is as defined in Figure 7-8.
1023

802.11b 50 10 20 The 31
subfields within
1023 the Frame Control field of control frames are set as illustrated in Figure 7-5.

802.11a 34 16 9 15 1023
B0 B15
802.11g 28 / 50 10 9 / 20 15 / 31 1023
Protocol Type Subtype To From More Retry Pwr More Protected Order
28 / 50 (2,4 GHz) 10 (2,4 GHz) 9 / 20 (2,4 GHz) 15 / 31 (2,4 GHz)
Version DS DS Frag Mgt Data Frame
802.11n 1023
34 (5 GHz) 16 (5 GHz) 9 (5 GHz) 15 (5 GHz)
Protocol Pwr
Control Subtype 0 0 0 0 0 0 0
802.11ac 34 16 9 15
Version 1023 Figure 7-8?ACK frame Mgt

Bits : 2 2 4 1 1 1 1 1 1 1 1
Standard Data rates (Mbps)
The RA field of the ACK frame is copied from the Address 2 field of the immediately previous directed
802.11 1; 2; data, management, BlockAckReq control,Control
Figure 7-56Frame BlockAck control,
field or PS-Poll
subfield valuescontrol
withinframe.
control frames

802.11b long 1; 2; 5,5; 11; For ACK7.2.1.1


framesRTS
sent frame
by non-QoS
formatSTAs, if the More Fragments bit was set to 0 in the Frame Control field of
802.11b short 2; 5,5; 11;
the immediately previous directed data or management frame, the duration value is set to 0. In all other ACK
frames, the
Theduration valuefor
frame format is the RTS
valueframe
obtained
is asfrom theinDuration/ID
defined Figure 7-6. field of the immediately previous data,
802.11a 6; 9; 12; 18; 24; 36; 48; 54; management, PS-Poll, BlockAckReq, or BlockAck frame minus the time, in microseconds, required to
transmit the ACK frame and its SIFS interval. If the calculated duration includes a fractional microsecond,
802.11g 6; 9; 12; 18; 24; 36; 48; 54;
that value is rounded up to the next higher integer.
20 MHz: (from 6,5 Mbps to 288,9 Mbps) [Std 802.11-2012, pag: 1771]
802.11n
40 MHz: (from 13,5 Mbps to 600 Mbps)
7.2.1.4 PS-Poll frame format
20 MHz: (from 6,5 Mbps to 693,3 Mbps) [Std 802.11ac-2013, pag: 324]
40 MHz: (from 13,5 Mbps to 1600 Mbps) IEEE
802.11ac The
PART
80 MHz: (from 29,3 Mbps to 3466,7 Mbps)
frame format forLAN
11: WIRELESS theMAC
PS-Poll
AND frame is as defined in Figure 7-9.
PHY SPECIFICATIONS Std 802.11-2007
160 MHz: (from 58,5 Mbps to 6933,3 Mbps) Figure 7-66RTS frame
Standard Physical header duration 7.2.1.2 CTS frame format
802.11 192 µs
The RA field of the RTS frame is the address of the STA, on the WM, that is the intended immedia
802.11b (long) The frame formatoffor
recipient thethe CTS frame
pending is as
directed defined
data in Figure 7-7.
or management frame.
802.11b (short) 96 µs
The TA field is the address of the STA transmitting the RTS frame.
802.11a (20 MHz) 16 µs + 4 µs + (16-bit service + 6-bit tail in MAC header)
802.11g (20 MHz) 16 µs + 4 µs + (16-bit service + 6-bit tail in MAC
For header) 6 µs sent by non-QoS STAs, the duration value is the time, in microseconds, required t
all RTS+frames
Figure 7-9?PS-Poll frame
802.11n transmit
30 µs + variable (depending on the number of the pending
MIMO streams) + data or management frame, plus one CTS frame, plus one ACK frame, plus three SIF
(variable number of bytes in MAC header) intervals. If the calculated duration includes a fractional microsecond, that value is rounded up to the nex
The BSSID is the address of the STA contained in the AP. The TA field is the address of the STA
higher
802.11ac 40 µs + variable (depending on the number of MIMOinteger. For+ all RTS frames sent by STAs under EDCA, following a contention access of the channe
streams)
transmitting
(variable number of bytes in MAC header) the frame.
the duration The
value AID
is set is the
in the value
following
Figureassigned
manner:
7-75CTSto the STA transmitting the frame by the AP in the
frame
association response frame that established that STAQs current association.
N If the NAV protection is desired for only the first or sole frame in the TXOP, the duration value
When thevalue set
CTS alwaysto the
frame time, an
follows in microseconds, therequired
fieldtoof
transmit theframe
pending frame,from
plus the
one TA
CTSfield
frame,
of plu
The AID its twoRTS
hasframe MSBs frame,
each
plusset
RA
to 1. the CTS is copied
the immediatelyone ACK
previous RTS if required,
frame to which three
the SIFS
CTS is intervals.
a response. When the CTS is the first frame in a
N Otherwise,
frame exchange, the is
duration value is set to the of
remaining duration of the TXOP.
7.2.1.5 CF-End the RA field
frame format set to the MAC address the transmitter.

For all For


CTSallframes
RTS frames
sent insent under HCCA,
response to RTStheframes,
durationthevalue is set to
duration one is
value of the following values:from the
value obtained
The frame format for the CF-End frame is as defined in Figure 7-10.
Duration Nfield Ifofthe
thepending
immediately
frame previous RTS
is the final frame,
frame, theminus the value
duration time, isinset
microseconds, required
to the time, in to
microsecond
transmit the CTS frame to
required and its SIFS
transmit theinterval.
pendingIfframe,
the calculated duration
plus one CTS frame,includes
plus onea fractional
ACK frame microsecond,
if required, plu
that value is rounded up tointervals.
three SIFS the next higher integer.
N If the pending frame is not the final frame in the TXOP, the duration value is set to the remainin
At a non-QoS STA, if the
duration of CTS is the first frame in the exchange and the pending data or management frame
the TXOP.
MASTER IN TELECOMMUNICATIONS ENGINEERING

DSAP SSAP Control Protocol Organization Protocol ID


8 bits 8 bits 8 bits 24 bits 16 bits

You might also like