1

Tutorial for lectures 1-7

i) What encoding is used in the figure?

clock

signal

1 1 1 0 1 0 1 0

Manchester

ii) What is the encoding in the rightmost figure called?

Amplitude Amplitude

Time

Time
Analogue Signal

PAM
 2
Tutorial for lectures 1-7

iii)What encoding is used in the figure below?

clock

signal

1 0 1 0 1 0 1 0

Full cycle encoding

iv) Is the first bit a 0 or a 1 or indeterminate in the figure below? Explain your reasoning.

clock

signal

1 1 0 0 1 0 1 0

Differential Manchester

v) Sketch the encapsulation at each stage of the stack below assuming Ethernet at the Link
level.

Application
messages

Transport
TCP packets
Network IP Datagrams

Link
Frames for specific
network technology
TCP Data passing
 3
Tutorial for lectures 1-7

Application Identical messages Application

Transport Identical TCP packets Transport

Network
Network Router
Identical Identical
datagram datagram Link
Link
hop 1 hop 2
Interface
Identical Identical
Frame m1 Frame m2

Medium 1 Medium 2
TCP/IP Data-passing between remote nodes

vi) Add peer to peer communication and router stack in use in the figure above

vii) Nyquist tells us that R=2Bc is the basic bit rate over a communication channel of bandwidth
Bc .

If a discriminator can identify M different amplitude levels in a signal element how does the
relationship between R and Bc change to give us a value for R?

If M is 4 in your new relationship between R and Bc how would a transmitter organize the
bits of the following bit stream and what do you think the transmitter would send in an
analogue carrier system? Bit stream 0101111000 .

viii) Explain how the same effect to increase R is accomplished in QPSK.

ix) Explain Spread Spectrum Direct Sequence and sketch an example of a binary sequence being
transmitted.
x) Explain with the aid of a sketch how Go Back N ARQ works in Sliding Windows.
 1

Tutorial Questions

i) What is the bandwidth of the signal f(t) = 10 +(5cos 2t) + (sin 4t)
ii) What is the bandwidth of the signal [cos2000t + cos4000t + cos6000t +
cos8000t]
iii) What is the bandwidth of the signal f(t) = (5cos 2t)2 + sin5t
iv) What is the bandwidth of the signal [cos4000t + cos2000t + cos3000t +
cos7000t]
v) Determine the bandwidth in Hertz of the signal f(t) = 2 [cos3000t] [sin 1000t]
vi) What is the rate in bits/sec to transmit a picture frame of size 460 by 300 bits in
33ms?
Asynchronous transmission is best for this type of data. TRUE or FALSE
vii) Show one way in which 100 different characters or symbols may be represented by a
6-bit code.
viii) Show one way in which 90 different characters or symbols may be represented by a
5-bit code.
ix) How many signaling levels are required under the Nyquist formula to support a
Shannon channel capacity for a 2MHz bandwidth with a SNRdB of 30?
x) How many bits per signal level are required under the Nyquist formula to support a
Shannon channel capacity for a 1MHz bandwidth with a SNRdB of 20?
xi) What is the ratio of signal energy per bit to noise power spectral density per Hertz ,
for a link transmitting at 6000 bits per second when the effective noise temperature
is 1000 K and the signal level is -100dBW?
xii) What is the ratio of signal energy per bit to noise power spectral density per Hertz ,
for a link transmitting at 3000 bits per second when the effective noise temperature
is 500 K and the signal level is -80dBW?
xiii) The ISO application layer is equivalent to the TCP/IP application layer. TRUE or
FALSE.
xiv) Calculate the baud rate to transmit 19600 bps if 16 different signal element levels
are used.
xv) What is the payload data rate for a PDU of 10,000 bits transmitted at
100Megabits/second over a physical level with a header and trailer adding 500
octets in total to form a transmission packet?
xvi) What is the maximum channel capacity of a 10MHz bandwidth with a SNRdB of
36.1245 and how many signal element levels are required to support that channel
capacity?
xvii) Calculate the quantizing noise level in a telephone system which samples a voice
channel at 1.5 times the Nyquist rate and transmits the PCM signal at 96,000 bits per
second.
 2

xviii) Indicate the bit periods and the binary ones and zeroes for the signal below if it is
a) Manchester Differential (Transition at the beginning of a bit for a binary zero and
always a mid-bit transition )
b) Manchester (positive going mid-bit transition for a binary one and negative going
for a binary zero) .

xix)
a)What is the number of uniform quantization levels required to produce a SNR quantization
of at least 60 dB for the effective bandwidth of a telephone voice channel sampled at
the Nyquist rate? What is the bit rate of the resulting PCM signal?
b)What is the number of uniform quantization levels required to produce a SNR quantization
of at least 80 dB for a 6KHz voice channel sampled at the Nyquist rate?
What is the bit rate of the resulting PCM signal?
xx)
Does an Ethernet star connected LAN implemented with switches require the
CSMA/CD protocol? YES or NO
xxi)
The ATM protocol reads an entire frame before deciding if there is an error. TRUE
OR FALSE
 1
Tutorial for Lecs 8-13.

i) For the sequence 00 01 F2 03 F4 F5 F6 F7 calculate the Internet Checksum.

Show the result of a check on reception assuming no errors in transmission.
ii) a) For the message 1010101101 with a polynomial 110101 calculate the FCS
b) Show the transmitted signal is received correctly in the absence of errors in transmission.
iii) Repeat ii a) with a shift register
iv) a) For communication channel using optical fibre sending multiple frames of 1000 bits at a
data rate of 1 Mbps between two nodes 2000km apart calculate the efficiency of stop and
wait? Sketch a picture of the transmission.
b) If sliding windows were to be used what would be the minimum window size if 100%
efficiency is desired? Sketch a picture of the transmission.
v) In a communication carrier channel of bandwidth 256KHz how many individual baseband
data channels of 3dB bandwidth 4KHz with full bandwidth of 5Khz can be modulated into
the carrier channel bandwidth if a minimum guard band width is used?

vi) If a TDM system is to accommodate 24 voice channels each sampled at 8KHz and each
having 8 bit samples and all multiplexed into the TDM system what is the minimum data
rate of the TDM system?
If a framing bit were added in every sample interval (of 125µs) what is the required data
rate?
vii) In TDMA/FDD Uplink and Downlink time slots are typically not normally overlapping in time
for a specific subscriber station – TRUE or FALSE?
viii) What is the main advantage of packet switching over circuit switching?
Whose theorem was the underlying technical innovation for the move to packet switching
for voice transmission and what did the theorem say?
ix) In a multimode transmission system what advantages are gained if a large packet is
subdivided into smaller packets for transmission?
What is the limitation on the gains made?
x) What are the 5 layers in the 802 standard?
What are the fields in an LLC PDU?
What are the fields in an 802 frame?

xi) What fields are in the MTU of an 802.3 frame?

What fields are in the Data Frame of 802.3?
What fields are in addition to the data frame in 802.3?
What is the maximum size of an 802.3 frame?

xii) For a fixed frame size of 1526 bytes does efficiency go up or down for transmission between
two nodes if a user moves from 100Mbps ethernet to 100Gbps Ethernet? Explain.
 Practise questions : use crib and trig sheets
Q1:

i) Calculate the bandwidth in cycles/second of the following signal.

f(t) = (10sin 2t)2 + cos8t – cos4t

ii) What is the maximum length of a coax cable if a 100W signal at 1MHz must be
delivered over the cable with at least 1W of power remaining?

iii) What is the bit length of the FCS if a P of 10010101 is used?

iv) Demonstrate how to have at least 250 different symbols accommodated in a 5 bit
coding scheme.
v) Calculate the theoretical maximum bit rate of a 100MHz channel with a SNR of 1023.
vi) What is the number of bits required to ensure a quantization noise of between 44dB
and 50dB?

Q2

i) Of the encoding techniques, Manchester, AMI, NRZ

a) which has the greatest DC component, and

b) which has the widest bandwidth requirement?

ii) Encode the binary signal, 1 0 0 1 0 1 0 1 0, below into each of the encoding schemes
of Differential Manchester (DM) and Pseudoternary (PT) in the space provided.
1 bit

DM

PT
 Q3

Complete the first diagram below for an AMI code, then in the second give the HDB3
signal representation for the bit stream. Assume the previous insertion was
immediately to the left of the leftmost bit drawn in the diagrams.

1 1 0 0 0 0 0 1 1 0 0 0 0 1 1 1 0 0 1
+V

-V
1 1 0 0 0 0 0 1 1 0 0 0 0 1 1 1 0 0 1
+V

-V
-V

Q4
-Vi) A data rate of 19,200 bps is required over voice grade lines. To accomplish this it has
been decided that a form of multiphase modulation must be used. The voice channel can
only accommodate a baud rate of 2400 signal elements/second.
a) How many bits must be transmitted per signal element?
b) If up to 128 phase angles can be supported is it possible to implement the system and
if so how might it be accomplished?

ii) For the above solution and representing the level number in each signal element by the
binary weight of the bits transmitted in the signal element show the sequence of level
numbers that represent the following binary information.
1000 1111 0001 0101 0010 0000 1001
Q5

A system is required to transmit the baseband analogue signals from 40 users over an
available bandwidth of 4.4MHz (from 10MHz to 14.4MHz). The system design parameters
require that the transmission of each user’s modulated signal must be isolated by a guard-
band frequency of at most 10% of the user’s useful signal bandwidth (3dB bandwidth.)
 a) What bandwidth would each user have available for its baseband signal’s
useful bandwidth if the guard bands are required to be implemented at the
3dB points (useful bandwidth of the baseband signals.)
b) Sketch a frequency spectrum diagram of the system showing where the first
3 users’ baseband signals are placed into their own portion of the 4.4 MHz
bandwidth.

Q6 A B
F0
F1
F2
F3
time

RR1

a) Complete the diagram above to show how a sliding window handles a “Go back N” ARQ
scheme using RR and REJ. The sliding window is of size 7 and F1 is lost. Show the frames
sent by A to B assuming there are no more lost frames, after F1 as shown. Show a total of 7
numbered frames correctly received by B after recovery from the lost frame. Assume
timeouts are longer than 2*RTT.
b) From an examination of the figure what size of window would give a good chance of 100%
line utilization. Why?
Q7

In the figure below the stations are communicating using different flow control protocols.

Station A uses sliding windows and station B uses stop and wait. There are no errors. There are

no internal processing delays in stations.

i) Determine the delay per kilometer between A and B assuming the propagation delay

is a multiple of 2 milliseconds as indicated in the figure. Use the same propagation

delay between B and C.

ii) For maximum sliding window efficiency in the absence of errors what would you

expect the minimum size of Station A’s sliding window to be? Why?

iii) Station B uses Stop and Wait for traffic from A which has a sliding window size of 3

and B must keep up with A transmitting 2000 bit frames by clearing out its receiving

window buffer before a new set of sequence numbered frames is received.. How long

does B have to send the 3 frames and what bit rate must be used from B to C?

Support your answer with an approximate sketch on the figure.

A 500Km B 125Km C

Time with 2
milli second
increments
Q8

A TCP/IP stack is used to send user data from Station A to Station D through intermediate nodes
B and C. Station A use a 802.3 transmission frame as does Station D.
The network for which B and C are edge routers is a packet switching X.25 network.
Station A’s IP layer is restricted to passing 500 octet IP-PDUs to the LLC.
Assume the TCP layer at Station A passes a TCP-PDUof 9,600 octets to its IP layer.
i) Determine the number of IP-PDU’s to be passed to the LLC layer.
ii) Determine the largest payload provided to an Ethernet frame (assume LLC control of 1
octet)
iii) Determine the TCP-PDU effective bit rate of A, relative to the 802.3 bit rate of R bps.

Q9
A user signal is to be sampled at 1.5 times the Nyquist rate. The signal’s highest frequency is
5000Hz.
i) For quantization noise level of less than 62dB but higher than 56dB determine how
many quantizing levels are required.
ii) Determine the resulting bit rate of the PCM signal.
iii) If a transmission channel bandwidth of 18 kHz is available how many bits per signal
element must be implemented. What actual bit rate would that number of bits
support?
iv) If the channel is noisy what is the maximum theoretical noise level in dB to carry the
required bit rate?
 1

Introduction to Computer Networks

Tutorial/Study questions:

Q1
i) A new standard is to be developed for voice sampling in the telephone
system. Sampling the 3000Hz bandwidth at twice the Nyquist rate and
using 64 quantization levels determine the bps to multiplex 24 PCM coded
voice channels using TDM. What is the required bandwidth for
transmission of the TDM signal so that the bits can be easily recovered?
Explain your reasoning.

ii) An analogue signal with a bandwidth of 5 kHz is to be sampled at 1.5

times the Nyquist rate and then transmitted over a line in a digital PCM
format. Determine the bit rate of the required PCM signal if the SNRquant
is to be at least 60dB and the minimum number of bits per sample is used.

iii) For ii) how many bits per signal element must be transmitted if the
available bandwidth is 12.5kHz? Under what line condition would it be
theoretically feasible to transmit that number of bits on the available
bandwidth? Do you think it would work in practice and if so/not why?

Q2
Ignoring switching delay determine the maximum possible data throughput when
HDLC is used to transmit 1024 bit frames with 3 bit sequence numbers and a 16
bit FCS across a 1Mbps data rate link with a propagation delay of 270ms?
Q3

Apply Dijkstra's algorithm to the network shown above and complete the least
cost routing table for node s = 5 as the source node.
 2

Iteration T L(1) path L(2) path L(3) path L(4) path L(6) path

Q4: Dedicated lines exist between location A and location B, and location B and
location C. The distance between A and B is 5000 km and between B and C is
1500 km. The technology used for the media provides a propagation time of 10
microseconds per kilometer between location A and location B, and 5
microseconds per kilometer between location B and location C.

If the data rate between A and B is 200 kbps and window size is 4 and packets of
125 8 bit bytes is it possible to choose a minimum data rate between B and C to
avoid B’s buffer’s from overflowing? Explain your reasoning. If it is possible
what is that data rate?
Q5.
A TCP PDU of length 60000 octets (including TCP header) is to be transmitted
over a physical link between two nodes using 10G Ethernet (10Gbps). Assume
for the purposes of this question that a standard Ethernet data frame is between 64
and 1518 bytes and the MTU of the payload is 1500 bytes. Assume the LLC
PDU has an overhead of 4 octets.

a) How many bits are sent over the 10Gbps Ethernet if standard Ethernet frames
are used? What is the effective data rate from the TCP PDU point of view?
b) If the propagation time between the two nodes is 12 microseconds what is the
line efficiency assuming stop and wait is used?
c) In one sentence say what can be done to increase line efficiency given that this
is 10G Ethernet? By calculation show how your suggestion can increase line
efficiency.

Q6
Using variable sub-nets create

i) A class B address 129.108.xxx.xxx that has 2 bits in the sub-net portion of the address.
Write out the binary pattern for the address
Write out the Subnet mask in binary and then in dotted decimal
How many subnets are supported?
How many hosts are supported?
ii) A class B address 129.108.xxx.xxx that has 3 bits in the sub-net portion of
the address.
Write out the binary pattern for the address
Write out the Subnet mask in binary and then in dotted decimal
 3

How many subnets are supported?

How many hosts are supported?

Q7
If a system is planned using the following parameters
N = number of hops between two given end systems = 4,
L = message length in bits 3000 bits
B = data rate in bits per second on all links= 9600 bps
P = fixed packet size in bits = 92 bits
H = header overhead in bits per packet 32 bits
S = call setup time (circuit switching or virtual circuit) in seconds = 0.2 secs
D = propagation delay per hop in seconds = .001 secs
assuming no acknowledgements and ignoring any processing delay at the nodes.
Then
i) calculate the end-to-end delay for circuit switching

iii) calculate the end-to-end delay for datagram packet switching.

Q8
Consider an optical slotted ring of length 15km with a data rate of 7.5Mbps and
750 repeaters. Each repeater introduces a 1-bit delay. Each slot has one source
address of 10 bits, one destination address of 10 bits, two data bytes, and five
control bits. How many slots can the ring support?

Q9 Sketch the Bipolar-AMI encoding for the binary stream 0 0 1 0 1 0 0 0 0 1 0 0 0 0

0 0 1 then draw the High Density Bipolar 3 zeroes encoding of the Bipolar-AMI
signal. Assume the previous HDB3 substitution was immediately to the left of the
leftmost binary zero in the above binary stream.

Q10 What is the overhead as a percentage of the total bits sent if an asynchronous
encoding scheme with 8 data bits, no parity, one start and 2 stop bits is used to
send 10,000 characters at 1Mbps? What is the effective data transmission rate?

If the 10,000 characters were sent in a synchronous manner by discarding the

asynchronous format and adding some synchronous control could this improve
the efficiency and the data rate? Support your answer with an example set of
numbers for payload efficiency and effective data rate for HDLC. Assume
HDLC carries 1000 characters per frame.

Q11. For a FCS the polynomial X5 +X4 + X2 + 1 is to be used on the data stream (MSB
first)

M = 1 0 1 0 1 0 1 1 0 1.
i) Find the FCS by long division
ii) Draw the shift register diagram for the calculating the FCS.
 4

Q12. A datagram transmission system has 3 hops between nodes 1 and 4. Each node uses its
own transmission bit rate as shown below. Assume multi-pair cables are used between
nodes with appropriate signal regeneration ( regeneration gives zero delay.)

UTP STP STP

1Km 2Km 2km
100Mbps 150Mbps 200Mbps

Determine the elapsed time to send 10 packets of 1000 bits from the 1st node (leftmost)
into the last node.
Assume the 1st node sends 10 packets one after the other without a break. Ignore ACKs.
Assume the remaining nodes send packets as quickly as possible after a packet arrives.
Comment on any buffering assumptions that are necessary.

Q13. A Transport PDU of 5108 octets (including TCP header) is to be transferred between
two users A and B using TCP/IP. Both users A and B’s LANs are Ethernet 10Mbps at
the physical level and are part of the same LAN connected by a bridge

Determine the effective data rate for the application that provided data to the Transport
level if the Maximum Ethernet Frame size permitted at the physical level of A's LAN is
846 octets and variable length packets are permitted.

Q14:
Assume a packet transmission system is designed to pass packets through intermediate
stations and over the links between the stations.

Station 1 to station 2 – fibre over a distance of 2.5km

Station 2 to station 3 - fibre over a distance of 5 km
Station 3 to station 4 - fibre over a distance of 10km

i) Assume the data rate used at each node is 800Mbps, that a sliding window large enough
to hold all the required packets is available in each node, that ACKS arrive as necessary
to maintain data flow, node processing time is negligible, and there are no errors.

Find the elapsed time end-to-end when sending 10,000 octets of data using an ethernet
frame of 1026 octets (assume the data load is sent directly to the physical level after a
fragmentation process at the boundary of the physical level only.) Frames must be full
size.

ii) Now assume the network manager changed only the node 3 to node 4 transmission to
stop and wait with a bit rate of 8Gbps with negligible sized ACK packets. Node 3 has
 5

sufficient buffering to hold as many frames as are required.

Show by calculation the effect on the end to end delay.

Q15. Complete the diagram below by replacement of the appropriate strings of zeroes
to give the HDB3 signal representation for the bit stream. Assume AMI and the last
insertion to the left of the diagram is an odd number of pulses away from your first
insertion.
+V
1 1 0 0 0 0 0 1 1 0 0 0 0 1 1 1 0 0 1

-V

Q16. How many signaling levels are required to support a Shannon communication
channel for a 2MHz communication bandwidth with a SNRdB of 30?

Q17. A new standard is to be developed for voice sampling in the telephone system.
Sampling the 4000Hz bandwidth at twice the Nyquist rate and using 64
quantization levels determine the bps to multiplex 24 PCM coded voice channels.
What is the required bandwidth for transmission of the multiplexed signal so that
the bits can be easily recovered? Explain your reasoning.

Q18 a) 5000 bytes (including TCP header) are to be transferred between two users A
and B using TCP/IP. Both users A and B’s LANs are 802.3 10Mbps at the
physical level. A’s LAN connects to an ATM network through a router. B’s
LAN also connects to the same ATM network through a router on B’s LAN.

Calculate the MTU if 8 frames are sent on A’s LAN to transmit the TCP-PDU
assuming error free transmission if the IP PDU is fragmented to be carried by
802.3. Determine the efficiency on A’s LAN from the viewpoint of the transport
level. Assume the only option used is the Internet timestamp and that one 32 bit
field is sufficient for this option.

Q19 Assume a packet transmission system is designed to pass packets through

intermediate stations and across the medium between the stations according to the
following specification (assume signal repeaters are installed as required and have zero
delay.).

Station 1 to station 2 - multipair twisted pair over a distance of 4km

Station 2 to station 3 - multipair twisted pair over a distance of 2km
Station 3 to station 4 - coax cable over a distance of 9km
 6

Repeaters delay the signal 0 bit time (assume attenuation is compensated for by
repeaters.)
The data rate used is 1Mbps

Find the end-to-end delay for 40,000 bits of data, a header of 100 bits, and a packet size
of 15000 bits. Packets must be padded to full size and are re-transmitted immediately
without delay by intermediate station (i.e. as soon as the last bit arrives.)

Q20: A data rate of 19,200 bps is required of voice grade lines. To accomplish this it
has been decided that a form of QFSK must be used. The channel can only
accommodate a baud rate of 2400 baud.

a) How many bits must be transmitted per signal element?

b) If only 64 phase angles can be supported is it possible to implement the
system and if so how might it be accomplished?

For the above solution, and representing the data in each signal element by the binary
weight of the bits transmitted in the signal element, show the series of weights which
would indicate the particular QFSK signal to be transmitted for the following binary
sequence:

1000 1111 0000 1010 1000 0000 0100 0001 0000 1111 1100 0000
 7

Q21 Complete the following diagram to show how a sliding window handles a
“Go back N” ARQ scheme using RR and REJ. The window is of size 4
and F1 is lost.
i) Show the ACK responses and the frames sent by A to B assuming
there are no more errors, after the one shown, until a total of 6 frames
are sent and correctly received and acknowledged.
ii) The ACK for the 7th frame is then lost. Complete the diagram to show
the flow of messages and ACKs until A receives a subsequent ACK
from B in response to a query from A of RR with P bit =1.
A B

F0
F1
F2
F3
X time

F4
RR3

c) Would selective reject be a better scheme to use? Explain your answer.

 8

Q22

For a burst period of 250ms with a single burst of 50ms duration delivering 0.5Mbits

and a departing packet rate of 10Mbps with a bucket capacity of 1Mbit determine

what R should be to enable the bucket to keep up with the incoming bursts. The

bucket is full to begin with and the 50ms burst occurs in the first part of the 250ms

period.
 9

Q23
The token bucket is full to begin but may not be full at the beginning or end
of a period once started and any bits not sent out in a period are buffered and
sent out first in the following period.
Given B = 2Mbits (bucket capacity) , N = 15Mbps (incoming burst rate),
R = 2.5Mtps (token rate), M = 5Mbps (outgoing rate) , S = 0,25secs
(period) , and t= 0.1 secs (burst duration) : For a token bit bucket with
maximum capacity B tokens, burst rate N bps, incoming token rate R
tps, output rate M bps, period S seconds , and burst duration t seconds
at the start of a period find
i)How many periods can the bucket keep working and sending out MS
bits per period? Hint: Each bit sent consumes a token,
ii) How many bits are buffered at the end of the last period identified in
i), and iii) How many tokens are left in the bucket at the end of the third
period?

Answers
i) 3
ii) 750,000 bits
iii) 125,000 tokens

Q24.
i) A new standard is to be developed for voice sampling in the
communication system. Sampling a channel of 5000Hz bandwidth at
twice the Nyquist rate and using 64 quantization levels determine the
resulting bps for the channel.
ii) Now determine the bps of a multiplexed line to multiplex 24 PCM coded
voice channels using TDM ignoring any overhead control bits
iii) Calculate the required bandwidth for transmission of the TDM signal so
that the bits can be easily recovered? Explain your reasoning.

Q25.
10,000 bytes of Transport PDU (including the TCP header) are to be transmitted
over the three hops described below when the MTU is 1000 bytes and 802.3 at
200Mbps is used with a TCP/IP stack. The LLC-control uses 1 byte and no
options are required for the IP.
 10

a) How many 802.3 frames must be sent?

b) Calculate the overhead in bytes, excluding the TCP Header, for the entire
transmission assuming no errors.
c) The four stations use sliding windows, the three hops use optical fibre with
20km between the nodes. The Go Back N ARQ is in use at all nodes and the
sequence number length is 3 bits.
Assume no transmission errors and each node ACKS the first frame received
and at least every 4th frame thereafter. Assume each node retransmits a frame
with no delay.

Is the transmission line between the first two nodes 100% utilized at any time or not?

Explain why or why not.

Q
The token bucket is full to begin but may not be full at the beginning or end of a
period once started and any bits not sent out in a period are buffered and sent
out first in the following period.
Given B = 2Mbits (bucket capacity) , N = 15Mbps (incoming burst rate), R =
2.5Mtps (token rate), M = 5Mbps (outgoing rate) , S = 0,25secs (period) , and t=
0.1 secs (burst duration) : For a token bit bucket with maximum capacity B
tokens, burst rate N bps, incoming token rate R tps, output rate M bps, period S
seconds , and burst duration t seconds at the start of a period find
i)How many periods can the bucket keep working and sending out MS bits per
period? Hint: Each bit sent consumes a token,
ii) How many bits are buffered at the end of the last period identified in i), and
iii) How many tokens are left in the bucket at the end of the third period?

Answers
i) 3
ii) 750,000 bits
iii) 125,000 tokens

Q.
i) A new standard is to be developed for voice sampling in the communication
system. Sampling a channel of 5000Hz bandwidth at twice the Nyquist rate and
using 64 quantization levels determine the resulting bps for the channel.
ii) Now determine the bps of a multiplexed line to multiplex 24 PCM coded voice
channels using TDM ignoring any overhead control bits
iii) Calculate the required bandwidth for transmission of the TDM signal so that the
bits can be easily recovered? Explain your reasoning.
Q.
10,000 bytes of Transport PDU (including the TCP header) are to be transmitted over the
three hops described below when the MTU is 1000 bytes and 802.3 at 200Mbps is used
with a TCP/IP stack. The LLC-control uses 1 byte and no options are required for the IP.

a) How many 802.3 frames must be sent?

b) Calculate the overhead in bytes, excluding the TCP Header, for the entire
transmission assuming no errors.
c) The four stations use sliding windows, the three hops use optical fibre with 20km
between the nodes. The Go Back N ARQ is in use at all nodes and the sequence
number length is 3 bits.
Assume no transmission errors and each node ACKS the first frame received and at
least every 4th frame thereafter. Assume each node retransmits a frame with no
delay.

Is the transmission line between the first two nodes 100% utilized at any time or
not? Explain why or why not..