TEST 1

HO CHI MINH CITY UNIVERSITY

OF TECHNOLOGY Course: System Performance
Faculty of Computer Science & Technology Evaluation (CO3007)
Class: 20201 Group: CC01
Time duration: 30 minutes
(open test)
Date: Oct 27, 2020

Student name: Student no.:

Score (number): Examiner:

Score (words): Invigilator:

(The test has 20 questions which must be answered by the best choice.)

Q. 1 [L.O.1]. Why is the step of workload selection conducted after the step of selecting evaluation
technique ?

A Not much important. Workload selection can be conducted before evaluation tech-
nique selection.
B Workload depends strongly on evaluation techniques.

C Different workload selection are for different evaluation techniques.

D Workload selection is performed only if performance metrics, factors, and evaluation
techniques are known.
Q. 2 [L.O.3] Consider a single queue with an infinite queueing length. The arrival process is bursty
(i.e., traffic with a sudden increase). Which of the following is correct to departing traffic of the
queue in steady condition ?
 
A Departing traffic is still bursty. B Departing traffic is changed to real-time
 traffic.
C De[arting traffic (rate) is less than arrival D The property of being bursty is decreased
traffic ]. in departing traffic.
Q. 3 [L.O.2]. We consider a local area network (LAN) providing a video conferencing service on virtual
circuits for end users (rectangle boxes). The LAN bounded within the rectangle is the system under
test (SUT).
u5
d The virtual circuits are given as follows.
u1 u6
u2 a c u7 Virtual circuit Path Allocated bandwidth
VC1 u1 − a − c − d − u5 10Mbps
u3 u8 VC2 u3 − b − c − d − u5 3Mbps
b
VC3 u7 − d − c − u8 8Mbps
u4
Which of the following is Jain’s fairness index of the bandwidths allocated to VCs by SUT ?
   
A 0.240 B 0.333 C 2.549 D 0.849

Q. 4 [L.O.3] Which of the following is correct with Little’s Law ?


A Applicable for any closed system.

B Applicable to any closed system without request/job lost inside the system.

C Applicable to any system working in transient state.

D To describe the relationship between arrival rate and departing rate.

Q. 5 [L.O.3] Consider 2 following single queues: (Q1) M/M/3/100/100 and (Q2) M/M/3/300/100.
Which of the following is correct w.r.t. performance comparison?
 
A Q1 is better than Q2. B Q2 is better than Q1.
 
C N/A. It depends on arrival rate. D Q1 is equivalent to Q2.

Head of department: . . . . . . . Test code CLC.2010 Page 1

 Q. 6 [L.O.3] Which of the following systems is impossible (or inappropriate) to be modeled as a queue-
ing system ?

A Operational processes of a software development company.

B A mechanism to manage processes in OS kernel.

C Flying activities of an UAV.

D A distribution network from manufacturers to their distributors.

Q. 7 [L.O.3] During an hour of measurement, the number of requests to a domain name server is 10800.
The average response time of the server is 1/3 second. What is the average number of requests
being processed within the server?
   
A 3600 B 1 C 32400 D 3

Two following questions use the same problem description as follows.

Consider a queueing system X as shown in the following picture. X receives the incoming traffic as a
Poisson process with rate λ. p1 , p2 , p3 , p4 are decomposing probabilities to 4 different flows (Note that
p1 + p2 = 1, p3 + p4 = 1). We denote λ1 and λ2 as rates of arrival process to S1 and S2, respectively.
System X
λ p3 p1
System 1 (S1) System 2 (S2)
p4
p2

Q. 8 [L.O.3] Suppose the system is under steady condition without the loss of any request. Which of
the following relations is invalid ?
 
A λ = λ2 p1 B λ1 = λ + λ1 p4 + λ2 p2
 
C λ1 = λ D λ2 = λ1 p3

Q. 9 [L.O.3] Given some value to the system parameters as follows.

• p1 = 4/5
• p2 = 1/5
• p3 = 1/2
• p4 = 1/2
• λ =8 requests/second.

Which of the following is correct ?


A λ1 = 20 requests/second and λ2 = 10requests/second.

B λ1 = 30 requests/second and λ2 = 15 requests/second.

C λ1 = 15 requests/second and λ2 = 15 requests/second.

D λ1 = 20 requests/second and λ2 = 30 requests/second.

Q. 10 [L.O.3]. In a bank, customers arrive with the rate 30 customers per hour. Each customer needs
to wait, on average, 6 (min) to be served by a teller. We do not know the distribution of customer
arrival. How many customers, on average, should be standing in line?
   
A 1/3 B 3 C 5 D 1/5

In Questions 11–15, we consider a data transmitter in which packets arrive at an output link due to
a Poisson process, the mean total data rate is 80Kbps, the mean packet length is 12Kb, and the link
speed is 100Kbps.

Head of department: . . . . . . . Test code CLC.2010 Page 2

Q. 11 [L.O.3]. The transmitter can be modeled by an M/M/1 model whose customers, server, and
service time are respectively

A The transmitter, packets, and transmission times.

B Packets, the transmitter, and transmission times.

C The transmitter, packets, and the data rate.

D Cannot fit the transmiter as an M/M/1 model.

Q. 12 [L.O.3]. The mean total data rate (λ) and link speed (µ), which were given, are bit rates (Kb
per second). We now want to change them into packet rates (packets per second). They should be
respectively
 
A λ = 3/20 and µ = 3/25 B λ = 25/3 and µ = 20/3
 
C λ = 20/3 and µ = 25/3 D λ = 3/25 and µ = 3/20

Q. 13 [L.O.3]. The mean response time (including time for queuing and transmission) is
   
A 5/3 B 3/5 C 4/5 D 5/4

Q. 14 [L.O.3]. The mean transmission time (in second) is

   
A 3/20 B 1 − 3/25 C 1 − 3/20 D 3/25

Q. 15 [L.O.3]. The mean number of packets is

   
A 1/4 B 5 C 4 D 1/5

Q. 16 [L.O.3]. The average delay experienced by a packet when traversing a computer network is 100
ms. The average number of packets that cross the network per second is 128 packets/s. What is
the average number of concurrent packets in transit in the network at any time?
   
A 1.28 packets B 12800 packets C 1280 packets D 12.8 packets

Q. 17 [L.O.3] Consider a MUX with 3 line-ins, in which their arrival processes are assumed to be
Poisson with the rate of 17 packets/second, 13 packets/seconds and 6 packets/second, respectively.
Note that it is assumed that signal processing speed is converted to packets/second. A speed of
multiplexing is set up to the rate 45 packets/second under exponential distribution. If the MUX
is modeled as an M/M/1 queue, what is the average time of a packet in the MUX ?
 
A 0.11 seconds. B 0.8 seconds.
 
C 0.03 seconds D None of the others.

Q. 18 [L.O.3] Which of the following is not a basic component of a single queue ?

 
A Arrival process. B Servers with their service distribution.
 
C Waiting queue with its discipline. D Relationships between servers and arrival
process.
Q. 19 [L.O.2, L.O.5]. If you want to compare two music player softwares, which methodology would
you choose?
   
A Simulation modeling B Analytical modeling C Measurement D All of the others.

Q. 20 [L.O.3]. Consider the following scenario: the interarrival time is exponentially distributed with
an average of 10 minutes and the service time is so with an average of 8 minutes. Which queuing
model the system should be?
   
A M/M/2 B M/M/1 C M/D/1 D M/G/1

Head of department: . . . . . . . Test code CLC.2010 Page 3

 TEST 1 SOLUTIONS
HO CHI MINH CITY UNIVERSITY Course: System Performance
OF TECHNOLOGY
Evaluation (CO3007)
Faculty of Computer Science & Technology
Class: ... Group: ...
Time duration: 30 mins
(open test)
Date: Oct 27, 2020

Code: 2010
  
Q. 1 D Q. 8 C Q. 14 D
  
Q. 2 D Q. 9 A Q. 15 C

Q. 3 D
 
 Q. 10 B Q. 16 D
Q. 4 B 
 Q. 17 A
Q. 5 D  
Q. 11 B
 Q. 18 D
Q. 6 C  
 Q. 12 C Q. 19 C
Q. 7 B  
Q. 13 B Q. 20 B

Head of department: . . . . . . . Test code CLC.2010 Page 1

 TEST 1
HO CHI MINH CITY UNIVERSITY
OF TECHNOLOGY Course: System Performance
Faculty of Computer Science & Technology Evaluation (CO3007)
Class: 20201 Group: CC01
Time duration: 30 minutes
(open test)
Date: Oct 27, 2020

Student name: Student no.:

Score (number): Examiner:

Score (words): Invigilator:

(The test has 20 questions which must be answered by the best choice.)

In Questions 1–5, we consider a data transmitter in which packets arrive at an output link due to a
Poisson process, the mean total data rate is 80Kbps, the mean packet length is 12Kb, and the link speed
is 100Kbps.

Q. 1 [L.O.3]. The transmitter can be modeled by an M/M/1 model whose customers, server, and
service time are respectively

A Cannot fit the transmiter as an M/M/1 model.

B The transmitter, packets, and transmission times.

C Packets, the transmitter, and transmission times.

D The transmitter, packets, and the data rate.

Q. 2 [L.O.3]. The mean total data rate (λ) and link speed (µ), which were given, are bit rates (Kb
per second). We now want to change them into packet rates (packets per second). They should be
respectively
 
A λ = 3/25 and µ = 3/20 B λ = 3/20 and µ = 3/25
 
C λ = 25/3 and µ = 20/3 D λ = 20/3 and µ = 25/3

Q. 3 [L.O.3]. The mean response time (including time for queuing and transmission) is
   
A 5/4 B 5/3 C 3/5 D 4/5

Q. 4 [L.O.3]. The mean transmission time (in second) is

   
A 3/25 B 3/20 C 1 − 3/25 D 1 − 3/20

Q. 5 [L.O.3]. The mean number of packets is

   
A 1/5 B 1/4 C 5 D 4

Q. 6 [L.O.3] Consider 2 following single queues: (Q1) M/M/3/100/100 and (Q2) M/M/3/300/100.
Which of the following is correct w.r.t. performance comparison?
 
A Q1 is equivalent to Q2. B Q1 is better than Q2.
 
C Q2 is better than Q1. D N/A. It depends on arrival rate.

Two following questions use the same problem description as follows.

Head of department: . . . . . . . Test code CLC.2011 Page 1

 Consider a queueing system X as shown in the following picture. X receives the incoming traffic as a
Poisson process with rate λ. p1 , p2 , p3 , p4 are decomposing probabilities to 4 different flows (Note that
p1 + p2 = 1, p3 + p4 = 1). We denote λ1 and λ2 as rates of arrival process to S1 and S2, respectively.
System X
λ p3 p1
System 1 (S1) System 2 (S2)
p4
p2

Q. 7 [L.O.3] Suppose the system is under steady condition without the loss of any request. Which of
the following relations is invalid ?
 
A λ2 = λ1 p3 B λ = λ2 p1
 
C λ1 = λ + λ1 p4 + λ2 p2 D λ1 = λ

Q. 8 [L.O.3] Given some value to the system parameters as follows.

• p1 = 4/5
• p2 = 1/5
• p3 = 1/2
• p4 = 1/2
• λ =8 requests/second.

Which of the following is correct ?


A λ1 = 20 requests/second and λ2 = 30 requests/second.

B λ1 = 20 requests/second and λ2 = 10requests/second.

C λ1 = 30 requests/second and λ2 = 15 requests/second.

D λ1 = 15 requests/second and λ2 = 15 requests/second.

Q. 9 [L.O.3]. Consider the following scenario: the interarrival time is exponentially distributed with
an average of 10 minutes and the service time is so with an average of 8 minutes. Which queuing
model the system should be?
   
A M/G/1 B M/M/2 C M/M/1 D M/D/1

Q. 10 [L.O.3] Consider a single queue with an infinite queueing length. The arrival process is bursty
(i.e., traffic with a sudden increase). Which of the following is correct to departing traffic of the
queue in steady condition ?
 
A The property of being bursty is decreased B Departing traffic is still bursty.
in departing traffic. 
C Departing traffic is changed to real-time D De[arting traffic (rate) is less than arrival
traffic. traffic ].
Q. 11 [L.O.1]. Why is the step of workload selection conducted after the step of selecting evaluation
technique ?

A Workload selection is performed only if performance metrics, factors, and evaluation
techniques are known.
B Not much important. Workload selection can be conducted before evaluation tech-
nique selection.
C Workload depends strongly on evaluation techniques.

D Different workload selection are for different evaluation techniques.

Head of department: . . . . . . . Test code CLC.2011 Page 2

Q. 12 [L.O.3] Which of the following is correct with Little’s Law ?

A To describe the relationship between arrival rate and departing rate.

B Applicable for any closed system.

C Applicable to any closed system without request/job lost inside the system.

D Applicable to any system working in transient state.

Q. 13 [L.O.2]. We consider a local area network (LAN) providing a video conferencing service on virtual
circuits for end users (rectangle boxes). The LAN bounded within the rectangle is the system under
test (SUT).
u5
d The virtual circuits are given as follows.
u1 u6
u2 a c u7 Virtual circuit Path Allocated bandwidth
VC1 u1 − a − c − d − u5 10Mbps
u3 u8 VC2 u3 − b − c − d − u 5 3Mbps
b
VC3 u7 − d − c − u8 8Mbps
u4
Which of the following is Jain’s fairness index of the bandwidths allocated to VCs by SUT ?
   
A 0.849 B 0.240 C 0.333 D 2.549

Q. 14 [L.O.3] Which of the following is not a basic component of a single queue ?

 
A Relationships between servers and arrival B Arrival process.
process. 
C Servers with their service distribution. D Waiting queue with its discipline.

Q. 15 [L.O.3]. The average delay experienced by a packet when traversing a computer network is 100
ms. The average number of packets that cross the network per second is 128 packets/s. What is
the average number of concurrent packets in transit in the network at any time?
   
A 12.8 packets B 1.28 packets C 12800 packets D 1280 packets

Q. 16 [L.O.2, L.O.5]. If you want to compare two music player softwares, which methodology would
you choose?
   
A All of the others. B Simulation modeling C Analytical modeling D Measurement

Q. 17 [L.O.3] During an hour of measurement, the number of requests to a domain name server is 10800.
The average response time of the server is 1/3 second. What is the average number of requests
being processed within the server?
   
A 3 B 3600 C 1 D 32400

Q. 18 [L.O.3] Which of the following systems is impossible (or inappropriate) to be modeled as a queue-
ing system ?

A A distribution network from manufacturers to their distributors.

B Operational processes of a software development company.

C A mechanism to manage processes in OS kernel.

D Flying activities of an UAV.

Q. 19 [L.O.3] Consider a MUX with 3 line-ins, in which their arrival processes are assumed to be
Poisson with the rate of 17 packets/second, 13 packets/seconds and 6 packets/second, respectively.
Note that it is assumed that signal processing speed is converted to packets/second. A speed of
multiplexing is set up to the rate 45 packets/second under exponential distribution. If the MUX
is modeled as an M/M/1 queue, what is the average time of a packet in the MUX ?
 
A None of the others. B 0.11 seconds.
 
C 0.8 seconds. D 0.03 seconds

Q. 20 [L.O.3]. In a bank, customers arrive with the rate 30 customers per hour. Each customer needs
to wait, on average, 6 (min) to be served by a teller. We do not know the distribution of customer
arrival. How many customers, on average, should be standing in line?
   
A 1/5 B 1/3 C 3 D 5

Head of department: . . . . . . . Test code CLC.2011 Page 3

 TEST 1 SOLUTIONS
HO CHI MINH CITY UNIVERSITY Course: System Performance
OF TECHNOLOGY
Evaluation (CO3007)
Faculty of Computer Science & Technology
Class: ... Group: ...
Time duration: 30 mins
(open test)
Date: Oct 27, 2020

Code: 2011
  
Q. 1 C Q. 7 D Q. 14 A
  
Q. 2 D Q. 8 B Q. 15 A
 
Q. 3 C  Q. 16 D
Q. 9 C
 

Q. 4 A Q. 17 C
Q. 10 A
  
Q. 5 D Q. 11 A Q. 18 D
 
 Q. 12 C Q. 19 B
Q. 6 A  
Q. 13 A Q. 20 C

Head of department: . . . . . . . Test code CLC.2011 Page 1