,

L-3ff -2/CSE Date: 02/10/2023

BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
L-3/T-2 B. Sc. Engineering Examinations 2021-2022

Sub: CSE 313 (Operating System)

Full Marks: 210 Time: 3 Hours
The figures in the margin indicate full marks.
USE SEPARATE SCRIPTS FOR EACH SECTION
-----------------------------------
SECTION -A
There are FOUR questions in this section. Answer any THREE.

1. (a) Consider the following workload: (18)

Process
Priority (Lowest Number Duration (sec) Arri val Time (sec)
has Highest Priority)
1'1 1 60 0
1'2 1 25 30
1'3 2 80 50
1'4 3 20 70
Draw the Gantt chart and calculate the average turnaround time for each of the following
scheduling algorithms:
• Shortest Remaining Time Next

• Priority Scheduling. Within the same priority class, schedule according to Round
Robin with quantum 20 sec. Each priority class maintains separate FIFO queue.
(b) Four jobs have arrived at the same time in a batch system. Their expected run times
are 9, 3, 5, and X. In what order should they be run to obtain optimal average turnaround

time? (i-lint: The jobs are non-preemptive) (10)

(c) Differentiate between an unsafe state and a deadlock state. (7)

2. (a) Peterson's solution for achieving mutual exclusion for USIng critical regIons IS

presented in Figure for Question 2(a): (12)

(i) Discuss priority inversion problem with a high-priority process, H, and a low-
priority process, L.
(ii) Does the same problem occur if round-robin scheduling is used instead of priority
scheduling? Justify.

~i~fJi1€fALSE 0
~oew'eTRUE 1
tdflofili€ ~J 2 /. nWllher oi pro.:PS$(:s 0/

lot turn, /- ",'hose turn IS ll"! -/

n'! interesledfNl: j- all valu8:s inl!lally 0 IrALSE. oj

'/Oi(j ante' _'99'00(1111 proc;'3'£sj; /" pwce~s IS 0 or 1 .f

!
In! other; /" number of the 0111::(process -/

oltier •••1 - process; ,- the OppoSi!fl 01 process "'

1I11eresled[process] ,., TAUE: /. show thai you arc Interested ,,/
twn •• process: /" selll.1g ,,/
..•••.
hlle (turn =- process 8& inleresled[other! •••• TRUE) j" null sialement ., ;

~'Old Iec'h'o_regi(ln(inl processJ j. process' who 15 leaving .1

InlerpSledlprocessJ • FALSE: ,- Indicate departure from crrtlCal regIOn _/

Contd 1'/2
Figure for QlIeslion 2(a): Pelcr. ...oll.S :-o!ulioll for 2 proccs\C's
 =2=

CSE 313
Contd. Q. No. 2(a)(ii)
(b) In the dining philosophers problem, let the following protocol be used: An even-
numbered philosopher always picks up his left fork before picking up his right fork; an
odd-numbered philosopher always picks up his right fork before picking up his left fork.

Investigate whether this modified protocol prevents deadlock. (10)

(c) State the four requirements that must be present in a solution to support mutual

exclusion among processes sharing resources. (6)

(d) Differentiate between micro kernel and monolithic kernel architectures. (7)

3. (a) Consider a system with 4 processes: PI through P4 and 3 resources types: A (9 units),
B (3 units), C (6 Units). Current allocation of resources and maximum requirement of a
process for each resource are given in Figure for Question 3(a). Determine whether the

current state of this system is safe or not. (15)

A B C
PI I 0 0
P2 6 I 2
P3 2 I I
P4 0 0 2
Current Allocation Matrix

A B C
PI 3 2 2
P2 6 I 3
P3 3 I 4
P4 4 2 2
Maximum Requirement Matrix
Figure for Question 3(a)

(b) Suppose that there is a resource deadlock in a system. Give an example scenario to
show that the set of processes deadlocked can include some processes that are not in the

circular chain in the corresponding resource allocation graph. (10)

(c) Explain how to attack the following conditions to structurally prevent deadlock

(i) Hold and wait condition (ii) Circular wait condition. (10)

4. (a) Write down the steps in making a system call. (7)

(b) Differentiate between process context switching and thread context switching Identify
which of the following are "per process item" and which are "per thread item"?
Program Counter, Stack, Address Space, Global Variables, Registers, Directories, Child

Process, Local Variables (13)

Contd P/3
 =3=
CSE 313
Contd ... Q. NO.4
(c) Illustrate the position of scheduler, process table and thread table in user-level thread
implementation and kernel-level thread implementation with diagrams. With the help of
your diagrams, justify that kernel-level thread implementation is better than user-level

thread implementation with respect to blocking. (15)

SECTION -8
There are FOUR questions in this section. Answer any THREE.

5. (a) The following sequence of Virtual Page Numbers (VPN) has been referenced in your

system. (7+7+14=28)
1,2,3,4,5,2,3,1,2,3,4,5, I
Now, answer the following questions.
(i) Explain Belady's anomaly. Show that this anomaly occurs for cache size 3 and 4
in the above case.
(ii) Calculate hit and miss rate for the optimal page replacement algorithm.
(iii) If the TLB can keep at most 4 entries and employs LRU for TLB replacement
and Page replacement uses CLOCK algorithm with cache size 4. Calculate the
memory access time for the above memory access sequence. TLB access takes 5ns,
memory access takes 60ns, and disk access takes 4ms.
Hint: The first 5 memory references in CLOCK algorithm gives the following cache
states:
Access
I 7(0) 7(0)
2 7(0) 7(0)
3 3 1)* 7(0)
4 3(1) 4(1 *
5 3 0) 4(0)

Here,
"7" refers to unused cache,
the number in bracket refers to use bit and
"*,, is the location of the clock hand.
(b) Both ffs and Ifs try to optimize on vsfs m some ways. State and elaborate the

difference in their optimization criteria. (7)

6. (a) The following function writes a sequence ofn numbers to a given file descriptor fd. (5x3+13=28)
void write fd(int fd, int n){
char buf[10l;
for (int i = 0; i < ni i++) {

itoa(i, buf, 10);

write (fd, buf, strlen(buf));
}
}
Contd P/4
 =4=
CSE 313
Contd ... Q. No. 6(a)(i)
(i) For each of the following possible values of n, design a device driver where
wri te _ fd is called frequently with n. Explain your design decisions using
illustrative pseudo codes.
(a) always 10
(b) always 1000000
(c) mixture of 10 and 1000000

(ii) What is the difference of output from the following two code blocks? Analyze
with necessary ligures and arguments showing what happens in the kernel.
Code block 1 Code block 2
int fdl = open("file.txtn) ; int fdl = open("file.txt") ;
int fd2 = open("file.txt") ; int fd2 = dup(fdl) ;
write
-fd(fdl); write
-fd(fdl) ;
write fd (fd2) ; write fd(fd2) ;
- -
(b) Memory allocation API consists mainly of the following two functions: (7)
void *malloc(size_t size);

void free (void *ptr);

With illustrative examples, show how the allocation API works. Mention the necessary
bookkeeping structure(s) it requires.

7. (a) You have executed the following commands in the root of an ffs file system with
4 byte block number and 4 KB block size. The disk has an average disk-arm positioning

time of 10 ms and max transfer rate of 100 MB/s. (4x3+ 16=28)

mkdir p

mv /q/foo.txt /p/bar.txt
I
Here, foo. txt is of size 1GB.
Now answer the following questions.
(i) Given, bitmaps are in block 4, inodes are in block 5 and data for the root directory
is in block 6. Also, the journal is kept in blocks 26 to 31. Determine the journaling
timeline if the system uses:
(a) Data Journaling

(b) Metadata Journaling

(c) Metadata Journaling with Checksum Optimization

(ii) How will this file be laid out in ffs? How does ffs get the relevant information
needed to decide that? Calculate how long it will take to sequentially read the whole
file. You can assume, there is enough space /0 save the file in ffs layout and only
movingfrom one block group /0 ana/her requires a disk-arm positioning.

Contd P/5
 =5=

CSE 313
Contd ... Q. No.7
(b) What is dynamic relocation and segmentation in the context of memory management?
Why would you choose one over another? Using pseudocodes, elaborate the address

translation process used by these two systems. (7)

8. (a) You have set up a RAID system with 10 disks and 4KB block size. You are using Ifs

as the default file system where segment size is 64 MB. The disk config is as follows:(3+9+6+5x2=28)
Capacity ITB
Rotation speed 10,000 RPM
Max seek time 12 ms
Max transfer rate 100 MB/s
Platters 2
Sector size I KB
Cache 16 MB
The following command has been executed in your file system.
rm Ifsfile.txt

Now, answer the following questions.

(i) How is the inode number problem solved in Ifs?
(ii) Write down the steps that will be executed while running the command. How will
Ifs invalidate any subsequent access to Ifsfile. txt?
(iii) Elaborate the recovery action that Ifs will take, if a crash occurs during the
operation.
(iv) Calculate the throughput for read and writes in your file system (lfs) for the
following RAID setups.
(a) RAID-O

(b) RAID-I

Hint:
• Ifs reads 1 block at a time and writes I segment at a time.
• You can read/write one sector at a time on a disk.
(b) What is thrashing in the context of memory management? Why is it a problem? List

some solutions for it. (7)

 "
L-3/T-2/CSE Date: 18/10/2023
BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
L-3/T-2 B. Sc. Engineering Examinations 2021-2022

Sub: CSE 317 (Artificial Intelligence)

Full Marks: 210 Time: 3 Hours
The figures in the margin indicate full marks.
USE SEPARATE SCRIPTS FOR EACH SECTION
--------------
SECTlON-A
There are FOUR questions in this section. Answer any THREE.

I. (a) "Surely computers cannot be intelligent - they can do only what their programmers
tell them." Is the latter statement true, and does it imply the former? Describe it in detail
using the concepts of evolutionary algorithm. (20)
(b) For each of the following activities, give a PEAS description of the task environment (15)
(i) Bidding on an item at an auction
(ii) Medical diagnosis system
(iii) Interactive English tutor

2. (a) Explain why problem formulation must follow goal formulation. (10)
(b) The traveling salesman problem (TSP) is described as follows: (25)
"Given a list of cities and the distances between each pair of cities, what is the shortest
possible route that visits each city exactly one time and returns to the origin city?"
Describe in detail how we can solve TSP by applying genetic algorithm with appropriate
representation scheme and operators.

3. (a) Describe simulated annealing algorithm with the help of its pseudo code. Explain how
this algorithm can avoid local optima problem of a greedy search algorithm. (20)
(b) Show that A * search algorithm is optimal if the heuristic used in it is admissible
and/or consistent. (15)

4. (a) Describe the main challenges of adversarial search as contrasted with single-agent
search. Explain why it is a good heuristic to choose the variable that is the most
constrained but the value that is the least constraining in a CSP search. (15)
(b) Show that the worst case time complexity the arc-consistency algorithm AC-3 is in
cubic order. (20)

SECTlON-B
There are FOUR questions in this section. Answer any THREE.

5. (a) We have a bag of three biased coins a, b, and c with probabilities of coming up heads
of 20%, 60%, and 80%, respectively. One coin is drawn randomly from the bag (with
equal chance of drawing each of the three coins), and then the coin is flipped three times
to generate the outcomes X" X2, and X3. (6+12=18)
Contd P/2
 =2=

CSE 317
Contd ... Q. No. 5(a)

(i) Construct a Bayesian network corresponding to this setup and state the necessary
conditional probability tables (CPTs).
(ii) Compute which coin was most likely to have been drawn from the bag if the
observed flips come out as Xl = head, X2 = head, and X3 = tail.

(b) Let Hx be a random variable denoting the handedness of an individual x, with possible
values 1 or r. A common hypothesis is that left- or right-handedness is inherited by a
simple mechanism; that is, perhaps there is a gene Gx, also with values 1 or r, and perhaps
actual handedness turns out mostly the same (with some probability s) as the gene an
individual possesses. Furthermore, perhaps the gene itself is equally likely to be inherited
from either of an individual's parents, with a small nonzero probability m of a random

mutation flipping the handedness. (5+6+6=17)

(i) Show which of the three networks shown III Figure 5 satisfy the following
equation:

peG jather,Gmolher,Gchild) = peG jalher)P(Gmolher)P(Gchild)

(ii) Compute the conditional probability table (CPT) for the Gchild node in network (a)
(of Figure 5), in terms of sand m.
(iii) Suppose the P(Gjalher = I) = P(Gmolher = l) = q. in the given network (a) (of Figure
5), compute an expression for P(Gchiid = I) in terms of m and q only, by conditioning
on its parent nodes.

(al (b) lei

Figure 5

6. (a) Discuss why overfitting is bad in machine learning. Briefly explain How you would
prevent overfitting in decision tree learning. (7)

(b) Why is a complex hypotheses space not always preferred in machine learning? Justify

with respect to Occam's razor principle. (5)

Contd P/3
 =3=
CSE 317
Coutd ... Q. No.6

(c) A professor wants to know if students are getting enough sleep. Each day, the
professor observes whether the students in the class have red eyes. The professor has the
following domain theory: The prior probability of getting enough sleep, with no
observations, is 0.7. The probability of getting enough sleep on night t is 0.8 given that
the student got enough sleep the previous night, and 0.3 if not. The probability of having
red eyes is 0.2 if the student got enough sleep, and 0.7 if not. Answer the following
questions. (8+15=23)
(i) Build a hidden Markov model (HMM) to formulate the above problem that the
professor could use to filter or predict from a sequence of observations. Show the
state transition probabilities (transition model) and observation probabilities (sensor
model).
(ii) Let the state variable be X, = EnoughSleep,. Given the evidence values el = not
red eyes, e2 = red eyes, and e3 = red eyes for the first three days, compute the filtered
estimates P(X,le/:J for each of 1= 1,2,3 using filtering algorithm. Using the filtered
estimate, compute smoothed probability estimates P(x'lel:3) for each of I = 1,2,3 using
forward-backward algorithm.

7. (a) Sometimes Markov decision processes (MOPs) are formulated with a reward function
R(s, a) that depends on the action a taken on state s or with a reward function R(s, a, s)
that also depends on the outcome state s '. Express the Bellman equations for these
formulations. (7)

(b) Figure 7 shows an instance of a grid world Markov decision process (MOP). Shaded
cells represent walls. In all states, the agent has available actions UP (t), Down (,j,), Left
(f-), and Right (~). Performing an action that would transition to an invalid state
(outside the grid or into a wall) results in the agent remaining in its original state. In states
with an arrow coming out, the agent has an additional action EXIT. In the event that the
EXIT action is taken, the agent receives the labeled reward and ends the game in the
terminal state T. Unless otherwise stated, all other states generate no reward, and all
transitions are deterministic (not stochastic). Let the discount factor be y = V2. (16)
Suppose that we are performing value iteration algorithm on the given grid world MOP.
Assume that value iteration begins with all states initialized to zero, i. e., U(s) = 0 \t.v.
Compute the optimal utility values U(s) for states (i.e., cells in the grid) A and B. Show
the utility values U(s) for all states after every iteration of the value iteration algorithm.
Show the optimal policy.

EXIT
r = •. 1

EXIT
r = •. 100

Figure 7

Contd P/4
 =4=

CSE 317
Contd ... Q. No.7

(c) Consider the following random process. A magician has three coins, each of which
has a distinct type. One is a fair coin (SO/50 odds of heads vs tails). The other two are
tricky coins: one has heads on both sides and the other has tails on both sides. At every
time step, the magician picks a coin randomly with the exception that a fair coin is not
picked at two successive time steps (in which case remaining coins are equally likely to
be chosen). After picking a coin (without actually showing you which coin was picked),
the magician flips (toss) it, and shows you the result, However, unfortunately, the
magician only shows you the coin very briefly, and 10% of the time you make a mistake
when you observe the true side of the coin (e.g., you see heads when it was actually tails).
Construct a hidden Markov model (HMM) to formulate the problem. Show the state
transition probabilities (transition model) and observation probabilities (sensor model). (12)

8. (a) Consider the following data set comprised of three binary input attributes (A I,AZ, and
A3) and one binary output y: (16)
Example Al Az A3 Outputy
XI 1 0 0 0
Xz I 0 1 0
X3 0 I 0 0
X4 1 I I I
Xs I I 0 1

Construct a decision tree for these data. Show the entropy and information-gain
computations made to determine the attribute to split at each node.
(b) Discuss how would you handle the following cases in decision tree learning? (i) No
examples left at a child node, (ii) No attributes are left, but both positive and negative
examples remain at a child node, and (iii) Missing values in an attribute. (6)
(c) Consider the Bayes net shown in Figure 8 with Boolean variables B =

BrokeElectionLaw, I = Indicted, M = PoliticallyMotivatedProsecutor, G = FoundGuilty,

J = Jailed. Using variable elimination algorithm, compute the probability that someone
goes to jail given that they broke the law, have been indicted, and face a politically
motivated prosecutor. (13)

BM f'(/)
1 1 .9
1 I .5

I/~B)
.9
I I 1
I I
.5
.1 I {'(,In,
.1

B I .II P(G)
1 1 1 .9
1
1
1 I
I 1
I I
.8
.0
.
lWJ
1 .0 IV)
I 1 1 .2 1 .9
f 1 f .1 f .0
f f 1 .0
f f f .0

Figure 8
L-3/T-2/CSE Date: 12/10/2023
BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
L-3/T-2 B. Sc. Engineering Examinations 2021-2022
Sub: CSE 321 (Computer Networks)
Full Marks: 2 I0 Time: 3 Hours
The figures in the margin indicate full marks.
USE SEPARATE SCRIPTS FOR EACH SECTION
_._ _-_. __ ...•
.•. --_ ..__ .•.•.•._-----_._--_._. __ .•.•.•.__ .•._---_._-------,----------------
SECTION -A
There are FOUR questions in this section_ Answer any THREE_

I. (a) Design a network with necessary agents to SUPPOl1 mobile hosts. Describe how we
can apply the network over different types of networks covering LANs, WANs, etc. (23)
(b) Classful IP addressing needs to be done for a company with different distant brances
each having a number of computers. Analyze whether the notion of subnetting will be
required here. Explain with necessary example(s) (12)

2. (a) In datagram subnets, DVR is applied with the notions of Forced Update and Split
Horizon Rule. You need to design an example network, where such applications cannot
remove the inefficiency of count-to-infinity problem_ Analyze your example with
necessary diagram(s). (23)
(b) Describe the types of networks where we need BGP to operate. Discuss peenng
options over such networks with necessary diagram(s). (12)

3. (a) Construct a network where we can experience congestion collapse. Analyze how
AIMD can lessen the extent of such congestion, with necessary figure(s). (23)
(b) Explain how a remote application can be called from a networked machine at
distance. Discuss the process with necessary figure(s). (12)

4. (a) In the networking word, there exist different types of computing machines. Some of
the machines are fast, whereas some other machines are slow. Thus, in the case of
networked data transmission, we can have slow receivers as well as slow senders.
Now, li'OIn the perspective of the Transport layer, you need to distinguish between
solutions lor the slow receivers as well as slow senders. Explain your answer with
necessary figure(s). (23)
(b) Develop a methodology for RTO estimation in a network, where packet drops are
common. Explain your methodology with necessary figure(s). (12)

Contd P/2
 =2=
CSE 321
SECTION -8
There are FOUR questions in this section. Answer any THREE.

5. (a) Suppose four active nodes -- nodes A, B, C, and D-- are competing for access to a
channel using slotted ALOHA. Assume each node has an infinite number of packets to
send. Each node attempts to transmit in each slot with probability p. The first slot IS

numbered slot I, the second slot is numbered slot 2. and so on. (8+7=15)
(i) Find the probability that the first success occurs in slot 4.
(ii) Calculate the efficiency of this four-node system.
(b) Draw the Ethernet frame format. You need to write the lengths of each field. (Hin/l:
There are seven fields. Hin/ 2: The paylaadfield is of variable leng/h ranging fi-OI1l -16-
1500 by/es). Illustrate the reason(s) behind keeping the last two bits of Preamble field
different than the rest. (10)
(c) Suppose. you (Hostname: ckrue/.edu) are querying the IP-address for cse.buet,edu.
However, BUET maintains an authoritative DNS server dns.bue/.edu for the hostnames
ending with bue/.edu. List all the DNS record(s) stored in 1).edu TLD DNS server and 2)
dns.bue/.edu DNS server that facilitate(s) your query. You need to mention the record(s)
in this format (Name, Value, Type). Feel free to use any IP address(es) of your choice, if

need arises. (10)

6. (a) Consider three LANs interconnected by two routers, as shown 111 Figure below.
Answer the following questions. (5+5+5+5=20)

Subnet 2

-
Figure for Question 6(a)
~._---..
...

(i) Assign IP addresses to all of the interfaces. For Sub net I use addresses of the form
172.J6.J.XXX; for Subnet 2 use addresses of the form 172.J6.2.XXX; and for Subnet
3 use addresses of the form J 72. 16.3.XXX
Contd P/3
..
=3=
CSE 321
Contd ... Q.No.6(a)
(ii) Assign MAC addresses to all of the adapters.
(iii) Consider sending an IP datagram from Host E to Host B. Suppose all of the ARP
tables are up to date. Enumerate all the steps.
(iv) Repeat (iii), now assuming that the ARP table in the sending host is empty (and
the other tables are up to date)
(b) We need to send data at a 5-Mbps rate. There are two encoding options: (8)
(i) A combination of 4B/5B and NRZ-I, or
(ii) Manchester coding
Argue which option should be followed to minimize the required bandwidth.
(c) Why switches are considered self-learning? Briefly explain. (7)

7. (a) Very briefly answer the following questions: (10)

What is the main improvement CSMA brought over ALOHA variants (Plain ALOHA,
Slotted ALOHA)? Could CSMA solve the problem entirely? What is the need for
CSMA/CD?
(b) Assume the generator G (= 1001) IS used III CRe. Now, answer the following
questions. (5+5=10)
(i) Illustrate how this generator can detect any single bit error in data D.
(ii) Determine whether the above G detects any odd number of bit errors. State the
reason(s) clearly.
Origin servers

15 Mbps access link

.!
Institutional network

Figure for Question 7(c)

(c) Consider the figure above, for which there is an institutional network connected to the
Internet. Suppose that the average object size is 850,000 bits and that the average request
rate from the institution's browsers to the origin servers is 20 requests per second. Also
suppose that the amount of time it takes from when the router on the Internet side of the
access link forwards an HTTP request until it receives the response is three seconds on
average.
Contd P/4
..

=4=

CSE 321
Contd ... Q.No.7(c)
Model the total average response time as the sum of the average access delay (that is, the
delay from Internet router to institution router) and the average Internet delay. For the
x
average access delay, use 1- xy where x is the average time required to send an object
over the access link and y is the arrival rate of objects to the access link. (8+7=15)
(i) Determine the total average response time.
(ii) Now suppose a cache is installed in the institutional LAN. Suppose the miss rate
is 0.4. Find the total average response time in this case.

8. (a) Suppose you walk into a room, connect to Ethernet, and want to download a Web
page. Identify all the protocol steps that take place, starting from powering on your PC to
getting the Web page. Assume there is nothing in our DNS or browser caches when you
power on your Pc. Explicitly indicate in your steps how you obtain the IP and MAC
addresses of a gateway router. (20)
(b) Consider sending over HTTP/2 a Web page that consists of one video clip, and five
images. Suppose that the video clip is transported as 2000 frames, and each image has
three frames. (5+5=10)
(i) If all the video frames are sent first withollt illter/eavillg, calculate the number of
frames to be transp0l1ed until all five images are sent.
(ii) If frames are illter/eave£!, calculate the number of frame(s) needed until all five
images are sent.
(c) How self-synchronization property helps the data transmission? Explain briefly. (5)
L-3/T-2/CSE Date: 07/1012023
BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
L-3/T-2 B. Sc. Engineering Examinations 2021-2022

Sub: CSE 325 (Information System Design)

Full Marks: 210 Time: 3 Hours
The figures in the margin indicate full marks.
USE SEPARATE SCRIPTS FOR EACH SECTION
---------_._----
SECTION-A
There are FOUR questions in this section. Answer any THREE.

I. (a) A company is very much reputed in completing their IT projects in time with desired
quality which meets functional and non-functional requirements correctly. The have pool
of experienced and qualitied programmers who are key to this success. If a project does
not get that much progress then those qualified programmers are assigned to solve the
problem. The programmers are low paid and generally without experience. The software
development company were awarded so many S/W development projects in the past.
Recently the company did not win a bid where there was a requirement to have a process
to deliver quality software. Assess the company in terms of CMMI level. What are the
necessary steps to be done to upgrade the company in the next CMMI level? (10)
(b) Compare different versioning models with examples focusing the following: (10)
(i) Locking
(ii) Conflict resolution
(iii) Practical use
(iv) Merging
(c) How are Push and Pull operations performed in Distributed Version Control?
Consider a scenario in a Distributed Version Control system when three users work
locally on a file A at the same time. Show the steps of committing their changes. (15)

2. (a) Explain the following rules with suitable examples in designing user interfaces. (12)
(i) Offer informative feedback
(ii) Design dialogue to yield closure
(iii) Permit easy reversal of action
(iv) Reduce short term memory load to the user
(b) What are the basic principles to make the effort estimation of software projects
accurate? (5)
(c) If is difficult to estimate the effort when you start a project. In the subsequent steps of
S/W development, the scope and complexity of the project becomes clear and we get
chance to do better estimation. Explain how COCOMO II model addresses this during
estimation. (6)
(d) Titas Gas Transmission and Distribution Company is trying to have a new billing
software. There are several alternatives to implement that. The first one is to acquire a
ready made billing software from a software development company with the cost of Tk
I crore.
Contd P/2
"

=2=

CSE 325
Contd 00' Q. No. 2(d)

It might require 20 additional changes/reports to customize the software for the use of the
company. The cost of each easy, moderately difficult or extremely difficult change/report
is Tk 50K, Tk 75K or Tk lOOK, respectively. The probability of the changes/reports are
easy and moderately difficult are 0.30 or 0.25, respectively. The second option is to reuse
an open source billing software by an in-house team of 10 members. There are actually
two steps in using such a software. In the first step, the in-house team must understand
the open source code and install it for proper understanding. The average salary of the in-
house team members is Tk 75K. It will take 6 months to understand the open source
system if it is straight forward with the probability of 0.60. Otherwise it will take 9
months if the open source system is complex. In the next step it may require another 10,
20 or 40 additional screens or reports with the probability of 0.25, 0.50 and 0.25,
respectively to fulfill the need of the client. The cost of developing one screen/report is 1
man-month by the recruited employees. The third option is to purchase a generalized
billing software and customize it for the use of the client. The cost of the generalized
billing software is Tk 50 lacs. But the cost of consultants for necessary configuration is
Tk 20 lacs. It might require 30 additional changes/reports to customize the software for
the use of the company. The cost of each easy or extremely difficult change/report is Tk
lOOK or Tk 150K. The probability of the changes/reports are easy is 0.30. Which one
would be best strategy for developing the billing software? (12)

3. (a) Consider an Automated Teller Machine (ATM) of a Bank. You can have the
following banking operations using the ATM card: (8)
(i) Withdrawing money
(ii) Checking Balance from Core Banking System
(iii) Changing PIN (Personal Identification Number)
(iv) Depositing money
(v) Transferring money to another account
(vi) Recharging your mobile phone
Draw a use case diagram showing the actors and use cases.
(b) Write down the use case for withdrawing money and changing PIN. (8)
(c) Draw collaboration diagram for the use case for withdrawing money using ATM
Card. Show the steps of deriving class diagram from the collaboration diagram. (19)

4. (a) As project manager how do you form your team for the following software projects: (9)
(i) A project for solving complex scientific problem
(ii) Development of ERP with many modules
(iii) A software which must be ready at the beginning of the next year.
(b) What is Software Failure? What are the components failure cost? How can you
prevent the failure in software development? (8)
Contd P/3
 =3=
CSE 325
Contd ... Q. No.4

(c) Explain the term measure and metric with examples in the discipline of software
measurement. Explain two measurement principles with necessary examples. (7)
(d) Your goal is to evaluate the performance of the programmers. Define the metrics to
achieve this goal using goal oriented software measurement. (11)

SECTION-B
There are FOUR questions in this section. Answer Question 5 (compulsory) and any other
two (2) questions.

5. (a) In Bangladesh, every school teacher is appointed based on MPO regulation.

According to this regulation, each candidate has to meet certain qualification criteria
defined for each post (of a particular subject) of a certain salary grade. The qualifications
are stated in terms of age, subject of graduation, and result at different levels. Example of
such requirement is shown below. This is for a particular type of institution, e.g., school,
college, etc. (13)
Table for Q 5(a)

~)~ fip••f~'il:i'jlJ m '{~fiJt (~) ~-;~W

6 R..t~
9'* oa m C"""""'< "

~o
I
I
~
(~Rvr-l)
,.,.,...,
r(1{l;''''~
~!ll """"'
ili~? ~
I
~

",...,
~~
<lnrl1rol1 f~~/)
..-:IT
I (:>>!Jooo,':'7\!:Bo/. )

(::l:) ~~.;);)
I ~) ~f'l mn"llT"" ~ 'ItfiJt I
I ! f"",,"1~=~.
~fil f'01'l1'l.
I
(:>~(]'OQ-..:JO~.~U/.)

I ~~!1
fii,"',"Hlt=l o~(.s~) h ~ ",11
'. I . fht'Vl:!:f",,,,,,,,", ~'It'It'l1
~;;)I
t--- i ~) . f'l . J11l1l ~
9'* ~am (""""'<
! I f~ ~ 1Ir•••• ~ I "'""" 5 ~'l'Iftm v:;r
i ~ f3fil1 """'" I
I I 'I~~ "lA1
~ f"lf-4'lt'l1'11)

I '8 i (""'") ~)~,.. R1f",<11'IlI «'ll 'ff~ (~) C!lO.~~

I fuoR ~
,I
' ""~ ~ f'01'l1'll ~aoo""~oCOl.

c I I 1~
J
fWl~i","t~ o~(,s1l] ~
R~'tf:r6if.nf'
••:r,;; ~Cftlr."lJ
«(fir-ow
m :;"II
i

When a candidate applies for a post, hislher qualifications are matched against such
requirements, If the qualification criteria satisfy, he/she is eligible to sit for a recruitment
test. Design a BPMN diagram to design the scenario discussed above.
(b) Draw an ERD to design database for storing the qualification criteria depicted in Q
5(a) for each type of institution. (12)
(c) Draw a Mock dashboard (considering scenano of Q 5(a)) to show the list of
candidates who are found not eligible to appear at the recruitment test for not satisfying
one or more criteria. You have to clearly demonstrate the criteria not matched so that the
viewer can easily understand the reason for disqualify. (10)
Contd P/4
"

=4=
CSE 325

6. (a) To purchase books, the sites like Amazon.com or Rokomari.com are widely used.
Draw a software architecture to implement such a system following Microservice
architecture. (15)
(b) The web portal of stock market experiences frequent write operation and also updated
data is expected by viewers. Which cache write strategy is most appropriate for such
application and why? (11)
(c) Give three example scenanos where Data-Centered software architecture IS

appropriate. (9)

7. (a) There is a common printing service for printing secured official documents In an
office which is used by different departments. Sometimes, multiple departments send lots
of printing request simultaneously and want to get assured of their request is being
addressed. What technology should be recommended by the designer of such a system?
Draw a block diagram for the required design. (9)
(b) What do you understand by Second Order SQLinjection? Give an example. What is
the primary (most recommended) way of preventing SQLinjection? (12)
(c) You have developed an application for different clients. You need to deploy that
quickly in different types of servers of the clients. What DevOps technology will serve
your purpose? (8)
(d) In your hardware infrastructure, servers of different capacity are being used. Which
load balancing algorithm is best suited for your case? (6)

8. (a) A university department organizes seminars on a number of research topics. These

seminars take place in September each year. There is a computer program to help with
the administration of these seminars. In the program there is an object representing a
seminar, and this Seminar Class has a number of possible states. (13)
A seminar is first proposed by a professor. Once dates, times and rooms for the seminar
have been arranged, it is considered as scheduled. On the second Monday in August the
seminar is advertised and it becomes open for enrolment to the students. When all places
are taken, the seminar is full and cannot allow any more enrolment. If a student drops out
(cancels his/her application) the seminar reverts to the Open For Enrolment position. Two
days before the start of the seminar, enrolment will be closed (no more applicants will be
considered).
Design a State diagram to show different states of a Seminar in the context discussed
above.
(b) What is the difference between Reflected and DOM-based XSS attack? Draw the
flow ofa Reflected XSS attack? (12)
(c) Discuss Interactor, Indirect and Domain viewpoints for the requirement analysis of a
Stock Price Prediction system. Mention three non-functional requirements for such a
system. (10)