COMPUTER

SCIENCE
Practice Book - III
(DSA | OS | DBMS | CN | COA)

GATE | IES | PSUs

 Index
No. SECTIONS PAGES

1. Data Structure & Algorithm ..................................................................01-24

• Programming & Data Structures
• Sorting & Searching Techniques, Divide & Conquer
• Greedy Method & Dynamic Programming
2. Operating System ........................................................................................25-45
• Process, Threads, CPU Scheduling
• Concurrency, Synchronization, Deadlock
• Memory Management, File System
3. Database Management System ............................................................46-68
• ER-Model, Relational model
• Normal forms, SQL
• File Organization, Indexing, Transactions & Concurrency Control
4 Computer Networks ...................................................................................69-92
• Concept of Layering, LAN Technologies (Ethernet), Flow &
Error Control Techniques
• Switching, Network Layer (IPv4, IPv6), Routing Algorithms
• TCP-UDP and Sockets, Congestion Control, Application Layer
Protocols, Basics of Wi i, Network Security
5 Computer Organization & Architecture .......................................93-111
• Memory Hierarchy, Machine Instructions, Addressing
Modes, ALU, Control Unit
• Instruction Pipelining, I/O Interface (Interrupt and DMA mode)
 Data Structure & Algorithm

Topics: Programming & Data Structures

-------------------------------------------------------------------------------------------
LEVEL 1 switch (i)
{
Q.1 What is the output of the following i=i+ 10;
code? Case 10: printf(“10”);
void main( ) Case 15 : printf("i is 15");
{ break;
int const *p = 10; Case 20: printf("i is 20");
(*p)++; break;
printf(“%d", ++(*p)); default : printf (“no choice”);
} }
a) 10 return 0;
b) 11 }
c) 12 What of the output of above code?
d) Compiler Error a) Compiler error
b) i is 20
Q.2 Consider the following C program: c) i is 10 i is 15
int main () d) i is 10 i is 15 i is 20
{
int x = 20; Q.4 Consider following C program:
static int y = x; #include <stdio.h>
if (x == y) int main ()
printf(“Equal”); {
else char * ptr = “ICE_GATE”;
printf(“Not Equal”); char a, b;
return 0; for (int i = 0; i < 3; i++)
} {
What will be the output for above C a= *++ ptr;
program? b= *ptr++;
a) Equal }
b) Not Equal printf (“%c, %c", a, b);
c) Runtime error return 0;
d) Compile time error }
a) A, T
Q.3 Consider the following code: b) T, E
#include <stdio.h> c) A, A
int main () d) T,T
{
int i = 10;

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 1
 Q.5 Consider the following code with
inputs given are {1, 2, 3, 4, 5, 6, 7, 8, 9, Q.7 Consider the following C fragment:
10} (Assume, integer takes 4 bytes): main( )
#include <stdio.h> {
int main () char x{ ] = “ICE-GATE”;
{ char *y =x;
int arr [10], x, sum = 0; printf(“%s”, x + y[2] - y[5]):
for (int i= 0; i < 10; i++) }
{ What is the output of above C
x = scanf("%a", & arr fragment?
[#]); a) ICE-GATE
sum = sum + x; b) -GATE
} c) GATE
printf (“%d %a”, x, sum); d) TE
return 0;
} Q.8 The keys 15, 17, 23, 32, 37, 45, 57, 63,
What will be the output of the above 2, 5, 18, 73 are inserted into an initially
code? empty hash table of length 7 with hash
a) 10, 10 function h(k) = k mod 7. The collisions
b) 4, 40 are resolved by chaining. The average
c) 1, 40 chain length of hash table is ______.
d) 1,10 (Upto 2 decimal place) (Assume every
element stored outside the hash table)
Q.6 In delete operation of binary search
tree, we need inorder successor (or Q.9 Given a hash table T with load factor
predecessor) of a node when a node to 80 that stores 2000 elements, the
number of slots is ______.
be deleted where it has both left and
right child. Which of the following is Q.10 Which of the following statement is
true about inorder successor needed in correct reason for “Balanced binary
delete operation? tree is better than a complete binary
a) Inorder successor is always either tree for sorting a set S of n elements”?
leaf node or a node with empty right S1: A balanced binary tree is more
child. space efficient.
S2: Insertion and deletion can be
b) Inorder successor maybe an ancestor
preformed faster.
of the node. S3: Implementation of balanced binary
c) Inorder successor is always a leaf tree requires less RAM space.
node. Which of the following is always true?
d) Inorder successor is always either a (a) Only S1 and S2
leaf node or a node with empty left (b) Only S2
child. (c) Only S1 and S3
(d) Only S3

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 2
 LEVEL 2
Q.1 Consider the following AVL tree.
Sequences are:
S1: abedfc S2: acdebf
S3: abcfde S4: bacedf
S5: cdabef
The number of sequences are DFS
traversal on the given graph _______.

Q.5 Consider the following statements:

Which of the following order of S1: Linear probing and Quadratic
elements are inserted into an empty probing suffers from primary
AVL tree, so that it is possible to get clustering.
the above AVL tree? S2: Linear probing and Quadratic
probing suffers from secondary
a) 94, 71, 86, 25, 98, 83, 27, 90
clustering.
b) 98, 94, 90, 83, 86, 25, 71, 27 S3: Double hashing do not suffers from
c) 86, 25, 98, 83, 27, 90, 71, 94 primary clustering but suffers from
d) None of these secondary clustering only to a small
extent.
Q.2 The post order traversal of a binary Which of the following is true?
search tree is (a) Only S3
7,12,11,15,14,13,20,22,25,24,21,17 (b) Only S1 and S3
Then the preorder traversal of this tree (c) Only S2 and S3
is (d) Only S1 and S2
(a) 17,13,11,12,7,14,15,21,20,24,22,25
(b) 17,13,11,7,12,15,14,21,20,24,25,22 Q.6 What is the postfix expression for
(c) 17,13,11,7,12,14,21,15,20,24,22,25 A + (B x C - D/E ↑ F) x G) x H
(d) 17,13,11,7,12,14,15,21,20,24,22,25 infix expression?
a) ABC x DE/F ↑ GH x x H -
Q.3 The maximum difference between
b) ABC x - DE/F ↑ G x H x +
number of nodes in AVL tree when
tree is complete binary tree and c) ABC x DEF ↑ /G x -H x +
minimally connected tree contains 4 d) ABC x - DEF/↑ G x H x +
level is ____ (Assume root present at
level 1) Q.7 The number of binary search tree's with
4 nodes (1, 2, 3, 4) possible where in
Q.4 Consider the following graph and every binary search tree ‘1’ is a leaf
sequences given below: node are ____________

Q.8 Consider the following code:

void find (Node * head)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 3
 { enqueue(Q1,
Node * P1 = head, * P2 = head; dequeue(Q2));
while (P2) }
{ Insert the element x into stack S
P1=P1 → next; void push(S, x){
P2 = (P2 → next)? P2 enqueue(Q1, x);
→ next → next : NULL; }
} How many enqueue and dequeue
printf(“%d", P1 → value); operations are required to perform a
} pop operation if Q1 contains n
Assume Node is the structure type with elements initially?
two members: ‘value’ and ‘next’. a) 3n+1, 2n-1
Identify the node value printed by the b) 2n-1, 2n
above code if non-empty linked list c) 2n-2, 2n-1
header is passed to the function find? d) 2n-2, 3n-3
a) First element of the list [i.e., value of
the first node] Q.10 Consider the following C-fragment
b) Second element of the list where size of int is 1 B (Assume
c) Middle element of the list starting address of array is 1000)
d) Last element of the list int main ()
{
Q.9 An implementaion of a stack S, using int S[6] = {10, 20, 30, 40, 50, 60};
two queues Q1 and Q2 is given below. int * Str = (int *) (&S + 1);
Delete the element from stack S and printf(“% d”, Str);
store it in x: }
void pop(S, x) The output generated by above code
{ is_______.
if (isEmpty Q(Q1))
{
printf(‘S is empty”); LEVEL 3
return;
} Q.1 An n-ary tree in which every node has
while (1) 0 or n children. If number of internal
{ nodes is 40 and leaf nodes is 401, then
the value of ‘n’ is _____.
x = dequeue(Q1);
if (isEmptyQ(Q1))
Q.2 Consider the following elements are
break;
inserted into an empty AVL tree in the
enqueue(Q2, x);
given order: 50, 40, 30, 20, 10, 5,7
}
The number of rotations are applied to
while (! isEmptyQ(Q2))
construct an AVL tree with the given

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 4
 elements are (Assume LR, RL, LL and strings. Assume finding an empty slot
RR a single rotation) is a successful probe, while non-empty
is a unsuccessful probe. The total
Q.3 What will be the output of the number of unsuccessful probes needed
using linear probing to hash all strings
following C program?
is ____.
#include <stdio.h>
int main ()
{
char c='A’;
int x = 5, y= 10, res1, res2;
res1 = c++ || x++ || --y;
res2 = ++x && --c && ++y;
printf("%d %d %d %c", res1,
res2, x + y, c);
return 0;
}
a) 1 0 17 B
b) 1 1 17 A
c) 1 0 16 B
d) 1 1 16 B

Q.4 Consider unsorted doubly linked list

data structure containing n items. For
decrease key operation a pointer is
provided to the record on which
operation is performed. An algorithm
performs the following operations on
the list in this order ( n ) insert, O(n
log n) decrease key,
O(n) find operations. What is the time
complexity of all these operations put
together?
(a) O(n)
(b) O(n2)
(c) O(n2logn)
(d) O(n2 log n)

Q.5 Consider the following, five binary

string each of length 8 bits:
11111011, 01101010, 01010010,
11011011, 10011010
A hash table of size S= 8 (0 to 7) using
open addressing for hashing the binary

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 5
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(d) (d) (c) (c) (d) (d) (c) (1.71) (25) (b)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(c) (d) (8) (2) (a) (c) (5) (c) (c) (1006)

LEVEL 3

1 2 3 4 5
(11) (4) (b) (b) (8)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 6
 SOLUTIONS

LEVEL 1
Q.1 (d)
Compiler error: Cannot modify a
constant value.
P is a pointer to a “constant integer”.
Q.7 (c)
But we tried to change the value of the
“constant integer”.

Q.2 (d)
In C, static variable can be initialized x = 2000, y=2000
only using constant literals. They can't y[2] = E(69 in ASCII)
be assign like this, so it will generate y[5] = A(65 in ASCII)
compile time error. x + y[2] - y[5] = 2000 + 69 - 65 = 2004
Therefore it prints from the array
Q.3 (c) starting at address: 2004 to the end.
After the switch statement, the control
directly transfers to the matching case, Q.8 (1.7- 1.8)
without executing any statement
between them.

Q.4 (c)
Both a and b points at the 6th location
(index 5) of the array because once in a
loop, pointer is increamented twice.

Q.5 (d) Average chain length =

Since scanf( ) returns the number of (1+2+3+3+2+1+0)/7 = 1.71
variables it has stored successfully. So
in every iteration, it always stores one Q.9 (25)
element of array successfully, so (Number of elemnts)
Load factor =
returns 1 each time to variable x and (Number of slots)
since, loop is iterated 10 times, sum = 2000
80 =
10. x
So, x = 25.
Q.6 (d)
Successor of Root element is always Q.10 (b)
the smallest element of the Right S1 is false, since a Balanced binary tree
is never more space efficient.
subtree.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 7
 S3 is false, since Balanced binary tree So, Preorder: 17, 13, 11, 7, 12, 14, 15,
can never require less space. It usually 21, 20, 24, 22, 25
takes more space since some time we
need to store flags also i.e. in red-black Q.3 (8)
tree for color. And complete binary
tree can be store in a RAM as an array
without using pointers.

LEVEL 2
Q.1 (c)

Total nodes when tree is complete

Binary tree = 8+ 4+ 2+ 1=15 node

The order: 86, 25, 98, 83, 27, 90, 71, Total nodes when tree is balanced but
94 will result the given AVL Option minimally connected = 7
(c) is correct. [Note: Option (a) & So, maximum difference should be =
option (b) will generate different AVL 15-7 =8
trees]
Q.4 (2)
Q.2 (d) S1: ‘abedfc': DFS traversal.
We know that in order of BST is S2: ‘acdebf’: DFS traversal.
increasing order S3: ‘abefde': is neither BFS traversal
In order: 7, 11, 12, 13, 14, 15, 17, 20, nor DFS traversal.
21, 22, 24, 25 S4: 'bacedf’: BFS traversal but not
Post order: 7, 12, 11, 15, 14, 13, 20, 22, DFS traversal.
25, 24, 21, 17. S5: ‘cdabef’: BFS traversal but not
DFS traversal.

Q.5 (a)
S1: If two keys started from same hash
address, they both will follow same
path in linear manner, to get empty
slot. Because of this problem average
searching time increases and this

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 8
 situation is called primary clustering.
This is possible only in LINEAR
probing. Hence S1 is false.
S2: If they follow same path in
quadratic manner to get empty slot, the
average searching time increases (less
than linear probing). This is called
secondary clustering and is possible
only in quadratic probing.
Only S3 is absolutely true because 99%
secondary clustering do not exist in
double hashing.
Q.8 (c)
Q.6 (c)
P1 traverses node by node. P2 traverses
Postfix expressions can be found using
by skipping one node. The node
operator stack operations.
pointed by P1 is the middle node in the
Push A A
linked list when P2 reaches to NULL.
+ + A
:. Middle element of the list is printed
( +( A
by P1 → value.
Push B +( AB
x +(x AB
Q.9 (c)
Push C +(x ABC In first while loop, there will be n
- +(- ABC x {pop x, because dequeue and n-1 enqueue operations
precedence of x is greater than -} performed.
( +(-( ABC x In Second for loop, there will be n-1
Push D +(-( ABC x D dequeue and n-1 enqueue operations
/ +(-(/ ABC x D performed.
Push E +(-(/ ABC x DE So, Total 2n-2 enqueue and 2n-1
↑ +(-(/↑ ABC x DE dequeue operations will be performed
F +(-(/↑ ABC x DEF for a pop operation.
) +(- ABC x DEF ↑/ {pop/ ↑}
x +(-x ABC x DEF ↑/ Q.10 (1006)
G +(-x ABC x DEF ↑/G
) (+ ABC x DEF ↑/ G x –
Push x +x ABC x DEF ↑/G x –
Push H +x ABC x DEF ↑/G x - H
) - ABC x DEF↑/G x-H x +

Q.7 (5)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 9
 LEVEL 3 evaluates all components and returns
true.
Q.1 (11)
Let I be the number of internal nodes, Q.4 (b)
L be number of leaf nods and T be total Unsorted doubly linked list:
number of nodes
I+L=T
n.I + 1 = T
From (i) and (ii);
I + L = nI + 1 Q.5 (8)
40 + 401 = 40.n + 1 Checks last 3 bits:
440 = 40.n
n = 11

Q.2 (4)

Total unsuccessful probes = [2 + 2 + 4]

= 8 probes.

Q.3 (b)
|| and && are also called short circuit
operators because:
|| operator, evaluation takes place from
left to night and stops evaluating next
component when it gets true value. So
far res1, c++ returns true value so
further components it does not
evaluate. And && operator, evaluation
takes place from left to right, it
evaluates all of its components until it's
getting true value, whenever it gets
false value, next components doesn't
evaluate and stops. So for res2, it

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 10
 Data Structure & Algorithm

Topics: Sorting & Searching Techniques, Divide & Conquer

-------------------------------------------------------------------------------------------
LEVEL 1 (b) T(n) = T(n – 1) + T(0) + cn
(c) T(n) = 2T (n – 2) + cn
(d) T(n) = T(n/2) + cn
Q.1 Which of the following sorting
techniques have minimum number of Q.6 Randomized quicksort is an extension
comparison in best case? of quicksort where the pivot is chosen
(a) Insertion sort randomly. What is the worst case
(b) Selection sort complexity of sorting n numbers using
(c) Merge sort randomized quicksort?
(d) Quick sort (a) O(n)
(b) O(n log n)
Q.2 Which of the following represents the (c) O(n2)
worst case time complexity to check (d) O(n!)
whether a given tree is min heap tree or
not with ‘n’ nodes? Q.7 Consider the Quick sort algorithm.
(a) O(logn) Suppose there is a procedure for
(b) O(n) finding a pivot element which splits the
(c) O(nlog n) list into two sub-lists each of which
(d) O(n2) contains at least one-fifth of the
elements. Let T(n) be the number of
Q.3 Which of the following sorting comparisons required to sort n
algorithms does not have a worst case elements. Then
running time of O(n2)? (a) T(n) <= 2T(n/5) + n
(a) Insertion sort (b) T(n) <= T(n/5) + T(4n/5) + n
(b) Merge sort (c) T(n) <= 2T(4n/5) + n
(c) Quick sort (d) T(n) <= 2T(n/2) + n
(d) Bubble sort
Q.8 In quick sort, for sorting n elements,
Q.4 For a linear search in an array of n the (n/4)th smallest element is selected
elements the time complexity for best, as pivot using an O(n) time algorithm.
worst and average case are ......., ....... What is the worst case time complexity
and ........ respectively of the quick sort?
(a) O(n), O(1), and O(n/2) (a) θ (n) (b) θ (n Log n)
(b) O(1), O(n) and O(n/2) (c) θ (n2) (d) θ (n2 log n)
(c) O(1),O(n) and O(n)
(d) O(1), O(n) and (n-1/2) Q.9 Given an array where numbers are in
range from 1 to n6, which sorting
Q.5 Which one of the following is the algorithm can be used to sort these
recurrence equation for the worst case number in linear time?
time complexity of the Quick sort (a) Not possible to sort in linear time
algorithm for sorting n(≥ 2) numbers? (b) Radix Sort
In the recurrence equations given in the (c) Counting Sort
options below, c is a constant. (d) Quick Sort
(a) T(n) = 2T (n/2) + cn

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 11
 Q.10 If the number of records to be sorted is to a join algorithm. The output is an
small, then ...... sorting can be efficient. array C[1,... 2n] which has all the value
(a) Merge of the arrays A[] and B[] and is in
(b) Heap increasing order. What is the worst
(c) Insertion case time complexity for join algorithm
(d) Bubble to join the two array?
(a) O(n2)
LEVEL 2 (b) O(n)
(c) O(n log n)
(d) O(1)
Q.1 Consider the following statements:
S1: Implementation of stack using Q.4 What is the worst case time complexity
queue, deletion of second element from to design unique binary search tree
top of stack time complexity O(n), with n elements, when its preorder and
when insertion take O(1) time. post order is given?
(a) O( n )
S2: In implementation of queue using (b) O(n)
stack, deletion of second element from (c) O(nlogn)
(d) Cannot be designed using BST
front take 0(1) time, when insertion
take O(n) time.
Q.5 Consider two sorted arrays A and B
Which of the following is correct ? has distinct elements, but there may be
a) Only S1 many elements common in both A & B
b) Only S2 What is the best case time complexity
c) Both S1 and S2 intersection of two array A and B?
d) None of the above (a) O(log n) (b) O(n)
(c) O(n log n) (d) O(n2)
Q.2 Consider the following graph:
Q.6 Using the best algorithm, what is the
time complexity to determine that a
number n is prime?
(a) O( n ) (b) O(n)
(c) O(log n) (d) O(n2)

Q.7 Assume that a merge sort algorithm in

the worst case takes 30 seconds for an
input of size 64. Which of the
If DFS algorithm applied starting from following most closely approximates
vertex ‘A’ which uses stack data the maximum input size of a problem
structure then the height of stack is that can be solved in 6 minutes?
needed in worst case for DFS traversal (a) 256 (b) 512
is ______. (c) 1024 (d) 2048

Q.3 Consider two arrays A[] and B[], if Q.8 Let P be a Quick sort Program to sort
A[1... n] is in increasing order, and numbers in ascending order using the
B[1... n] is in decreasing order is input first element as pivot. Let t1 and t2 be

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 12
 the number of comparisons made by P int i, X[N], Y[N], Z[N];
for the inputs {1, 2, 3, 4, 5} and {4, 1, X[0]=Y[0]=Z[0]=0;
5, 3, 2} respectively. Which one of the X[1]=1; Y[1]=2; Z[1]=3;
following holds? for i=2, i≤ n, i++ {
(a) t1 = 5 (b) t1< t2 X[i]= Y[i-1] + Z[i-2];
(c) t1> t2 (d) t1 = t2 Y[i]= 2 * X[i];
Z[i] = 3 * X[i];
Q.9 State True or False for internal sorting }
algorithms: return X[n];
(i) Internal sorting are applied when the }
entire collection if data to be sorted is
small enough that the sorting can take Q.2 The running time of f1(n) and f2(n) are
place within main memory. (a) Ѳ(n) and Ѳ(n)
(ii) The time required to read or write (b) Ѳ(2n) and Ѳ(n)
is considered to be significant in (c) Ѳ(n) and Ѳ(2n)
evaluating the performance of internal (d) Ѳ(2n) and Ѳ(2n)
sorting.
(a) (i)- True, (ii)- True Q.3 f1(8) and f2(8) returns the value of
(b) (i)- True, (ii)- False (a) 1661 and 1640
(c) (i)- False, (ii)- True (b) 59 and 59
(d) (i)- False, (ii)- False (c) 1640 and 1640
(d) 1640 and 1641
Q.10 .......... is putting an element in the
appropriate place in a sorted list yields Q.4 Which of the following is correct
a larger sorted order list. recurrence for worst case of Binary
(a) Insertion (b) Extraction Search?
(c) Selection (d) Distribution (a) T(n) = 2T(n/2) + O(1) and T(1) = T(0) = O(1)
(b) T(n) = T(n-1) + O(1) and T(1) = T(0) = O(1)
(c) T(n) = T(n/2) + O(1) and T(1) = T(0) = O(1)
LEVEL 3 (d) T(n) = T(n-2) + O(1) and T(1) = T(0) = O(1)

Q.1 Let A, B, C, D, E are sorted sequences Q.5 Given a sorted array of integers, what
having length 70, 74, 80, 85, 102 can be the minimum worst case time
respectively. They are merged into a complexity to find ceiling of a number x
single sequence by merging together in given array? Ceiling of an element x is
two sequences at a time. The maximum the smallest element present in array
number of comparison that will be which is greater than or equal to x.
needed by best algorithm for doing Ceiling is not present if x is greater than
merging is ____. the maximum element present in array.
For example, if the given array is {12,
Common Data for Questions 2 & 3 67, 90, 100, 300, 399} and x = 95, then
int f1 (int n){ output should be 100.
if (n= =0 || n= =1) return n; (a) O (Log Log n)
else return (2*f1 (n-1) + 3*f1 (b) O (n)
(n-2)); (c) O (Log n)
} d) O (Log n * Log n)
int f2 (int n){

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 13
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(a) (b) (b) (c) (b) (c) (b) (b) (b) (c)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(c) (7) (b) (c) (b) (a) (b) (c) (b) (a)

LEVEL 3

1 2 3 4 5
(962) (b) (c) (c) (c)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 14
 elements, T(4n/5) is for the rest 4n/5
SOLUTIONS elements, and n is for finding the pivot.
If there are more than n/5 elements in
LEVEL 1 one set then other set will have less
than 4n/5 elements & time complexity
will be less than T(n/5) + T(4n/5) + n
Q.1 (a) because recursion tree will be more
If input is already sorted, then insertion balanced.
sort will give O(n) comparison i.e., for
every element 1 comparison. i.e., n
Q.8 (b)
element = O(n) comparison.
The recursion expression becomes:
While selection sort have O(n2), quick T(n) = T(n/4) + T(3n/4) + cn
sort O(n log n) and merge sort have
After solving the above recursion, we
O(n log n). get θ (nLogn).
Q.2 (b) Q.9 (b)
Visit each node one time by comparing
Q.10 (c)
parent node with its children, will take
O(n) time in worst case.
LEVEL 2
Q.3 (b)
Q.4 (c) Q.1 (c)
Average time required is (n+1)/2 but In implementation of stack using queue:
time complexity is O(n). If we do some extra work in inserting the
element then deletion of second element
Q.5 (b) from top of stack required queue (n - 1)
In worst case, the chosen pivot is
always placed at a corner position and element in another queue then again (n -
recursive call is made for following. 1) i.e. (n - 2) element to another queue so
a) for subarray on left of pivot which is it take O(n) time.
of size n-1 in worst case. In implementation of using stack; will
b) for subarray on right of pivot which take O(1) time to delete second element
is of size 0 in worst case. from front will take O(1) time.
Q.6 (c)
Q.2 (7)
Randomized quick sort has expected
Maximum height of stack needed,when
time complexity as O(nLogn), but
depth of the graph is maximum, so,
worst case time complexity remains
same. In worst case the randomized
function can pick the index of corner
element every time. So, it is O(n2).

Q.7 (b)
For the case where n/5 elements are in
one subset, T(n/5) comparisons are
needed for the first subset with n/5 Which is 7, so maximum height of
stack is 7.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 15
 Q.3 (b) In every step of quick sort, numbers are
This Join algorithm is similar to merge divided as per the following ecurrence:
algorithm in merge sort. It takes O(n + T(n) = T(n-1) + O(n)
n) = O(2n) = O(n).
Q.9 (b)
Q.4 (c) Q.10 (a)
To create unique BST, we need either
inorder and preorder or inorder and LEVEL 3
postorder.
Since inorder of BST is element in
Q.1 (962)
increasing order. So, sorting of
Given file size: 70, 74, 80, 85, 102
elements take O(n logn) time then
create min heap.
design unique BST from inorder and
Algorithm: Take 1st 2 min element,
preorder will take O(n) time.
merge it, and put back result into heap,
So total time = O(nlogn) + O(n) =
again find next two min element,
O(nlogn)
merge them, repeat until one element
left.
Q.5 (b)
Since both arrays are sorted, so
applying merge procedure will take (n
+ n - 1) = O(n) time.

Q.6 (a)
If (n%2==0) then {return "not prime";}
Else {
For(i=3;i<(√n)+1;i=i+2;){
If (n % I == 0) then
{return "not prime";}
}
}return "prime";

Q.7 (b)
Time complexity of merge sort is
Θ(nLogn)
c*64 Log 64 = 30
c*64*6 = 30
c = 5/64
For time 6 minutes: Q.2 (b)
5/64*n Log n = 6*60 The recursive and iterative approach of
the same algorithm.
n Log n = 72*64 = 512 * 9
n = 512.
Q.3 (c)
Q.4 (c)
Q.8 (c)
When first element or last element is Q.5 (c)
The time complexity T(n) can be written
chosen as pivot, Quick Sort‘s worst
as: T(n) ≤ T(n/2) + O(1) ⇒ O (Log n).
case occurs for the sorted arrays.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 16
 Data Structure & Algorithm

Topics: Greedy Method & Dynamic Programming

-------------------------------------------------------------------------------------------
LEVEL 1 (c) Exponential time
(d) Quadratic time
Q.1 Does greedy approach give an optimal Q.6 Which of the following is true about
solution for every instance of knapsack Huffman Coding?
problem? (a) Huffman coding may become lossy
(a) Yes in some cases
(b) No (b) Huffman Codes may not be optimal
(c) Never gives an optimal solution lossless codes in some cases
(d) Greedy approach is not applicable (c) In Huffman coding, no code is
for knapsack problem prefix of any other code.
(d) All of the above
Q.2 Which of the following is not an
application of greedy method? Q.7 Assume that enqueing and dequeing
(a) Minimum – spanning tree algorithm operation take O(V) time each, during
(b) Dijkstra’s algorithm for shortest BFS algorithm on a graph. What would
paths from a single source. be the worst case time complexity of
(c) Job Scheduling Problem BFS algorithm? (V is the number of
(d) Longest Common Subsequence vertices in the graph and graph is given
as input to BFA algorithm in the form
Q.3 Consider the following statements: of adjacency list).
S1: Kruskal’s algorithm and Prim’s (a) O(V3)
algorithm both are greedy algorithms. (b) O(V2E)
S2: Kruskal’s algorithm and Prim’s (c) O(V2)
algorithm both are dynamic (d) O(V)
programming.
(a) Only S1 is true Q.8 Which of the following are
(b) Only S2 is true asymptotically tight.
(c) Both S1 and S2 is true I. 2n2 = O(n2)
(d) Neither S1 nor S2 is true. II. 2n = O(n2)
III. n2 = O(n5)
Q.4 A greedy algorithm always makes the IV. n4 = Ω(n3)
choice that looks ______ at the (a) Only I
moment. (b) Only II
(a) Worst (c) Only II, III
(b) Best (d) Only II, III and IV
(c) Average
(d) Can’t say anything. Q.9 Consider the following knapsack
instance.
Q.5 What is the time complexity of • Objects: (x1, x2, …, x7)
travelling salesman problem with n
vertices using dynamic programming? • Weights: (W1, W2, W3, …, W7) = (3,
(a) Linear time 4, 6, 7, 2, 4, 2)
(b) Logarithmic time

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 17
 • Profits: (P1, P2, …, P7) = (12, 8, 20, 8, P Q R S
10, 24, 5) (a) (iv) (i) (ii) (iii)
• Capacity of the bag (m) = 17 (b) (iv) (i) (iii) (ii)
(c) (iii) (i) (ii) (iv)
If knapsack problem is solved using (d) (iii) (i) (iv) (ii)
maximum profit per unit weight then
the overall profit is __________. Q.3 The cost of minimal spanning tree
(using prim’s algorithm) for below
Q.10 A graph G is such that every vertex is graph is ______ (root vertex is ‘a’).
connected to every other vertex via
direct link. We need to find shortest path
between two vertices using Bellman ford
algorithm. How much time does it take?
(a) O(V2 log V)
(b) O(V2)
(c) O(V3)
(d) (V+E) log V (a) 37
(b) 38
(c) 36
LEVEL 2 (d) 28
Q.1 Consider f(n), g(n) and h(n) be Q.4 Consider the following data:
function defined as follows: W1 = 2, W2 = 3, W3 = 4, W4 = 1
P1 = 10, P2 = 15, P3 = 5, P4 = 12 and
Knapsack capacity = 5
Which of the following is the possible
combination of weights for maximum
Which of the following represents profit using 0/1 knapsack concept?
correct asymptotic solution for f(n) + (a) (W1, W2)
[g(n) x h(n)]? (b) (W2, W4)
(a) Ω(n3 ) (c) (W3, W4)
(b) Ο(n4) (d) (W1, W4)
(c) Θ(n4)
(d) O(n3) Q.5 The correct matching for the following
pairs is
Q.2 Consider a list of problem and a list of (a) All pairs shortest path
corresponding time complexity using (b) Merge sort
dynamic programming: (c) Minimum weight spanning tree
Problem Time complexity (d) Connected Components
P. Matrix chain multiplication (i) O(n)
Q. Fibonacci number (ii) O(n2) (1) Greedy
R. LCS problem (m, n) (iii)O(2n) (2) Depth-first search
S. 0-1 Knapsack (m, n) (iv)O(n3) (3) Dynamic programming
(4) Divide and conquer
Which of the following is the correct
match between the problem and their (a) A-2, B-4, C-1, D-3
time complexity? (b) A-3, B-4, C-l, D-2

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 18
 (c) A-3, B-4, C-2, D-1 (a) X[m] ≥ Y[n], m, m – 1
(d) A-4, B-1, C-2, D-3 (b) X[m] ≤ Y[n], m, m – 1
(c) X[m] = = Y[n], m, m + 1
Q.6 If graph contains negative weight (d) None of these
edges then which of the following is
correct when we run dijkstra’s Q.10 Using Kruskal’s algorithm how many
algorithm? distinct minimum cost spanning trees are
(a) It may not terminate possible?
(b) It terminates but may produce
incorrect result
(c) It never terminates due to cycle in
graph
(d) None of these
(a) 3
Q.7 What is the time complexity of (b) 4
Huffman Coding? (c) 5
(a) O (N) (d) 6
(b) O (NlogN)
(c) O (N (logN) ^ 2) LEVEL 3
(d) O (N^2)
Q.1 Assume Dijkstra’s algorithm to find
Q.8 Assume Dijkstra’s algorithm is used to
the shortest paths from node ‘A’ in
find the shortest paths from node ‘A’ in
the following graph. following graph:

The number of edges are there which are

not included in any of the shortest paths
from node A is ___________. Find the no. of edges not included in
any of the shortest paths from G.
Q.9 In the longest common-subsequence
(LCS) problem we are given two Q.2 A networking company uses a
sequences X=〈x1,x2,…xm〉 and compression technique to encode the
Y=y1,y2,…yn and we wish to find a message before transmitting over the
maximum length common subsequence network. Suppose the message contains
of X and Y. The following is the the following characters with their
recursive formulation of LCS problem. frequency:

Identify the respective expressions for

A, B and C.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 19
 (b) 44000
(c) 19000
(d) 25000

Q.5 Consider two strings A = “qpqrr” and B

= “pqprqrp”. Let x be the length of the
longest common subsequence (not
necessarily contiguous) between A and B
and let y be the number of such longest
common subsequences between A and
B. Then x + 10y = _______.
(a) 33
If the compression technique used is (b) 23
Huffman Coding, how many bits will (c) 43
be saved in the message? (d) 34
(a) 224
(b) 800
(c) 576
(d) 324

Q.3 Consider a set of 4 messages (M1 –

M4) whose frequency of occurrences
in the text is as given:
(0.37, 0.51, 0.05, 0.07)
Using frequency dependent Huffman
coding the codes of the messages M2
and M3 respectively is?
(a) 0, 110
(b) 0, 011
(c) 1, 000
(d) 1, 001

Q.4 Four matrices M1, M2, M3 and M4 of

dimensions pxq, qxr, rxs and sxt
respectively can be multiplied is
several ways with different number of
total scalar multiplications. For
example, when multiplied as ((M1 ×
M2) × (M3 × M4)), the total number of
multiplications is pqr + rst + prt. When
multiplied as (((M1× M2) ×M3) ×
M4), the total number of scalar
multiplications is pqr + prs + pst.
If p = 10, q = 100, r = 20, s = 5 and t =
80, then the number of scalar
multiplications needed is
(a) 248000

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 20
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(b) (d) (a) (b) (c) (c) (c) (a) (71) (c)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(a) (a) (a) (b) (b) (b) (b) (4) (d) (b)

LEVEL 3
1 2 3 4 5
(4) (c) (c) (c) (d)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 21
 Adjacency list is scanned for every
SOLUTIONS vertex, when it is dequed. So overall
O(E) time to scan adjacency lists
LEVEL 1 ∴ O(V2 + E) = O(V2)

Q.8 (a)
Q.1 (b)
II, III and IV are not asymptotically
It does not give optimal solution of 0/1
tight. They can be expressed with more
knapsack problem.
tightness.
2n = o(n2)
Q.2 (d)
n2 = o(n5)
Longest common subsequence is the
n4 = ⍵(n3)
example of dynamic programming.
I is already asymptotically tight.
Q.3 (a)
Q.9 (71)
Both prim’s and kruskal’s are greedy
algorithm.  P1 P2 P3 P7 
 , , ,..., 
 W1 W 2 W 3 W7
Q.4 (b) = (4, 2, 3.33, 1.142, 5, 6, 2.5)
A greedy algorithm always makes the Decreasing order of Pi / Wi is x6, x5,
choice that looks best at the moment. x1, x3, x7, x2, x4
Since capacity of bag is 17 we can only
Q.5 (c) add the following items using the
Travelling salesman problem with Pi / Wi ratio.
dynamic programming has exponential x6(4) + x5(2) + x1(3) + x3(6) + x7(2)
time complexity. Total profit = 1 × 24 + 1 × 10 + 1 × 12
+ 1 × 20 + 1 × 5
Q.6 (c) = 24 + 10 + 12 + 20 + 5
Huffman coding is a lossless data = 71
compression algorithm.
The codes assigned to input characters Q.10 (c)
are Prefix Codes, means the codes are In Bellman ford algorithm for every
assigned in such a way that the code round ‘E’ edges are relaxed. Each edge
assigned to one character is not prefix relaxation takes constant time. In
of code assigned to any other character. general there are (V – 1) rounds
This is how Huffman Coding makes excluding the verification round.
sure that there is no ambiguity when Total time = (V – 1) × O(E) × O(1)
decoding. = O(VE)
But for complete graph E = O(V)
Q.7 (c) ∴ Total time = O(V3)
In normal BFS both enqueing and
dequeing operation takes O(1) but in LEVEL 2
this algorithm it takes O(V) time.
All the vertices will be enqued and Q.1 (a)
dequed once.
∴ V(V + V) = O(V2) times

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 22
 Q.4 (b)
Knapsack capacity is 5
In 0/1 knapsack, either you have to
consider full weight or no weight,
fraction is not allowed.
So, W4 = 1
W2 = 3
--------------------
Weight = 4
We will get max profit = 27.
So, (W2, W4) is the combination.

Q.5 (b)

Q.6 (b)
It terminates after |E| relations and |V|
Q.2 (a)
Matrix chain multiplication using + |E| priority queue operations, but may
dynamic programming take O(n3) time produce incorrect results.
while without dynamic programming
(nn). Q.7 (b)
Fibonacci number take O(n) using O (nlogn) where n is the number of
dynamic programming while O(2n) unique characters. If there are n nodes,
without dynamic programming. extractMin() is called 2*(n – 1) times.
LCS problem take O(n2) with dynamic extractMin() takes O(logn) time as it
programming and without dynamic called minHeapify().
programming O(22n). 0-1 Knapsack So, overall complexity is O(nlogn).
take O(n2) with dynamic programming
i.e. one of NP-complete problem, Q.8 (4)
because of very less repetition O(m, n) A → E : 3[edge e1]
≈ O(2n) and without dynamic A → D : 6[edge e3]
programming O(2n). A → B : 4[edge e2]
A → C : 9[edge e2, edge e5]
Q.3 (a) The edges e4, e6, e7 and e8 are never
In prim’s algorithm we start with root used in any of the shortest paths.
vertex and grow until the tree spans all
vertices in v. At each step an edge with Q.9 (d)
The following is the respective
lowest cost is added to the tree.
expressions for A, B and C.
A: X[m] = = Y[n]
B: m
C: m – 1

Q.10 (b)
Cost = 4 + 8 + 2 + 4 + 2 + 1 + 7 + 9
= 37

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 23
 Only these four minimum cost Iteration 4:
spanning trees are possible.

LEVEL 3
Q.1 (4)
Edge (T, M), (M, E), (M, A) and (A, E)
are not included.

Q.2 (c)
Total number of characters in the
message = 100.
Each character takes 1 byte. So total
number of bits needed = 800. Q.4 (c)
After Huffman Coding, the characters It is basically matrix chain multiplication
can be represented with: problem. We get minimum number of
f: 0 multiplications using ((M1 × (M2 × M3))
c: 100 × M4).
d: 101 Total number of multiplications =
a: 1100 100×20×5 (for M2 × M3) + 10×100×5
b: 1101 + 10×5×80 = 19000.
e: 111
Total number of bits needed = 224 Q.5 (d)
Hence, number of bits saved = 800 – The LCS is of length 4. There are 3 LCS
224 = 576 of length 4 “qprr”, “pqrr” and “qpqr”.

Q.3 (c)
Iteration 1:

Iteration 2:

Iteration 3:

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 24
 Operating System

Topics: Process, Threads, CPU Scheduling

-------------------------------------------------------------------------------------------
LEVEL 1 d) Increase/decrease degree of
multiprogramming
Q.1 The time required for the process to
complete its execution is called? Q.6 Which Disk scheduling policy may
a) Completion time suffer from starvation?
b) Turn Around Time a) FCFS
c) Burst Time b) SSTF
d) Waiting Time c) SCAN
d) Both (b) and (c)
Q.2 Which of the following is/are not
shared by threads in a process? Q.7 Consider the following statements with
I. Program counter respect to FCFS scheduling
II. Address space 1. Waiting time can be large if short
III. Registers requests wait behind the long ones. 2.
a) I and III only It is not suitable for time sharing
b) II only systems where it is important that each
c) I only user should get the CPU for an equal
d) II & III only amount of time interval.
3. A proper mix of jobs is needed to
Q.3 Which of the following scheduling achieve good results from FCFS
algorithms is non-pre-emptive? scheduling.
a) Multilevel Queue Scheduling Which of the above statements are
b) Multilevel Queue Scheduling with true?
Feedback a) 1 and 2 only
c) Shortest Remaining Time First b) 2 and 3 only
d) Shortest Job First c) 1 and 3 only
d) All statements are correct
Q.4 Which of the following scheduling
algorithms could result in starvation? Q.8 Four processes arrive in the order A, B,
I. FCFS C, D at time 5th; with Burst time
II. SJF requirements of 4, 1, 8, 1 respectively.
III. Round Robin Using Round Robin with time
IV. Priority Quantum of 1 unit, the completion time
a) I, II and III only of process A is ___.
b) II, III and IV only
c) II and III only Q.9 Which of the following scheduling can
d) II and IV only be done by thread library?
a) Process scheduling
Q.5 Which of the following is NOT a CPU b) User thread scheduling
scheduling criteria for performance c) Kernel thread scheduling
comparison? d) None
a) Turn Around time and waiting time
b) CPU utilization and throughput
c) Interactiveness and response time

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 25
 Q.10 Consider the following set of processes Q.3 Consider the exponential average
that need to be scheduled on a single formula used to predict the length of
CPU. the next CPU burst.
If α = 0 and T0 = 100 ms. The next
CPU burst time of the process is?
where T0 represents predicted burst
time of process.

Q.4 int i;
for(i=0;i<3;i++) {
Assume all times are given in
fork( );
milliseconds and round robin CPU
fork( );
scheduling is used with time quantum 2
}
msec. The average turnaround time is
If a process executes the above code,
________ (in msec).
after the execution of the code, the
number of child processes are _____.
LEVEL 2
Q.5 An operating system uses SRTF
Q.1 Let the time taken to switch between algorithm. Consider arrival time and
two process is t1, time taken to switch execution time for the following
between two kernel level threads is t2 processes:
and time taken to switch between two
user level threads is t3.
Then which of the following is true?
a) t1 < t2 < t3
b) t1 > t2 = t3
c) t1 > t2 > t3
d) t1 = t2 > t3 What is total waiting time for P1?
a) 5
Q.2 Match List1 with List2. b) 6
List1 c) 7
(A) FCFS d) 8
(B) Round Robin
(C) Working set algo. Q.6 A scheduler which selects partial
(D) SJF executed processes from secondary
storage device is called
List2 a) Short term scheduler
(1) Least response time b) Long term scheduler
(2) Min. average waiting time c) Medium term scheduler
(3) Belady’s anomaly d) Process scheduler
(4) Principle of Locality
Q.7 Considering the exponential average
a) A – 4, B – 3, C – 2, D – 1 behavior used to predict the next CPU
b) A – 3, B – 1, C – 4, D – 2 burst. If α = 0.80 and τ0 = 25 ms and
c) A – 2, B – 1, C – 3, D – 4 previous (T0, T1, T2, T3) runs were as
d) A – 1, B – 2, C – 3, D – 4 10, 12, 15, 20. The predicted value of
τ4 is _______

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 26
 d) m, 0, n
Q.8 Consider the following set of processes
that need to be scheduled on a single Q.2 Which of the following is TRUE?
CPU. a) User level threads are scheduled by
the kernel.
b) When a user level thread is blocked,
all other threads of its process are
blocked.
c) Context switching between user
level threads is slower than context
What is the throughput of the system if switch between kernel level threads.
SRTF CPU scheduling is used? d) Kernel level threads cannot share
a) 0.30 data segment.
b) 0.25
c) 0.35 Q.3 A process executes the following code
d) 0.40 for (i=0 ; i < n-1 ; i + +)
fork ( );
Q.9 Disabling interrupts solution: The number of child process created is
(1) Guarantees Mutual Exclusion a) 1
(2) Guarantees Bounded Waiting b) n-1
(3) Guarantees Progress c) 2n-1 – 1
(4) Most feasible solution for real time d) 2n - 1
Which of the above statements are
true? Q.4 Suppose in a system processes arrive at
a) 1, 3 an average rate of 6 per minute. On
b) 1, 2, 3 average each process takes 8 seconds
c) 2, 4 of service time. Estimate the % of time
d) 1, 2, 3, 4 CPU is idle in a single processor
system is _____.
Q.10 What is the common problem seen in
Strict Alteration & Peterson’s solution? Q.5 Consider 3 processes A, B, C to be
a) Progress is not guaranteed scheduled as per the Shortest
b) Busy waiting Remaining Time First (SRTF)
c) Both A & B algorithm. The process 'A' is scheduled
d) None of the above first and when 'A' has run for 7 units of
time, the process 'C' has arrived.
LEVEL 3 Process C gets its CPU immediately
after arrival and has run for 1 unit of
Q.1 The maximum, minimum and time; then process B arrives and
maximum number of processes that preempts process C. Process B takes 2
can be in running, blocked and ready time units to complete its execution.
state respectively for a computer After that process C completes and at
system with n process and m CPUs is, the end process A completes. The
a) 1, n, n minimum burst time of process A is __
b) n, 0, m
c) m, 0, m .

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 27
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(c) (a) (d) (d) (d) (d) (d) (14) (b) (7.75)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(c) (b) (100) (63) (b) (c) (18.9) (a) (a) (b)

LEVEL 3

1 2 3 4 5
(d) (b) (c) (20) (12)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 28
 SOLUTIONS
LEVEL 1

Q.1 (c)
By definition, the time required for the Q.9 (b)
process to complete its execution is Thread management is done in user
called Burst Time. space by thread library.

Q.2 (a) Q.10 (7.75)

All the threads share address space but
other entities like, stack, PC, resisters
are not shared.

Q.3 (d)
Shortest Remaining Time First is Pre-
emptive version of SJF.
SJF is non pre-emptive.
Average = 31 / 4 = 7.75 msec
Q.4 (d)
FCFS & Round Robin are starvation LEVEL 2
free.
Q.1 (c)
Q.5 (d) Time taken for processes to switch is
Increasing or decreasing degree of more than time taken for kernel level
multiprogramming is the constraint of threads switching which is more than
long term scheduler. user level threads switching time.

Q.6 (d) Q.2 (b)

FCFS suffers from convoy effect. It FCFS suffers from Belady’s anomaly.
does not suffer from starvation, while Round robin has least response time
SSTF and SCAN suffers from among all scheduling algorithms.
starvation. Working set algorithm follows
Principle of Locality.
Q.7 (d) SJF has minimum average waiting time
Waiting time can be more if short among all scheduling algorithms.
processes wait for long ones.
For time sharing systems Round Robin Q.3 (100)
scheduling is suitable. Use the formula
A proper mix of jobs results in less Tn+1= α tn+(1 - α)Tn= 0+1(100)=100ms
waiting time & turnaround time.
Q.4 (63)
Q.8 (14) fork() statement will be executed 6
Completion time of process A = 14. times.
Hence child processes = 26 – 1 = 63.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 29
 Q.5 (b)
Q.2 (b)
(a) is false, because user level threads
At t = 5, P1 arrives and executed for 2 are not scheduled by kernel. Kernel
ms until P2 arrives (so no waiting time level threads are scheduled by kernel.
here) At t = 13 again P1 gets the (b) is true, because OS is unaware of
execution. So waiting time for P1 = 6. user level threads. So, when one thread
is blocked entire process gets blocked.
Q.6 (c) (c) is false, because context switching
is faster in user level threads compare
Q.7 (18.9) (18.7 - 19.5) to kernel level threads.
τn+1 = α Tn + (1-a) τn (d) is false, because all the threads
τ1 = 0.8(10) + (0.2) (25) = 13 share code and data segments.
τ2 = 0.8(12) + (0.2) (13) = 12.2
τ3 = 0.8(15) + (0.2) (12.2) = 14.44 Q.3 (c)
τ4 = 0.8(20) + (0.2) (14.44) = 18.999 ≈ We know that the total number of
19.0 processes is 2n-1 but we subtract the
parent process. Therefore, total number
Q.8 (a) of child processes = 2n-1 – 1
GANTT Chart
Q.4 (20)
CPU is busy for 48 seconds.
Throughput = 6 processes × 8 seconds = 48 sec
60 seconds 100%
No. of processes
48 seconds → 48/60 × 100% = 80%
Max(completion time)- Min(Arrival time) CPU idle = 20%
4
= = 0.307
13 − 0 Q.5 (12)
Process C runs for 1 unit of time and is
Q.9 (a) pre-empted by Process B (Which takes
Disabling interrupt guarantees Mutual 2 unit of time), that means remaining
Exclusion and Progress. time of Process C when Process B
It does not guarantee Bounded waiting. (with burst time 2) arrives should be
It is not suitable for real time system minimum 3. Hence minimum burst
because interrupts are very useful in time of Process C = 1 + 3 = 4.
real time systems. Now, Process A runs for 7 unit of time
and preempted by Process C with burst
Q.10 (b) time remaining 4, that means
remaining time of Process A should be
LEVEL 3 minimum 5.
Hence minimum burst time of process
Q.1 (d) A = 7 + 5 = 12.
Ready Running Block
state state state
Min. 0 0 0
Max. n m n

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 30
 Operating System

Topics: Concurrency, Synchronization, Deadlock

-------------------------------------------------------------------------------------------
LEVEL 1 a) <P3 P0 P2 P1>
b) <P3 P2 P0 P1>
Q.1 Consider there are 3 processes P1, P2 c) <P2 P3 P1 P0>
and P3 each require 2 unit, 3 unit and 4 d) <P2 P3 P0 P1>
units of resource 'R' respectively.
Maximum number of resources needed Q.5 Consider X, Y, Z are shared
but still system may be in deadlock is? semaphores consider following three
concurrent processes:
Q.2 Let P and V be semaphore operations.
P represents wait and V represents
signal operation. Counting semaphore
variable S is initialized to 1 and no
blocked processes are present in the
system. If the following operations are
performed in the given order then what
is the value of S? If on running these processes
P, V, P, V, V, P, P, V, V, V, P, V, V, concurrently, possible outputs are:
V, P, P 1. CFBEAD 2. ACFDBE
3. ADCFBE
Q.3 Dijkstra’s banking algorithm in an What can be the initial value of X, Y &
operating system solves the problem of Z so that above outputs are possible?
a) Deadlock avoidance a) X = 2, Y = 0 and Z = 2
b) Deadlock recovery b) X = 1, Y = 0 and Z = 1
c) Mutual exclusion c) X = 2, Y = 0 and Z = 1
d) Context switching d) X = 1, Y = 0 and Z = 2

Q.4 Consider a system has 4 processes and Q.6 The necessary condition for Deadlock,
three resource types A, B, C. Resource which states that at least one resource
type A has 6 instances, B has 14 must be held in a non-sharable mode
instances and C has 3 instances. The a) Hold and wait
maximum requirement for each process b) Circular wait
and its current allocation are c) Preemption
d) Mutual exclusion

Q.7 A cycle in resource allocation graph is

a) necessary & sufficient condition for
multi-instance resource system
b) Not necessary and neither sufficient
condition for any system
c) Necessary but not sufficient
condition for system with multi
What is the safe sequence for the instance resource
processes so they don't enter into the d) Necessary but not sufficient for
deadlock state? system with single instance resource

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 31
 Q.8 Consider following statements with LEVEL 2
respect to interprocess communication:
S1: Message passing is useful for Q.1 A system contains 30 programs and
exchanging smaller amount of data. each requires 5 tape units for its
S2: Message passing require kernel operation. The minimum number of
support. tape units which the system must have
S3: Shared memory is faster than such that deadlocks never arise is ____.
message passing mechanisms.
S4: Shared memory is implemented Q.2 Which of the following is true for Lock
with the help of system calls and variable mechanism?
continuous requires system call support a) Mutual Exclusion is guaranteed
for exchanging data. Which of the b) Progress is guaranteed
statements are NOT TRUE? c) Bounded Waiting is achieved
a) S2 & S4 only b) S3 only d) None of this
c) S4 only d) S1 & S4 only
Q.3 Consider a non-negative counting
Q.9 What is the drawback of banker's semaphore S. The operation P(S)
algorithm? decrements S and V(S) increments S.
a) In advance process rarely knows that During an execution, 11 V(S)
how many resources that they need operations and 21 P(S) operations are
b) The number of processes changes as issued in some order. The largest initial
time progresses value of S for which at least one P(S)
c) Resource once available can operation will remain blocked is?
disappear
d) All of the above Q.4 Race condition in IPC environment
a) results when several threads try to
Q.10 Consider program for P1 and P2: read the same data concurrently
b) results when several threads try to
read & modify the same Data
concurrently
c) implies cooperation between non-
concurrent execution
d) outcome does not depend on
execution sequence

Q.5 Consider the following piece of code

executed by a set of concurrent
Here, m and n are binary semaphore
processes: (Assume mutex = 0)
variables whose values are initially
Signal (mutex);
initialized to 1. x and y are shared
<CS>
resources whose values are initialized
Wait (mutex);
to 0. Which of the following holds by
a) Mutual exclusion is always
above processes?
guaranteed
a) Deadlock and no mutual exclusion
b) Mutual exclusion is sometimes
b) Deadlock and race condition
guaranteed
c) No deadlocks and no race condition
d) Race condition and no deadlock

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 32
 c) Mutual exclusion is never x + +; y + +;
guaranteed wait(n); wait(m);
d) Progress is not guaranteed y + +; x + +;
signal(n); signal(m);
Q.6 Assume P1 and P2 are two concurrent signal(m); signal(n);
processes and that the following data } }
are shared between processes and Which of the following is true about
initialized as given for execution. above processes?
Semaphore m = 1 a) Race condition and deadlock both
Semaphore n = 1 occurs.
int x = 0 b) Deadlock but no race condition.
int y = 0 c) Race condition but no deadlock.
Here x and y are shared resources. d) No deadlock and no race condition.
Consider the following programs for
P1 and P2. Q.8 Consider the following resource
P1 ( ) P2( ) allocation graph:
{ {
wait(m); wait(n)
x + +; y + +;
wait(n); wait(m);
y + +; x + +;
signal(n); signal(m);
signal(m); signal(n);
} }
Which of the following is true about
above processes? Which of the following is true?
a) Race condition and deadlock both a) Graph represents unsafe state.
occurs. b) The system is safe and safety order
b) Deadlock but no race condition. is P2, P4, P1, and P3.
c) Race condition but no deadlock. c) The system is safe and safety order
d) No deadlock and no race condition. is P1, P2, P3 and P4.
d) The system is safe and safety order
Q.7 Assume P1 and P2 are two concurrent is P3, P2, P1 and P4.
processes and that the following data
are shared between processes and Q.9 In dining philosopher problem with 5
initialized as given for execution. philosophers and 6 chopsticks
Semaphore m = 1 a) Deadlock is possible but not
Semaphore n = 1 starvation
int x = 0 b) Neither Deadlock nor starvation is
int y = 0 possible
Here x and y are shared resources. c) No Deadlock but starvation is
Consider the following programs for possible
P1 and P2. d) Both Deadlock & starvation is
P1 ( ) P2( ) possible
{ {
wait(m); wait(n) Q.10 Consider following statements about
deadlock:

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 33
 S1: If a system is in unsafe state, the begin
process may complete its execution x=2
without entering a deadlock state. y=y+x
S2: If a process release all resources end
before request new resource, the begin
deadlocks and starvation both are y=3
possible. x=x+y
S3: Deadlock avoidance is less end
restrictive than deadlock prevention. parend
S4: In deadlock prevention, the request What can be the final value of ‘x’ and
for resources is always granted if the ‘y’ after computing the above
resulting state is safe. concurrent program.
Which of the statements are true? 1. x = 5, y = 3
a) S1, S2 and S3 2. x = 5, y = 8
b) S1 and S3 3. x = 2, y = 5
c) S3 and S4 4. x = 2, y = 8
d) S2, S3 and S4 Which of the above claims are
possible?
LEVEL 3 a) Only 1, 2, 3
b) Only 2, 3, 4
Q.1 Consider a system with 4 processes c) Only 1, 3, 4
that share 6 instances of the same d) All
resource type. Each process can request
a maximum of K instances. Resource Q.4 To provide synchronization for
instances can be requested and released producer - consumer problem with
only one at a time. The largest value of buffer size 1; minimum number of
K that will always avoid deadlock is? semaphores are required is? (Solution
should be using only semaphores)
Q.2 Consider the following statements:
S1: A thread running in critical section Q.5 Starvation problem can be solved in
may get context switched. priority based process scheduling
S2: Hardware access to devices is algorithms by
usually unavailable in user mode. a) Assigning static priorities to each
S3: Modification to the page tables are process
only possible in kernel mode. Which of b) Assigning dynamic priorities to each
the above statements are correct? process with priority of a process
a) S1, S2 directly proportional to waiting time of
b) S2, S3 the process
c) S1, S3 c) Assigning dynamic priorities to each
d) S1, S2, S3 process with priority of a process
inversely proportional to waiting time
Q.3 Consider the following concurrent of the process
program d) None of these
int x = 0
int y = 0
parbegin

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 34
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(6) (3) (a) (c) (a) (d) (c) (c) (d) (a)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(121) (b) (9) (b) (c) (b) (b) (b) (c) (b)

LEVEL 3
1 2 3 4 5
(2) (d) (a) (2) (c)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 35
 SOLUTIONS Now remaining resources = (3, 5, 1) +
(3, 5, 1) = (6, 10, 2)
Finally request for P0 can be
LEVEL 1
completed. So only possible safe
sequence is <P2 P3 P1 P0>
Q.1 (6)
P1 P2 P3
Q.5 (a)
1. If we take X= 1, Y= 1, then after
2R 3R 4R
process-1 or process-3 if both are
1R 2R 3R = 6R concurrently running, then both of the
So, Maximum number of resources that
processes will block and will stuck in
are needed but still deadlock occurs = deadlock situation.
6.
2. If we take X = 2, Z = 1, Y = 0 or X =
1, Y = 0, Z = 2 then only one of the
Q.2 (3) output above mentioned will print.
7 P operations, 9 V operations 3. An X=2, Y=0, Z=2, if either
S = 1 => S = 1 – 7 + 9 = 3. So, S=3
process-1 or process-3 starts, it will set
after the given operations. value to X= 1, Y=0, Z=1 and if either
process want to interrupt, then it will
Q.3 (a) set value to X = 0, Y = 0, Z = 0 and
Deadlocks avoidance problem in an
then output will be printed.
operating system can be solved by
Dijkstra's banking algorithm.
Q.6 (d)
By Definition of mutual exclusion.
Q.4 (c)
Need matrix for processes is calculated
Q.7 (c)
as:
A cycle in resource allocation graph is
necessary and sufficient for system
with single instance resource, but for
multi instance it is necessary but not
sufficient.

Q.8 (c)
System calls are required only to
establish shared memory region. After
establishing shared memory region, all
Total allocated resources are (A, B, C)
accesses are treated as routine memory
= (5, 12, 3)
access and no assistance from the
So, remaining resources are (1, 2, 0)
kernel is required.
Here, only request for P2 can be
completed.
Q.9 (d)
Now remaining resources = (1, 2, 0) +
(0, 1, 1) = (1, 3, 1)
Q.10 (a)
This satisfies only resources for P3
P1 first executes wait(m) i.e. m = 0; P2
Now remaining resources = (1, 3, 1) +
executes wait(n) then n = 0;
(2, 2, 0) = (3, 5, 1)
(3, 5, 1) satisfied the need for P1.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 36
 Now the process which will perform Signal (mutex) will never result in
wait( ) an any of the two semaphore blocking a process, when concurrent
will block. process access mutex.
Here P1 and P2 both again executes
wait( ) and goes to the block condition. Q.6 (b)
Hence deadlock occurs. No mutual No Race Condition: x can be accessed
exclusion occurs. by one process only at a time and same
as y. Both processes cannot access x or
LEVEL 2 y parallel. Hence mutual exclusions
holds; means No race condition.
Q.1 (121) Deadlock:
If 120 resources are there then it may P1( ) executes: wait(m);
be possible that all 30 process have 4 x + +;
resources and waiting for 1 more P2( ) executes: wait(n);
resource. If 121 resources are there, y + +;
then atleast one must have 5 resources Now both P1 and P2 are blocked.
so deadlock can never occur. Hence deadlock.

Q.2 (b) Q.7 (b)

For Lock variables, No Race Condition: x can be accessed
a) Mutual Exclusion fails. by one process only at a time and same
b) Progress is achieved. as y. Both processes cannot access x or
c) Bounded waiting fails. y parallel. Hence mutual exclusions
So, ans is b. holds; means No race condition.
Deadlock:
Q.3 (9) P1( ) executes: wait(m);
S = -21 + 11 = -10. If initial value of S x + +;
= 10 then 21 P(S) operations will P2( ) executes: wait(n);
succeed. As we want to block at least y + +;
one P(S) operation by keeping “S” the Now both P1 and P2 are blocked.
largest initial value. If we keep initial Hence deadlock.
value of S=9, then 20 P(S) operations
will succeed and only 21th P(S) Q.8 (b)
operation will get blocked. So, the Allocation and need table according to
largest initial value of S for which at given graph:
least one P(S) operation remains
blocked is 9.

Q.4 (b)
Race condition occurs when two or
more processes try to modify same data
item.

Q.5 (c)
P2, P4, P1 and P3 is safe sequence.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 37
 x=x + y //D
Q.9 (c) end
Deadlock never happens, but starvation For 1. ⇒A, B, C, D
to a particular philosopher may be 2. ⇒A, C, D, B
possible. 3. ⇒C, D, A, B

Q.10 (b) Q.4 (2)

2 semaphores are required,
Assume, name of the semaphores: S1,
S2
Initialization: S1 = 1, S2= 0

LEVEL 3 Q.5 (c)

According to option (c),
Whichever process has waiting time
Q.1 (2) highest, will be given more priority.
P1 P2 P3 P4 So, processes will not suffer from
| | | | starvation.
1 1 1 1
+1
So, a process can request for maximum
of 2 resources. More than 2 resource
request will lead to deadlock.

Q.2 (d)
Thread running in critical section may
get context switched, because, ‘locks’
(user-level) independent to the
scheduler and if thread running in the
kernel with interrupts disabled would
get context switched. S2 & S3 are also
correct statements.

Q.3 (a)
Consider
begin
x=2 //A
y=y+x //B
end
begin
y=3 //C

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 38
 Operating System

Topics: Memory Management, File System

-------------------------------------------------------------------------------------------
LEVEL 1 after servicing the number of request
________
Q.1 Interrupts are used for
a) Invoking system call Q.6 In order to change from privileged to
b) Indicating completion of I/O non-privileged:
c) Sometimes preempting running a) a hardware interrupt is needed
process from CPU b) a software interrupt is needed
d) All of the above c) a privileged instruction is needed
d) a non-privileged instruction is
Q.2 Limit information is provided in needed
segment table for every segment
a) To limit number of memory accesses Q.7 Consider a machine 32 MB physical
for a segment memory and 32 bit virtual address
b) To provide memory protection space. If the page size is 8 KB, what is
c) To limit the size of a segment to be the approximate size of the page table
more than a specific size in MB?
d) None of these
Q.8 Consider a disk system with 101(0 to
Q.3 Which of the following memory 100) cylinders. The requests to access
management technique suffers from cylinders occurs in sequence as: 15, 75,
fragmentation? 49, 65, 93, 80, 10
a) Paging If the head is currently at 60. What is
b) Segmentation the additional distance that will be
c) Contiguous traversed by head when C-SCAN
d) All of the above algorithm is used compared to SCAN
algorithm is _____ (Assume SCAN
Q.4 Consider a system uses 3 page frames algorithm head moves towards 100
for storing process pages in main when it starts execution).
memory. It uses the optimal
replacement algorithm policy. Assume Q.9 A computer system implements a 36
that all page frames are initially empty. bit virtual address. Page size of 4 KB
The total number of page fault that will and the size of physical memory is 30
occur while processing is 5, 6, 7, 8, 6, bits. The approximate size of page
5, 10, 7, 6, 8, 10, 7. table in the systemis___ MB. [Assume
memory is byte addressable]
Q.5 Consider a disk has 100 numbered
from 0 to 99. At some time the disk Q.10 The purpose of applying paging in
arm is at cylinder 51 and there is a segmentation is
queue of disk access request for a) To overcome thrashing
cylinders 16, 43, 46, 50, 55, 57, 73 and b) To overcome external fragmentation
83. If shortest seek time first (SSTF) is c) To reduce segment table size
being used for scheduling disk access, overhead
the request for cylinder 55 is serviced d) Both (B) & (C)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 39
 LEVEL 2 Q.5 A Disk of size 40MB uses Bit Map for
free space management with a disk
Q.1 Page size has impact on which of the block of size 4KB, the number of
following: blocks needed to store the Bit Map of
I. Internal fragmentation this disk is ______.
II. Page Table size
III. Page-fault rate Q.6 On a system using simple
a) All the three segmentation, compute the physical
b) Only I & II address for each of the logical address,
c) Only II given the following segment table
d) None of these

Q.2 Which of the following is NOT part of

UNIX I-Node structure?
a) File Name
b) Direct Block addresses
c) Link count
d) Type of file

Q.3 Match the following groups using the

disk operations Match the following addresses using
Group-I the segmentation
P. Seek Time
Q. Latency Time
R. Transfer Time
Group-II
1. Time taken to get the desired track 2.
Time taken to get the desired sector
3. Time to read the data from the disk
P Q R
a) 1 2 3
b) 2 1 3
c) 3 1 2
d) 3 2 1 a) (i)-T, (ii)-Q, (iii)-S, (iv)-R, (v)-P
b) (i)-Q, (ii)-R, (iii)-T, (iv)-S, (v)-P
Q.4 If an instruction takes x secs to execute c) (i)-R, (ii)-T, (iii)-S, (iv)-P, (v)-Q
without page fault and takes an d) (i)-R, (ii)-T, (iii)-S, (iv)-Q, (v)-P
additional 'z' secs with a page fault then
what page fault rate is needed to Q.7 Consider the following statements
achieve an effective instruction S1: The main reason to have multi-
execution time of 'k' secs. level page table is to speedup address
a) (k - x) / z translation
b) (k + x) / z S2: Using a multi-level page table
c) (k - z) / x increases TLB hit time
d) None of these

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 40
 S3: The main reason to have a Page table entries are 32 bits wide. The
hardware TLB is to speed up address page size in KB of the system should
translation be _____ in order to make sure that the
Which of the above statements are page table exactly fits in one frame of
correct? memory.
a) S1 and S2
b) S1 only Q.2 Consider the following heap in which
c) S3 only blank regions are not in use and shaded
d) S2 and S3 regions are in use.

Q.8 Which of the following operations

(actions) is typically not performed by
the OS when switching context from
process A to process B?
a) Saving current register values for The sequence of requests for block of
process A and restoring saved register size 200 KB, 65 KB and 110 KB are
values for process B. satisfied using best fit policy. Then the
b) Changing address translation tables internal fragmentation present after
c) Invalidating the TLB requests as _______ (in KB) (A fixed
d) None of the above partition scheme).

Q.9 Mapping of logical address to physical Q.3 Consider a logical address space of
address is done with the help of eight pages of 1024 words each
a) Page table mapped onto a physical memory of 32
b) Segment table frames. How many bits are there in the
c) Both (a) and (b) logical address?
d) Either (a) or (b) a) 13 b) 15
c) 12 d) 14
Q.10 Consider a 2-level paging system with
TLB support. The page table has Q.4 In a paging system, page size is 2048
divided into 2 K pages each of size 8 K bytes. A process might need 15 frames.
words. If the physical address space What is the maximum possible internal
has 32 M words which are divided into fragmentation size?
8K frame. TLB access time is 20 ns a) 2048 bytes
and main memory access time is 200 b) 1024 bytes
ns. The CPU finds 135 page references c) 2047 bytes
in the TLB out of total reference of d) No internal fragmentation
180. Then what is the effective
memory access time? Q.5 Consider linked allocation system, a
a) 315 ns b) 270 ns program has just read the 15th disk
c) 220 ns d) 320 ns block. After some point of time, if it
next want to use the 20th block, then
LEVEL 3 how many disk blocks must the
program access?
a) 5 b) 10
Q.1 A system supports 32 bit virtual
address, with a page size of power of 2. c) 20 d) 15

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 41
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(d) (b) (d) (7) (3) (d) (1) (59) (48) (d)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(a) (a) (a) (a) (1) (c) (c) (d) (d) (d)

LEVEL 3
1 2 3 4 5
(128) (125) (a) (c) (c)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 42
 SOLUTIONS 225
Number of frames = = 212
213
LEVEL 1 That's each page table entry will
contain 12 bits address of page frame
Q.1 (d) so 2 byte (byte-addressable memory)
Interrupts are useful for all of the Size of the page table = Number of
above. entries x each entry size
= 219. 21 = 220 = 1 MB
Q.2 (b)
Limit denotes number of words/bytes Q.8 (59)
in a segment, to limit a segment to SCAN:
access the memory allocated to other
segment which is protection of
memory.

Q.3 (d) Distance travelled in SCAN: 5 + 10 + 5

+ 13 + 7 + 51 + 34 + 5 = 130
Q.4 (7)
C-SCAN:
(i.e. when the head reaches other end, it
immediately returns to the beginning of
the disk, without servicing any requests
on the return trip).

Q.5 (3)

Distance travelled in C-SCAN: 5 + 10

+ 5+ 13 + 7 + 100 + 10 + 5 + 34 = 189
So, additional distance travelled is 189
So service 55 is serviced after servicing – 130 = 59
3 requests.
Q.9 (48)
Q.6 (d) Virtual memory = 36 bits = 236 B
Changing from privileged mode to Physical memory = 30 bits = 230 B
non-privileged mode doesn't require an #frames = 230 / 212 = 218 B
interrupt. Size of page table = # entries in page
table x page table entry size
Q.7 (1) = (236 B / 212 B ) x 18 bits
Number of entries in page table = = 16MB x 3B = 48MB
Virtual address space size 232
= 13 = 219 Q.10 (d)
Page Size 2
entries

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 43
 LEVEL 2
Q.7 (c)
Q.1 (a) S1: Multi-level page table is to save
All the above three depends on page space in memory.
size directly or indirectly. S2: TLB hit time is not affected by
page table structure
Q.2 (a) S3: TLB is used for cache address
File name is not a part of UNIX-I node. translations and hence it will speed up
the entire process.
Q.3 (a) So, S1 and S2 are false.
Seek time: Time taken to get the track. S3 is true.
Latency Time: Time taken to get
desired sector. Q.8 (d)
Transfer time: Time to read data. All actions are performed during
context switching.
Q.4 (a)
If page fault rate is = p then, Effective Q.9 (d)
access time In paging → page table
k = p(x + z) + (1 - p)x In segmentation → segment table
k = px + pz + x – px For particular implementation either of
p = (k - x) / z these two is used.

Q.5 (1) Q.10 (d)

Number of disk blocks = 40MB/4KB = TLB hit ratio (x) = 135/180 = 0.75
10K Effective memory access time = x(C +
Size of bitmap = 10K bits = M) + (1 - x) (C + 3M)
1.25KBytes = 0.75 (20 + 200) + (1 - 0.75) (20 +
For 1.25KB bitmap 1 disk block is 3(200)) = 165 + 155 = 320ns
required.
LEVEL 3
Q.6 (c)
(i) For segment 0, length = 124 > Q.1 (128)
logical address 99 Suppose the page size is 2x Bytes
.•. Physical address = 330+99 = 429 Number of pages = 232 / 2x = 232 – x
(ii) For segment 2, length = 99 > Page table size =
logical address 78 number of pages * 1 entry size
.•. Physical address = 111+78 = 189 = 232 – x * 4Bytes
(iii) For segment 1, length = 211 < = 234 – x
logical address 265 As given page table should fit in one
.•. Trap page, hence
(iv) For segment 3, length = 302 > page size = page table size
logical address 222 2x = 234-x
.•. Physical address = 498+222 = 720 x=17
(v) For segment 0, length = 124 > Page size = 217 = 128K
logical address 111
.•. Physical address = 330+111 = 441 Q.2 (125)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 44
 So, internal fragmentation is [50 + 35 +
40] KB = 125 KB

Q.3 (a)
Logical address space as:

If logical address space in m bits, we

have n bit for offset and m – n bits for
pages.
Here, page size = 1024 words = 210 (10
bit for offset)
Number of pages = 8 = 23 (3 bit for
page number)
So, logical address space requires 13
bits.

Q.4 (c)
In the worst case, a process would need
n pages plus 1 byte. Then it would be
allocated n + 1 frames i.e. in the last
frame, only 1 byte is used resulting in
2048-1, 2047 bytes of internal
fragmentation.

Q.5 (c)
Since in linked allocation, to find the ith
block, we must start at the beginning of
that file and follow the pointer until we
get to the ith block. No random access
is possible. So to read the 20th block, it
has to come from first again.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 45
 Database Management System

Topics: ER-model, Relational model

-------------------------------------------------------------------------------------------
LEVEL 1 Q.7 An instance of relational schema R (A,
B, C) has distinct values of A including
NULL values. Which one of the
Q.1 E-R model uses ___ symbol to
following is true?
represent weak entity set?
a) A is a candidate key
a) Dotted rectangle
b) A is not a candidate key
b) Diamond
c) A is a primary Key
c) Doubly outlined rectangle
d) Both (a) and (c)
d) None of these
Q.8 A table joined with itself is called
Q.2 Relational Algebra is
a) Join
a) Data Definition Language
b) Self Join
b) Meta Language
c) Outer Join
c) Procedural query Language
d) Equi-Join
d) None of the above
Q.9 In context with relational Algebra,
Q.3 Key to represent relationship between
Which of the following are unary
tables is called
operations?
a) Primary key
1. Select
b) Foreign Key
2. Project
c) Secondary Key
3. Union
d) None of these
4. Product
a) 1 and 3 only
Q.4 What is a relationship called when it is
b) 2 and 4 only
maintained between three entities?
c) 1 and 2 only
a) Unary
d) All are binary
b) Binary
c) Ternary
Q.10 The column of a table is referred to as
d) Quaternary
the
a) Tuple
Q.5 Which of the following operation is
b) Attribute
used if we want to select only certain
c) Entity
columns of a table?
d) Degree
a) PROJECTION
b) SELECTION
c) UNION LEVEL 2
d) JOIN
Q.1 Which of the following statements
Q.6 NULL is about ER models is/are correct?
a) The same as 0 for integer I. Many-many relationships cannot be
b) The same as blank for character represented in ERD.
c) The same as 0 for integer and blank II. Relationship sets can have attributes
for character of their own.
d) Not a value III. All many to one relationships are
represented by the relationships

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 46
 between a weak & a non weak entity S (C B) C primary key and attribute
set. “B” of relation “R” is foreign key
a) II only references to S and no NULL values in
b) III only both relations and both R and S
c) II, III only consists more than two records. Which
d) I and II only of the following statement is false?
a) ΠB (R) – ΠC (S) always empty.
Q.2 Consider the following relations: b) Πc (S) – ΠB (R) may not empty.
R1 (PQR) and R2 (PST) c) ΠB (R⋈R.B=S.c S) – Πc (S) always
R1 has 500 records and R2 has 800 empty.
records. The non-Null attribute ‘P’ in d) ΠB (R⋈R.B≠S.C S) – Πc (S) may not
R2 is referencing attribute ‘P’ in R1. empty.
The maximum number of records
present in R1 ⋈ R2 is _____. Q.6 Which of the following statements
is/are true?
Q.3 The following table has two attributes S1: In relational algebra σ (selection)
A and C where A is the primary key operation is NOT commutative.
and C is the foreign key referencing A S2: In relational algebra Π (projection)
with on-delete cascade. operation is commutative.
a) Only S1
b) Only S2
c) Both S1 and S2
d) Neither S1 nor S2

Q.7 In Railway reservation system, the

entities are date, Train number, place
of departure, Arrival, coach no. and
seat no. The primary key is
The number of tupples that must be a) Train number
additionally deleted to preserve b) Train number + place of departure
referential integrity when the tupple c) Train number + date
(4, 3) is deleted is? d) All entities

Q.4 Consider the following ER diagram: Q.8 Processor and I/O Ports is an example
of
a) One-to-one relationship
b) One-to-many relationship
c) Many-to-one relationship
d) Many-to-many relationship

The minimum number of relational Q.9 Let R(a, b, c) and S(d, e, f) be two
table required for above ER diagram is relations in which d is the foreign key
_______ of S that refers to the primary key of R.
Consider the following four operations.
Q.5 Consider the following relations: (1) Insert into R
R (A B) A primary key (2) Insert into S

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 47
 (3) Delete from R
(4) Delete from S
Which of the following is true about
the referential integrity constraint
above?
a) None of them can cause any If ‘n’ entries in E1 and ‘m’ entries in
violation E2. How many entries in relationship
b) All of the can cause violation set (R)?
c) Operations (1) and (2) can cause a) At least n
violation b) At most n
d) Operations (2) and (3) can cause c) Exactly n
violation d) At least n and At most m
Q.10 Find out the degree of the relationship. Q.3 Consider the following relation:
A student uses exactly one notebook Supplier (Sid, Sname, Address)
for each project. Each notebook Parts (Pid, Pname, Color)
belongs to one student for each project. Catalog (Sid, Pid, Cost)
Note that a student may still work on Retrieve pairs of Sids such that the
many projects and maintain different supplier with the first Sid charges more
notebooks for different projects. for some part than the supplier with the
a) Unary second Sid.
b) Binary Q1: πsid,s (Catalog
c) Ternary ⋈ (Sid≠S ∧ Pid=P ∧ cost>c) ρS, P, C (Catalog))
d) Two binary relations Q2: πsid,s (Catalog
⋈ (Sid ≠ S ∧ Cost > S) ρS, P, C (Catalog))
LEVEL 3
Which of the following is correct about
Q.1 Consider the following relations: above queries?
R(A B) A is primary key a) Q1 correct but not Q2
S(C D E) C is primary key and D is b) Q2 correct but not Q1
alternative key. An attribute “B” of c) Q1 and Q2 both correct
relation R is foreign key references to d) Q1 and Q2 both false
alternative key D of relation S. Which
of the following statement is true? Q.4 Which is the optimized version of the
a) Every record of R must references to relational algebra expression?
some record of S. ПA1 (ПA2 (σF1 (σF2 (r)))
b) Each record of R can references to at Where A1, A2 are sets of attributes in r
most one record of S. with A1< A2 and F1, F2 are
c) Each record of R can reference zero Boolean expression based on attributes
or more record of S. in r.
d) Each record of R must reference one a) ПA1 (σF1⋂F2 (r))
or more record of S. b) ПA1 (σF1⋃F2 (r))
c) ПA2 (σF1⋂F2 (r))
Q.2 Consider the following ER model: d) ПA2 (σF1⋃F2 (r))

Q.5 Consider the following relations:

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 48
 How many tuples will be in the result
of the following relational algebra
expression?
(STUDENT ⨝AGE ≥ 25 ⋂ STUDENT.SID =
RESERVE.SID RESERVES ⨝
GATEEXAM)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 49
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(c) (c) (b) (c) (a) (d) (b) (b) (c) (b)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(a) (800) (6) (4) (d) (d) (d) (d) (d) (c)

LEVEL 3

1 2 3 4 5
(b) (c) (a) (a) (4)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 50
 SOLUTIONS LEVEL 2
Q.1 (a)
LEVEL 1
Q.1 (c)

Q.2 (c) So, ER-diagram represents many

Relational Algebra is Procedural query students can enroll many courses.
Language • In ER-diagram Relationship set can
have own attribute.
Q.3 (b) • Many to one relation can be
Foreign key is used to represent represented by the relation between
relation between two tables. non-weak entity set.

Q.4 (c) Q.2 (800)

Relationship between three entities is Since ‘P’ in R2 is not key, hence all
called Ternary relation. value of ‘P’ may or may not be unique.
Hence every entry under ‘P’ in R2 will
Q.5 (a) match with ‘P’ in R1. Hence maximum
PROJECTION extracts only certain is 800.
columns which we want to select.
Q.3 (6)
Q.6 (d) If (4, 3) is deleted then 4 is the primary
Null cannot be represented by any key but in (2, 4), (6, 4) and (3, 4), 4 is
value. the foreign key so these must be
deleted.
Q.7 (b) The primary key for (2, 4), (3, 4) and
Candidate key should be not null and (6, 4) is 2, 3 and 6 respectively must be
unique. deleted but in (5, 2) and (7, 2) with
primary key 5 and 7 is the foreign key
Q.8 (b) also deleted and (9, 5) is also deleted
because 5 is primary key which is
Q.9 (c) already deleted.
Select and Project operations are
applied only on one operand. Whereas, Q.4 (4)
Union and product can be applied on E1R1 (ABCD)
more than one operand. R2 (AD)
E2 (DEF)
Q.10 (b) E3R3 (X D Y Z) candidate key
The row of a table is also known as So, minimum 4 relational tables are
tuple and the column is known as required.
attribute.
Q.5 (d)
ΠB (R⋈R.B≠S.C S) which selects all
foreign key values.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 51
 ΠB (R⋈R.B≠S.C S) – Πc (S) is always Remaining records of R references to
empty. at most one record of S.
All foreign keys values – All primary
key values = Always empty

Q.6 (d)
Both the statements are incorrect.
The select operation is commutative
i.e. σc1 (σc2 (R)) ↔ σc2 (σc1 (R)).
Ultimately only those tupples will be
selected which satisfy both C1 and C2.
Hence order does not matter.
But Π (projection) operation is not Q.2 (c)
Here, E1 has Key + Participation
commutative.
Π a1 (Π a2 (R)) ↔ Π a2 (Π a1 (R)) if and Constraint. So E1 must relate with
exactly one entry of E2.
only if a1 is substring (or subset) of a2,
otherwise operation would be E2 has Only Key Constraint. So E2
incorrect. relate with at most one entry of E2.
E2 has more or equal entries in relation
(table of E2).
Q.7 (d)
With any proper subset of the entity set So, exactly n entries in relationship set
(R).
we cannot uniquely identify any
passenger. Thus all entities are required
to form primary key. Q.3 (a)
Queries 1 will return pairs of Sids such
that the supplier with the first Sid
Q.8 (d)
Processor can have more than one I/O charges more for some part than the
ports and I/O process can be assigned supplier with the second Sid. Which is
to more than one processor in multi- possible by the condition that is when
core processor system. two supplier id is different but part id is
same and charges of first supplier is
more than second supplier.
Q.9 (d)
Insertion into child table and deletion Queries 2 will return empty set because
from parent table can cause violation. we compare cost with Sid which
always returns empty set.
Q.10 (c)
Degree of relationship is ternary Q.4 (a)
because relationship will keep three Since, A1 < A2, final relation R
entity students, notebook and project. displaying values for attributes will be
present in set A1.

LEVEL 3 Q.5 (4)

The first join produces 5 tuples.
Q.1 (b) Second join forms 4 tuples.
Null value of B column record not So, 4 tuples in the result.
references to any record of S.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 52
 Database Management System

Topics: Normal Forms, SQL

-------------------------------------------------------------------------------------------
Q.5 Why do we have to create a primary
index on a primary key
LEVEL 1 a) Ease of searching
b) Sequential ordering of a file
c) Both (a) and (b)
Q.1 Consider the relation R (A, B, C, D, E,
d) None of these
F) holds following FD's
AB → C
Q.6 A functional dependency of the form
C→D
X → Y is trivial if
D→E
a) Y ⊆ X
E→F
b) Y ⊂ X
F→B
c) X ⊆ Y
What will be the total number of
d) X ⊂ Y and Y⊂ X
candidate keys presents in above
relation?
Q.7 In SQL, which command is / are used
a) 2
to remove rows from a table.
b) 3
a) DELETE
c) 4
b) REMOVE
d) 5
c) TRUNCATE
d) Both (a) & (c)
Q.2 If a relation is in BCNF, it is also in:
a) 1NF
Q.8 In SQL, which of the following is not a
b) 2NF
data definition language commands?
c) 3NF
a) DROP
d) All of the above
b) RENAME
c) GRANT
Q.3 A DBMS query language is designed
d) CREATE
to
a) Support end users who use English-
Q.9 Which dependency is not followed
like commands
from the following table?
b) Support in the development of
complex applications software
ENAME PNAME DNAME
c) Specify the structure of a database.
d) All of the above. Smith X John
Smith X Anna
Q.4 A primary key is combined with a Aman Z Anna
foreign key creates Aagam Y John
a) Parent-Child relationship between
the tables that connect them a) ENAME → PNAME
b) Many to many relationship between b) ENAME → DNAME
the tables that connect them c) PNAME → ENAME
c) Network model between the tables d) Both (b) and (c)
that connect them
d) None of the above Q.10 A table T1 in a relation database has
following rows and columns.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 53
 Employee ID Salary How many records resulted by above
101 10k query on given instant of relation R?
202 25k a) 1
333 17k b) 2
454 NULL c) 3
The following sequence of SQL d) 4
statements was successfully executed
in table T1. Q.4 Given relation R(P, Q, R, S, T) and set
Update T1 set salary = salary + 5k; of functional dependencies
select avg(marks) from T1; F = {PQ → R, PQ → S, S → P, QR →
What is the output of the select S, QR → T} The highest normal form
statement? satisfied by R is
a) 18 a) 1NF
b) 16.75 b) 2NF
c) 25 c) 3 NF
d) NULL d) BCNF

Q.5 Consider A(P, Q, R, S, T, V, W) and

LEVEL 2 the following FD's:
W → VS
Q.1 Consider the Two sets of FD's for the T→S
relation R(ABCDE): WS → RT
F1 = {A → B, AB → C, D → AC, QS → P
D → E) Which of the following is minimal
F2 = {A → BC, D → AE) cover of the given FD’s?
Which of the following statement true a) {W → V, T → S, W → R, WS → T,
about FD set? QS → P}
a) F1 ⸦ F2 b) {W → V, W → S, T → S, W → R,
b) F1 ⸧ F2 QS → P}
c) F1 ≡ F2 c) {W → V, T → S, W → R, WS → R,
d) None of these QS → T}
d) {W → V, T → S, W → R, W → T,
Q.2 Consider the following relation R(A, QS → P}
B, C, D, E, F, G, H). The maximum
number of super keys possible if 4 Q.6 Which of the following statement is
simple candidate keys in relation R are false of relation R is in 3NF but not
__________ BCNF?
a) Relation R must consist at least two
Q.3 Consider the following relation SQL over-lapped candidate keys.
query: b) Relation R must consist proper
subset of candidate key determines
proper subset of some other candidate
key.
c) Relation R must consist at most one
compound candidate key and other

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 54
 candidate keys are simple candidate
key.
d) Relation R must consist at least two
compound candidate keys.

Q.7 Given the function dependencies

A → Z; A → B; B → C and C → PQ
Which of the following does not hold?
a) A → C
b) C → Z
c) A → BZ
d) None of the above

Q.8 Let R (A, B, C, D, E, P, G) be a

relational schema in which the
following functional dependencies are
known to hold: Roll_no is the Primary Key of student
AB → CD, table
DE → P, Dept_id is the Primary Key of
C → E, Department table
P → C and B → G. Student.Dept_id is a Foreign Key
The relational schema R is refers Department.Dept_id
a) In BCNF What will happen if we try to execute
b) In 3NF, but not in BCNF the following 2 SQL statements?
c) In 2NF, but not in 3NF (i) Update Student set Dept_id =
d) Not in 2NF NULL where Roll_no = 1;
(ii) Update Department set Dept_id =
Q.9 Consider the two relations student NULL where Dept_id = 1;
(name, rollno, deptno, rank) and course a) Both (i) and (ii) will fail
(deptno, rollno, deptname, teacher, b) (i) will fail but (ii) will succeed
marks). Which of the following queries c) (i) will succeed but (ii) will fail
cannot be expressed using the basic d) Both (i) and (ii) will succeed
relational algebra operations (U, –, ×,
π, σ, ρ)?
a) Rank of every student in a LEVEL 3
department
b) Students whose name is the same as Q.1 Consider following 2 queries with
their department name relation schema:
c) The sum of all marks for a particular Student (ID, name, DEPT)
course Courses (CID, ID, Grade, DEPT)
d) All students of a given department Q1: Select * From Student S Where
not exists (select * from Course C
Q.10 Consider the two tables in a relational Where C.ID = S.ID and C.Grade > 4.0)
database with columns and rows as Q2: (πid (Student) – πid (σGrade > 4.0
follows: (Course))) ⋈ Student.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 55
 Which of the following query results {AB → C, BC → A, AC → B, B → D,
“all students who never got a grade D → E}.
above 4.0”? The minimum relations required to
a) Only Q1 decompose R into BCNF which satisfy
b) Only Q2 lossless join and dependency
c) Both Q1 and Q2 preserving decomposition is ______.
d) None of them
Q.5 What is the output of following SQL
Q.2 What does the following SQL query query?
find if applied on relation Select * from student S where not
emp (ssn, name, salary); exists (select * from course C
SELECT MAX (E1.salary) FROM where C.ID = S.ID and C.MARKS >
emp E1, emp E2 WHERE 60)
E1.salary < E2.salary; a) Displays all students who never got
a) It finds the highest salary available marks above 60
in the given relation. b) Displays all students who never got
b) It finds the second highest salary marks less than equal to 60
available in the given relation. c) Displays all students who got marks
c) It finds the lowest salary available in above 60
the given relation. d) Displays all students who got marks
d) It finds the second lowest salary less than 60
available in the given relation

Q.3 There is a database of the dealers of a

Products_deal. Following is the
snapshot.

Consider the following query:

SELECT A.Color-id, A.Dealer-no
FROM
Product_deal A, Product_deal B
WHERE A.Dealer-no = B.Dealer-no
and A.Part-no < > B.Part-no;

The number of tupples contained in the

output will be _____.

Q.4 Consider the following relational

schema R(ABCDEFG) with FD set

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 56
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(d) (d) (d) (a) (c) (a) (d) (c) (d) (b)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(c) (240) (b) (c) (d) (c) (b) (d) (c) (c)

LEVEL 3

1 2 3 4 5
(c) (b) (4) (4) (a)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 57
 SOLUTIONS Q.9 (d)
Both (b) and (c) cannot be derived.
Because, for one value Smith there are
LEVEL 1 two different DNAME. Same with
PNAME.
Q.1 (d)
Q.10 (b)
(AB)+ = ABCDEF Average = (15 + 30 + 22 + null) / 4
(AC)+ = ACDEFB
= 16.75
(AD)+ = ADEFBC
(AF)+ = AFBCDE
(AE)+ = AEFBCD LEVEL 2
There are 5 candidate keys are present
for given relation. Q.1 (c)
F1 covers F2: True
Q.2 (d) F2 covers F1: True
BCNF satisfied all the conditions that
are followed from 1NF and 2NF and is Q.2 (240)
more strict than 3NF. If there are m attributes in relation and
n simple candidate keys then maximum
Q.3 (d) number of super keys possible
= [2n – 1] × 2m-n
Q.4 (a) = [24-1] × 24
Foreign key is used to connect two = 15 × 16 = 240
tables with parent-child relationship
between them. Q.3 (b)
Given query will result tupple where A
Q.5 (c) ≥ 5 (i.e. A =5, 6, 10, 8) and B ≥ C (i.e.
Main objective of primary key is to B = 8 because we cannot compare Null
search data easily and by indexing it values) or C ≥ 5 (i.e. C= 5, 6, 5 because
we can order the data in sequential we cannot compare Null values) so the
manner. resulting tupples are 2 i.e. A=6, A=8.

Q.6 (a) Q.4 (c)

If Y ⊆ X, then X → Y is trivial. Candidate keys for the relation are: PQ,
QS and QR
Q.7 (d) S → P, prime attribute → prime
The DELETE command is used to attribute (not allowed in BCNF but
delete rows from a table. allowed in 3NF).
The TRUNCATE command is used to ⇒ Relation R is in 3NF but not in
delete all the rows from the table and BCNF since S → P does not have a
free the space containing the table. super key on the left hand side.

Q.8 (c) Q.5 (d)

CREATE, ALTER, DROP, RENAME, Checking QS → P, Q+ = Q, S+ = S,
and TRUNCATE are DDL commands. Hence QS → P is essential.
Checking WS → R, WS → T

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 58
 W+ → WVSRT, Hence it can be second statement will not execute,
decomposed to W → R, W → T because primary key cannot be
So, the dependencies remained are NULL.
W → V, W → S, T → S, W → R, W
→ T, QS → P LEVEL 3
Now, {W → T, T → S} by transitive
rule W → S can be obtained.
Q.1 (c)
Hence minimal cover is: W → V, T →
Both query results students who never
S, W → R, W → T, QS → P.
got a grade above 4.0 but 1st query also
give those students who never enrolled
Q.6 (c)
any course.
If relation R is in 3NF but not BCNF
then at least two compound keys must
Q.2 (b)
exist where non-trivial FD with
Condition E1.salary < E2.salary; will
determinant is not super key.
find all salaries accept highest.
Aggregate function MAX(E1.salary)
Q.7 (b)
will find maximum of all salaries
(a) and (c) can be derived by given
which are less then highest salaries i.e.
FDs. But (b) cannot be derived by any
second highest salary.
FD.
Q.3 (4)
Q.8 (d)
The output Table will be
It follows partial dependency, so it is
not in 2 NF.

Q.9 (c)
The six basic operators of relational
algebra are the selection (σ), projection
(π), Cartesian product (×), Set union
(U), Set difference (–), and Rename
(ρ). Many other operators have been Q.4 (4)
defined in terms of these six. Among Given relation is in 1NF.
the most important are set intersection, For 2NF:
division, and the natural join, but
aggregation is not possible with these R1(ABCDFG)
basic relational algebra operations. So, AB → C
we cannot run sum of all marks with BC → A
the six operations. AC → B

Q.10 (c) R2(BDE)

First query will successfully execute B→D
and after execution attribute D→E
Dept_id of Student table with Roll_no
= 1 becomes “NULL” which is foreign For 3NF:
key from Department.Dept_id. Now,
R1(ABCFG)
AB → C

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 59
 BC → A
AC → B

R2(BD)
B→D

R3(DE)
D→E

For BCNF(Lossless join

Dependency
Preserving)

R1(ABC) R2(ABFG)
AB → C
BC → A
AC → B

R2(BD)
B→D

R3(DE)
D→E

Q.5 (a)
Inner query results all students who got
marks above 60. Now Not exist will
display all students who never got
marks above 60.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 60
 Database Management System

Topics: File Organization, Indexing, Transactions & Concurrency Control

-------------------------------------------------------------------------------------------
LEVEL 1 Q.6 What are the potential problems when
a DBMS executes multiple transactions
concurrently?
Q.1 A File is organized so that the ordering a) Phantom problem
of data record is same as or close to the b) The dirty read problem
ordering of data block in some index. c) Unrepeatable read problem
Then that index is called d) All of the above
a) Dense
b) Sparse Q.7 Which among the following 2-phase
c) Clustered locking protocols is a deadlock free?
d) Unclustered a) Basic 2PL
b) Strict 2PL
Q.2 Assume transaction A holds a shared c) Rigorous 2PL
lock R. If transaction B also requests d) Conservative 2PL
for a shared lock on R, it will
a) Result in a deadlock situation Q.8 A clustering index is defined on the
b) Immediately be granted fields which are of type
c) Immediately be rejected a) Non-key and ordering
d) B granted as soon as it is released by b) Non-key and non-ordering
A c) Key and ordering
d) Key and non-ordering
Q.3 The index consists of
a) A list of keys Q.9 The timestamp schedule which accepts
b) Pointers to the master list view serializable schedule is:
c) Both (a) and (b) a) Basic timestamp
d) All of the above b) Thomas write rule
c) Both (a) and (b)
Q.4 Isolation is managed by d) None
a) Transaction management component
b) Recovery management component Q.10 Match List-I with List-II and select the
c) Concurrency control component correct answer using the codes given
d) None below the lists:
List-I
Q.5 The following property makes sure that A. Primary index
more than one transaction are being B. Clustered index
executed simultaneously and as if it is C. Secondary key index
the only transaction in the system. D. Secondary non key index
a) Atomicity
b) Consistency List-II
c) Isolation 1. Dense
d) Durability 2. Sparse

a) A-2, B-1, C-1, D-2

b) A-1, B-1, C-2, D-1

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 61
 c) A-1, B-2, C-2, D-1 which are working on two data items
d) A-2, B-2, C-1, D-1 namely A and B:
S: R4 (A), R2 (A), R3 (A), W1 (B),
W2 (A), R3 (B), W2 (B)
LEVEL 2 Which of the following is correct?
a) The schedule cannot be serialized
Q.1 Which of the following claims are
b) The schedule is equivalent to T3,
correct about strict 2PL? T4, T1, T2
I. Guaranteed conflict serializability
c) The schedule is equivalent to T1,
II. Guaranteed strict recoverable T4, T3, T2
III. Guaranteed deadlock avoidance
d) The schedule is equivalent to T2,
a) I, II only T3, T1, T4
b) I only
c) I, II, III Q.5 Consider the following schedules:
d) II, III only
S1: r1(x), r2(z), r1(z), r3(x), r3(y),
w1(x), c1, w3(y), c3, r2(y), w2(y),
Q.2 Suppose the process P has been
w2(z), c2
running for several days when a new S2: r1(x), r2(z), r1(z), r3(x), r3(y),
process Q starts up and begins
w1(x), w3(y), r2(y), w2(z), w2(y), c1,
contending with P for resources. Which c2, c3
of the following is true?
S3: r1(x), r2(z), r3(x), r1(z), r2(y),
a) In a wait-die system, if P needs a r3(y), w1(x), c1, w2(z), w3(y), w2(y),
resource held by Q, then P waits.
c3, c2
b) In a wait-die system, if Q needs a The number of above schedules
resource held by P, then Q waits.
recoverable is _______.
c) In a wound-wait system, if P needs a
resource held by Q, then Q yields and
Q.6 Consider the following schedules:
waits.
d) In a wound-wait system, if Q needs
a resource held by P, then Q dies.

Q.3 Which of the following statements

is/are true?
S1: 2-phase locking protocol ensures
both conflict serializability and
deadlock avoidance.
S2: Time-stamp ordering protocol do
Which of the following is correct about
not ensure both conflict serializability
above schedule?
and deadlock avoidance.
a) Only view serializable
a) Only S1
b) Only conflict serializable
b) Only S2
c) Both conflict and view serializable
c) Both S1 and S2
d) Neither view serializable nor
d) Neither S1 nor S2
conflict serializable
Q.4 Consider the following schedule S on 4
Q.7 Consider the following schedule:
transactions i.e. T1, T2, T3 and T4

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 62
 S: R2(A) W1(B) W1(C) R3(B) R2(B)
R1(A) C1 R2(C) C2 W3(A) C3 Q.10 Which of the following is true about
Which of the following is correct about the given schedule ‘S’?
above schedule?
a) Schedule(S) is not conflict
serializable schedule.
b) Schedule(S) is allowed by 2PL.
c) Schedule(S) is strict recoverable
schedule.
d) Schedule(S) is allowed by strict
2PL.

Q.8 Which of the following statements are

true for B tree and B+ tree index?
a) B tree index are faster for range
queries compare to B+ tree index.
b) If disk block allocated for B+ tree
index and same size disk block
allocated for B tree index then number
of index blocks and I/O cost of B+ tree
index is less than or equal to B tree
index for given distinct keys.
c) If disk block allocated for B+ tree
index and same size disk block
allocated for B tree index, then B tree
index access cost is less than or equal a) It is conflict serializable
to B+ tree index for given distinct keys. b) It is view serializable but not
d) If number of keys that can be stored conflict serializable
in B tree and B+ tree index is same, c) It is conflict serializable but not
then I/O cost of B+ tree index is less view serializable
than or equal to I/O cost of B tree d) It is not serializable
index for random access of same key
from set of distinct keys.
LEVEL 3
Q.9 Consider a table T in a relational
database with a key field K. A B-tree Q.1 Consider the following schedule:
of order p is used as an access structure S: r1(A), r3(D), w1(B), r2(B), r4(B),
on K, where p denotes the maximum w2(C), r5(C), w4(E), r5(E), w5(B)
number of tree pointers in a B-tree The number of serial schedules which
index node. Assume that K is 12 bytes are view equal to schedule (S) ____.
long; disk block size is 1024 bytes;
each data pointer PD is 10 bytes long Q.2 Consider the following modified 2
and each block pointer PB is 7 bytes phase locking protocol: Before a
long in order for each B tree node to fit transaction T writes a data object A, T
in a single disk block, then maximum has to obtain an exclusive lock on A.
value of p is ____. Before a transaction T reads a data
object A, T has to obtain a shared lock

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 63
 on A. If exclusive locks are held until
the end of transaction and shared locks
can be released at any time then which
of the following properties are
guaranteed?
a) Conflict serializability
b) Recoverability
c) Avoids cascading rollbacks
d) All of the above
The minimum number of key
Q.3 A database relation has 5000 records insertions that causes a new level to be
block can hold either 10 records or 15 introduced in the above B+ tree _____.
keys and pointer pairs. If sparse index (Assume key redistribution is not
is used at 1st level and multilevel allowed)
indexing is used in system, then the
number of disk block required to store
relation and index is ________

Q.4 Consider a file of 8192 records. Each

record is 16 bytes long and its key field
is of size 6 bytes. The file is ordered on
a key field, and the file organization is
unspanned. The file is stored in a file
system with block size 512 bytes, and
the size of a block pointer is 10 bytes.
If the primary index is built on the key
field of the file, and a multilevel index
scheme is used to store the primary
index, the number of first-level and
second-level blocks in the multilevel
index are respectively
a) 8 and 1
b) 16 and 1
c) 8 and 2
d) None of these

Q.5 Consider the following B+ tree with the

order of internal and leaf nodes as 3
and 2 respectively:

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 64
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(b) (b) (b) (c) (c) (d) (d) (a) (c) (d)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(a) (a) (d) (c) (2) (d) (b) (b) (36) (d)

LEVEL 3
1 2 3 4 5
(10) (d) (538) (a) (3)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 65
 Primary index and clustered index are
SOLUTIONS sparse. Secondary key and non-key
index are dense.
LEVEL 1
LEVEL 2
Q.1 (b)
Q.1 (a)
In sparse index number of entries
Every 2PL is conflict serializable, and
depends on number of data block in
every strict 2PL is always guaranteed
previous level.
to be strict recoverable but not
In dense index number of entries
deadlock free.
depends on number of data record in
previous level.
Q.2 (a)
In wait-die scheme, when transaction
Q.2 (b)
Ti request a data items currently held
Shared locks can be shared by different
by Tj, Ti is allowed to wait only if it
transactions simultaneously.
has a time stamp smaller than that of Tj
otherwise Ti is rolled back (die). Here
Q.3 (b)
process P is running so it has time
stamp less than process Q now if
Q.4 (c)
process P need a resource held by
Isolation is managed by concurrency
process Q then process P has to wait.
control component.
Q.3 (d)
Q.5 (c)
2PL ensures conflict serializability but
Isolation ensures that no transaction
may lead to deadlock.
will affect the existence of any other
Time-stamp concurrency control
transaction.
algorithm is a non-lock concurrency
control method. In time-stamp based
Q.6 (d)
method, deadlock cannot occur as no
The above all and lost update problem
transaction ever waits.
are all effects of executing multiple
transactions concurrently. Q.4 (c)
The precedence graph of the given
Q.7 (d) schedule is
Conservative 2PL is deadlock free.

Q.8 (a)

Q.9 (c)
Both Basic timestamp and Thomas
write rule timestamp schedule accepts
view serializable schedule. Therefore schedule is equivalent to
(T1, T4, T3, T2), (T1, T3, T4, T2) and
Q.10 (d) (T4, T1, T3, T2).

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 66
 Q.5 (2)
In S2: w3(y) is first and r2(y) appears
second
Hence, c2 should appear after c3.
In recoverable [If Ti reads a value
written by Tj, then Ti must commit
after Tj commits] Schedules S1 and S3
are recoverable.

Q.6 (d)
Checking for conflict serializable:

Cycle is present so not conflict

serialiazable. Q.8 (b)
Since there is blind write between If disk block allocated for B+ tree index
W2 (a) to W3 (a), so it may be view and same size disk block allocated for
serializable.
B tree index then number of index
Checking for view serializability: blocks and I/O cost of B+ tree index is
1. Final write: a = T1, b = T3 ⋯ (1) less than or equal to B tree index for
2. Initial read: a = T1, T2 b = T3, c = T1 given distinct keys.
⋯ (2)
3. Write read: No write read Q.9 (36)
(T2, T3) → T1 from (1) Typical structure of a B-tree non-leaf
T1 → T2, T3 from (2) node is follows:
Both at a time not possible, so not view
serializable.

Q.7 (b) In order for each B-tree node to fit in a

single disk block the maximum value
of p is.
(p × PB) + ((p – 1) × (PD + K)) ≤ B
Shared disk block size B = 1024 byte
Block pointer P=7B
Record (data pointer) P = 10 B
Search field K= 12 B
(p × 7) + ((p – 1) × (10 + 12)) ≤ 1024
7p + 22p – 22 ≤ 1024
29p ≤ 1046
1046 
p ≤  = 36
 29 

Q.10 (d)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 67
 The given schedule is not serializable,
since it is not conflict serializable, not Q.4 (a)
view serializable. Content of index <key, BP> = 6 + 10 =
16
Block factor of database = 512 / 16 =
LEVEL 3 32
Number of block in database =
Q.1 (10) 8192 / 32 = 256
In first level entry for each record,
Number of blocks in first level =
Number of database block
Entry of 1st level
 256 
=  =8
 32 
In second level
Number of blocks in second level =
Number of 1st level Block
Entry size of 2nd level
8
=   =1
 32 

Q.5 (3)
If we insert keys 45, 48, 55 in same
Q.2 (d)
order, then on insertion of key 55, root
The given locking protocol follows the
will be overflow and new level will be
properties of strict 2 PL which is
created.
conflict serializable, recoverable and
avoids cascading rollbacks.

Q.3 (538)
Disk block size = 5000 records
Block size = 10 records or 15 (keys +
Pointers)
Sparse index at 1st level. So number of
disk block at 1st level is number of
block in database.
Database = 5000 / 10 = 500 blocks
 500 
1st level = =
 15  = 33.33 34 blocks

 34 
2nd level = =  15  =
2.26  3 blocks

3rd level = 1 block

Total = [500 + 34 + 3 +1] blocks = 538
blocks.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 68
 Computer Networks

Topics: Concept of Layering, LAN Technologies, Flow & Error

Control Techniques
-------------------------------------------------------------------------------------------
LEVEL 1 a) stop-and-wait
b) go-back-n
c) selective-reject
Q.1 Method of communication in which d) both (a) and (b)
transmission takes place in both
directions, but only in one direction at
Q.6 PURE ALOHA
a time, is called
a) Does not require global time
a) Simplex
synchronization
b) Four wire circuit
b) Does require global time
c) Full duplex
synchronization
d) Half duplex
c) Both (a) and (b)
d) None of these
Q.2 To allow the sender to detect a
collision in CSMA/CD
Q.7 For stop-and-wait flow control, for n
a) Frames must include a checksum.
data packets sent, how many
b) Frames need to be shorter than some
acknowledgments are needed?
propagation time in the network.
a) n
c) Frames must be encrypted and
b) n-1
authenticated.
c) 2n
d) Frames need to be longer than
d) n + 1
maximum propagation time in the
network.
Q.8 Decryption and encryption of data are
responsibility of
Q.3 Bit stuffing refers to
a) Physical layer
a) inserting a ‘0’ in user stream to
b) Data link layer
differentiate it with a data
c) Presentation layer
b) inserting a ‘0’ in data stream to
d) Session layer
avoid ambiguity
c) appending a nibble to the data
Q.9 In which topology, if there are n
sequence
devices in a network, each device has
d) appending a nibble to the use data
n-1 ports for cables?
stream
a) Mesh
b) Star
Q.4 Slotted ALOHA
c) Bus
a) divide time into discrete intervals
d) Ring
b) require global time synchronization
c) both (a) and (b)
Q.10 In OSI reference model which layer
d) none of these
supports port addressing?
a) Data link
Q.5 In which ARQ, when a NAK is
b) Physical
received, all frame sent since the last
c) Transport
frame acknowledged and retransmitted
d) Presentation

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 69
 LEVEL 2 b) It can be used with slotted aloha but
not with carrier sense multiple access.
c) Over a short time scales, it improves
Q.1 Which of the following MAC address the fairness of the throughput achieved
represent multicast address?
by different nodes compared to not
a) 24:2B:34:59:48:1B using the mechanism.
b) B4: B3:C7:C5:89:10
d) It ensure that two or more nodes that
c) B3:B4:C5:C7:10:89 experience a collision in a time slot
d) Both (a) and (c)
will experience a lower probability of
colliding with each other when they
Q.2 Consider the following statements:
each retry that packet.
S1: Data link layer define the
boundaries of frames because framing
Q.5 The maximum length of the cable
is always of fixed size. (min) for transmitting data at a rate of
S2: Bit stuffing is the process of adding
600 Mbps in IEEE 802.3 LAN with a
one extra 0 after 6th bit whenever five frame size of 18,000 bits is ______
consecutive 1’s follow a 0 in the data,
meter. (Assume the signal speed in the
if 0111110 is assumed as a flag. cable to be 2,50,000 km/s)
Which of the following option is
correct?
Q.6 If a network designer wants to connect
a) Only S1 is true.
5 routers as point-to-point simplex line,
b) Only S2 is true. then total number of lines required
c) Both S1 and S2 are true.
would be?
d) Neither of S1 or S2 is true. a) 5
b) 10
Q.3 Which of the following are not true in c) 20
IEEE 802.3?
d) 32
S1: The maximum frame size that is
supported by IEEE 802.3 is 1500 bytes
Q.7 Start and stop bits are used in serial
S2: The maximum payload that is
communication for
supported by IEEE 802.3 is 1518 bytes
a) Error detection
S3: Minimum frame size and payload
b) Error correction
size are 46 and 64 bytes respectively.
c) Synchronization
a) Only S3
d) Slowing down the communication
b) Only S2 and S3
c) Both S1 and S2
Q.8 Select incorrect statement in error
d) All of these
retransmission used in continuous
ARQ method.
Q.4 Binary exponential back off is a
a) Go-back-N method requires more
mechanism used in some MAC
storage at the receiving side
protocols. Which of the following is
b) Selective repeat involves complex
true?
operation than GO-back-N
a) It ensures that two nodes that
c) Go-back-N has better line utilization
experience a collision in a time slot
d) Selective repeat has better line
will never collide with each other when
utilization
they each retry that packet.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 70
 Q.9 Consider a 10 km long broadcast LAN transmissions of a frame to make it
has 1010 bps bandwidth and uses success is _____.
CSMA/CD. The signal travels along
Q.3 Assume there are 100 nodes connected
the wire at 2.5 ×108 m/s. What is the
to a 1000 meter length of the coaxial
minimum packet size that can be used
cable. Using some protocol, each node
on this network?
can transmit 50 frames/second, where
a) 106 bits the average frame length is 2500 bits.
b) 107 bytes The transmission rate at each node is
c) 104 bits 108 bps. The efficiency of this protocol
is ____. (Upto 1 decimal place)
d) 105 bytes
Q.4 In synchronous transmission, 5 eight
Q.10 Choose the best matching between bit characters are included in 30 eight
List-I and List-II bit information characters. If bit rate of
List-I (Protocol) sender is 4200 bits/sec, what is the bit
A. MAC, LLC, PPP rate of receiver (in bits per sec)?
B. DHCP, ICMP, BGP
C. Telnet, FTP, SMTP Q.5 Consider a network connecting two
nodes A and B having propagation
List-II (Layers)
delay of 3 ×104 µ sec . Bandwidth of
1. Application Layer
2. Transport Layer network is 50 Mbps. If each frame size
3. Network Layer is 5000 bytes and both uses Go-Back-N
4. Data link layer sliding window protocol, where
maximum 100 frames can be sent at a
Codes: time then the maximum possible data
a) A-3, B-2, C-4 rate is ____ (Mbps)
b) A-3, B-4, C-1
c) A-4, B-3, C-1
d) A-1, B-2, C-3

LEVEL 3
Q.1 Consider a 40 Kbps satellite link has a
propagation delay of 504 ms. The
transmitter employs the “go back n
ARQ” scheme with n set to 15.
Assuming that each frame is 105 bytes
long, the maximum data rate possible
is ______ (in kbps). (Upto 2 decimal
places)

Q.2 If probability of frame reaching safely

is 0.1 then mean number of

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 71
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(d) (d) (a) (c) (b) (a) (a) (c) (a) (c)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(c) (d) (d) (d) (3750) (b) (c) (c) (d) (c)

LEVEL 3

1 2 3 4 5
(12.24) (10) (12.5) (3500) (66.67)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 72
 Encryption and decryption of data are
SOLUTIONS responsibility of presentation layer.

LEVEL 1 Q.9 (a)

In mesh topology, every node is
connected with every other node. So, if
Q.1 (d) there are n devices in a network, each
A half-duplex (HDX) system provides device has n-1 ports for cables.
communication in both directions, but
only one direction at a time (not
Q.10 (c)
simultaneously).
Port addressing is the functionality of
transport layer.
Q.2 (d)
In CSMA/CD to detect a collision
condition will be T.T. ≥ 2 T.P. i.e. LEVEL 2
frame size should be greater than
maximum propagation time between Q.1 (c)
two stations. For MAC address to be multicast
address, LSB of 1st byte (from MSB
Q.3 (a) side) will be 1.
Bit stuffing is required when there is a Here, 2A = 00101010
flag of bits to represent one of the B4 = 10110100
incidents like start of frame, end of B3 = 10110011
frame, etc, If same frame bits appear in Only B3 has LSB as 1, so, only (c)
the data stream, a zero is inserted. The represents multicast MAC address.
receiver deletes this zero from the data
stream. Q.2 (d)
Variable size framing is also used apart
Q.4 (c) from fixed-size framing. Hence, S1 is
false.
Q.5 (b) If the flag is 0111110, while bit
In the case of go-back-n protocol, stuffing ‘0’, it will be stuffed after 5th
when a NAK is received, all the frames bit. So, S2 is false.
sent since the last frame acknowledged
are retransmitted. Q.3 (d)
Maximum frame size = 1518 bytes
Q.6 (a) Maximum payload size = 1500 bytes
Pure ALOHA Does not require global Minimum frame size = 64 bytes
time synchronization. Slotted ALOHA Minimum payload size = 46 bytes
requires global time synchronization.
Q.4 (d)
Q.7 (a) By binary exponential back off
For stop-and-wait flow control for n mechanism, after collision if both
data packets sent, n acknowledgements stations try to send same frame, their
are needed. probability of collision will decrease
but not guaranteed that collision will
Q.8 (c) never happen. Binary exponential back

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 73
 off mechanism used in CSMA
protocol. Q.9 (d)
Minimum packet size =
Q.5 (3750) 2×Distance×Bandwidth
Tt ≥ 2Tp Velocity
Distance = 10 km
Frame length Distance
⇒ ≥2* Velocity = 2.5 ×108 m/s
Bandwidth Speed
Bandwidth = 1010 bps
18000 2x
⇒ 6
≥ Minimum packet size =
600*10 250000 km
2×10×103 ×1010
18 × 10 × 25 × 104
3
= 8 ×105 bits =
⇒ Km ≥ 2x 2.5 ×108
600 × 106 5
⇒ x = 3.75 km 10 bytes
So, x = 3750 m
Q.10 (c)
Q.6 (b) Option (c) is the best matching.
If 1st router connects with other 4
routers then total for 1st = 4. Same as
2nd router connects with 3 routers, so LEVEL 3
total for 2nd = 3. Same as for 3rd router
connects with 2 routers, so total = 2. Q.1 (12.24) [12.23 - 12.25]
Same as 4th router connects with 1. So Tx = 105 x 8 bits / 40 Kbps = 21 ms
total simplex lines will be = 4 + 3 + 2 + Tp = 504 ms
1 = 10. So answer is ‘b’. Tp
a= = 504 / 21 = 24
Tx
Q.7 (c)
Start and stop bits are used to W
Efficiency of GBN =
synchronize sender and receiver in (1 + 2a )
serial communication. where, W = Window size
= 15 / (1 + 48)
Q.8 (c) = 15 / 49
Go-back-N has better line utilization is BW utilization or throughput or
incorrect statement. Because if there is maximum data rate = Efficiency × BW
an error at the receiver the frames must = 15 / 49 × 40 kbps = 12.24 Kbps
be retransmitted where the error
occurred in the transmission. Q.2 (10)
Option (a) is correct because in Go- The mean number of transmissions
Back-N method sometimes when before successful transmission is 1 / p,
acknowledgement is not received by where p is the probability of frame
sender then sender assumes that packet reaching safely. Therefore required
is not received by receiver so he again number of transmissions are 1/0.1 = 10.
sends the packet. So receiver requires
more memory to store. Q.3 (12.5) [12.4- 12.6]
Selective repeat involves complex Node throughput = 50 frames/sec
processing than GO-back-N is correct. = 50 x 2500 bits/sec

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 74
 = 125000 bits/sec
System throughput = 125000 × 100
nodes = 125 × 105 bps
Maximum system rate = 108 bits/sec
125*105
Efficiency = × 100 = 12.5 %
108

Q.4 (3500)
Synchronous bits are not taken by
receiver.
They are sent by sender just to alert the
receiver about the incoming data.
5 eight bit character = 40 bits
30 eight bit character information =
240 bits
40 synchronous bits → 240 info bits
X Synchronous bits → 4200 info bits
X = 40 * 4200 / 240 = 700
synchronous bits

Total data bits = 4200 – 700 = 3500

bits/sec

Q.5 (66.67)
Bandwidth = 50 Mbps
Frame size = 5000 bytes
Propagation delay = 3 ×104 µ sec
= 30 msec
So in 1RTT =
30×2×10-3×50×106 bits
= 3000×103bits
But maximum bits that can be
transferred in one time = 100 frame
= 5000 × 8 × 100 bits
= 4000000 bits
So, effective bandwidth =
4000000
×50 Mbps
3×106
4
= ×50×106 bps
3
= 66.67×106 bps
= 66.67 Mbps

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 75
 Computer Networks

Topics: Switching, Network Layer, Routing Algorithms

-------------------------------------------------------------------------------------------
LEVEL 1 Q.5 Which of the following is not a class C
host address?
a) 224.4.5.6
Q.1 How many class-A networks are
b) 192.4.5.10
possible?
c) 203.0.0.0
a) 128
d) 198.8.19.0
b) 256
c) 127
Q.6 If length of header is 48 then what is
d) 126
the HLEN field of IP?
a) 1100
Q.2 Consider the following statements
b) 1010
about routing:
c) 1111
S1: Distance vector and link state are
d) 1011
inter domain routing protocols.
S2: Routing information protocol is
Q.7 Consider statements S1 and S2 and
based on Bellman ford algorithm and
choose the correct answer from the
open source shortest path first is based
options.
on Dijkstra algorithm.
S1: Default Subnet Mask for class A
Which of the statements are true?
network is 255.255.255.0.
a) Both are true
S2: Default Subnet Mask for class C
b) S1 is true, S2 is false
network is 255.0.0.0.
c) S1 is false, S2 is true
a) Only S1 is true.
d) Both are false
b) Only S2 is true.
c) Both S1 and S2 are true.
Q.3 Consider the following statements:
d) Both S1 and S2 are false.
S1: In IPv6, provider based unicast
address is identified by prefix bits of
Q.8 In the network 143.128.67.235 / 20,
address as 001.
what is the decimal value of 3rd octet?
S2: In IPv6, geographic based unicast
a) 176
address is identified by prefix bits of
b) 79
address as 010. Which of the following
c) 51
is true?
d) 49
a) Only S1
b) Only S2
Q.9 IP-address of a class B is given as
c) Both S1 and S2
130.60.32.0. Find Network id for this
d) Neither S1 nor S2
IP address.
a) 130.60.32.255
Q.4 ICMP (Internet Control Message
b) 130.60.0.255
Protocol) is
c) 130.60.0.0
a) a protocol that handles error and
d) 130.60.15.32
control messages
b) a protocol used to monitor
Q.10 Limited broadcast address of a network
computers
is always?
c) both (a) and (b)
a) 255.255.255.255
d) none of these

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 76
 b) 255.255.0.0 d) Under some circumstance, a circuit
c) 192.127.255.255 switched network may prevent some
d) 255.0.0.0 sender from starting new conversation.

LEVEL 2 Q.4 Consider the following graph:

Q.1 Consider station “A” send data to

station “B”. Data frame contain MAC
address, IP address and port address.
Which of the following address may
change on intermediate router?
a) Destination IP address Find the link state routing table for
b) Destination MAC address node V using OSPF.
c) Source MAC address a)
d) Source IP address

Q.2 Consider the following statements:

S1: In IPv4 header total length field
denote the size of payload (excluding
header of IP) comes from TCP layer.
S2: In IPv6 header payload length
indicate the size of pay load (header of
IP+ data from TCP layer). b)
Which of the following is true?
a) Only S1
b) Only S2
c) Both S1 and S2
d) Neither S1 nor S2

Q.3 Which of the following is incorrect?

a) In circuit switched network,
switches require connection c)
establishment and teardown phases,
where as in packet switched network
doesn't.
b) Unlike packet switching, circuit
switched networks do not need any
information about the network
topology for correct functionality.
c) Once a connection is correctly
established, a switch in circuit switched d)
network can forward data directly
without requiring data frames to
compute a destination address.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 77
 transmission delay of 20 ms. Assume
that the transmission time for the
acknowledgment and the processing
time at nodes are negligible. It's needed
to design a selective repeat sliding
window protocol for this network.
Then the minimum size in bits of the
sequence number field to achieve a link
Q.5 A router uses the following routing utilization of atleast 50% has to be __.
table:
Q.8 Two computers C1 and C2 are
configured as follows:
• C1 has IP address 197.168.76.24 and
network mask 255.255.192.0.
• C2 has IP address 197.168.114.28
and network mask 255.255.224.0.
Find the next hop (where router will Which one of the following statements
send the packet), if the router has IP is true?
address “205.32.16.63” a) C1 and C2 both assume they are on
a) R0 the same network.
b) R1 b) C2 assumes C1 is on same network,
c) R2 but C1 assumes C2 is on a different
d) R3 network.
c) C1 assumes C2 is on same network,
Q.6 Consider a source computer (S) but C2 assumes C1 is on a different
transmitting a file of size 105 bits to a network.
destination computer (D) over a d) C1 and C2 both assume they are on
network of three routers (R1, R2 and the different networks.
R3) and four links (L1, L2, L3 and L4).
L1 connects S to R1; L2 connects R1 Q.9 Following specifications are given for
to R2; L3 connects R2 to R3 and L4 a packet switch network.
connects R3 to D. Let each link be of Bandwidth = 200 bps
length 100 km. Assume signals travel Packet Size = 5000 bits
over each link at a speed of 108 meters Distance between two hopes = 500 m
per second. Assume that the link Speed of signal = 80 m/s
bandwidth on each link is 1 Mbps. Let Number of hopes = 11
the file be broken down into 100 Find the total time for a packet to travel
packets each of size 1000 bits. The from sender to receiver.
total sum of transmission and
propagation delays in transmitting the Q.10 How many class-B Hosts per network
file from S to D is _______ (in msec). (Usable Addresses) are possible?
a) 254
Q.7 Consider a network connecting two b) 65534
systems 10000 km apart. The c) 65535
propagation speed of the media is 4 x d) 255
106 meters per second. Link has a

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 78
 LEVEL 3 d) 172.89.0.1/26 and
172.89.255.254/26
Q.1 Consider hierarchal routing with 4800 Q.5 Consider the following scenario:
routers, they are divided in hierarchy as
clusters, regions, routers. What is the
minimum size of routing for three level
hierarchy?
a) 5, 48, 20
b) 3, 80, 20 Suppose a TCP message that contains
c) 15, 16, 20 900 B of data and 20 B of TCP header
d) 5, 24, 40 is passed to IP code at host A for
delivery to B. IP header is of 20 B.
Q.2 Which of the following statements is Assume that link A-R1 has maximum
TRUE about IPv4 protocol? frame size of 1024 B including a 14-B
a) A best-effort delivery service, best- frame header, link R1 – R2 can support
effort means that IPv4 provides no a maximum frame size of 512 B,
error control or flow protocol. including an 8 B frame header and link
b) TTL field is used to optimize R2 – B can support a maximum frame
throughput. size 512 B including a 12 B frame
c) IPv4 supports only unicast and header.
broadcast addressing mode. The lengths of the last fragment at each
d) ARP is used to acquire the IP of the three links respectively are:
address of a host whose MAC address a) 920; 440; 440
is known. b) 960; 460; 440
c) 960; 460; 460
Q.3 Consider the flow specification having d) 940; 460; 460
the maximum packet size 800 bytes,
token bucket rate of 8 × 106 bytes/sec.
Token bucket size is 4 MB and the
maximum transmission rate 16 × 6
bytes/sec. The time for which burst be
send at maximum speed will be ____
(in sec). (Up to 2 decimal places)

Q.4 An organization is granted the block

172.89.0.0/16. The administrator wants
to create 1024 subnets using 10 bits.
The first and last addresses of any host
in subnet 1024 respectively are
a) 172.89.255.1/26 and
172.89.255.245/26
b) 172.89.255.1/26 and
164.76.255.255/26
c) 172.89.0.1/26 and
164.76.255.255/26

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 79
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(d) (c) (d) (a) (a) (a) (d) (b) (c) (a)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(b) (d) (b) (c) (d) (107) (8) (c) (312.5) (b)

LEVEL 3

1 2 3 4 5
(c) (a) (0.50) (d) (d)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 80
 SOLUTIONS Q.8 (b)
IP address of network: 143.128.67.235
/ 20
LEVEL 1 143.128.0100 0011.11101011
Last IP address assigned to any host:
143.128.01001111.11111110
Q.1 (d)
3rd octet: 01001111 = 79
Total number of networks = 27 , from
which 2 networks are reserved. So, 126 Q.9 (c)
numbers of networks are possible. Network id is computed by ANDING
of Network mask and IP address.
Q.2 (c) Here, Network mask is 255.255.0.0,
Distance vector and link state are intra because of Class B network.
domain routing protocol (routing inside By doing AND operation, answer will
an autonomous system is intra domain be 130.60.0.0.
routing)
Routing information protocol is based Q.10 (a)
on distance vector routing follows In a limited broadcast packet
Bellman ford algorithm and OSPF is destination address is all ones.
based on link state routing follows
Dijkstra algorithm.
LEVEL 2
Q.3 (d)
In IPv6 address: Q.1 (b)
Prefix bits: 001 indicate geographic Since packet encapsulate datagram and
unicast address. frame encapsulate packet. So, cannot
Prefix bits: 010 indicate provider change IP address and port number but
unicast address. destination MAC address may change
So, both statements are false. on intermediate routers.

Q.4 (a) Q.2 (d)

In IPv4: Total length indicates size of
Q.5 (a) header + data from TCP layer. Since
Class C address range is 192.0.1.1 to header size is variable.
223.255.254.254. So, option (a) cannot In IPv6: Payload length indicates size
be the answer. of data comes from TCP layer.
So, both statements are false.
Q.6 (a)
HLEN × 4 = Header Size Q.3 (b)
Here, header size = 48. Since in circuit switched network
So, HLEN = 12. Which is 1100. before sending any information
connection is established. So
Q.7 (d) information about whole network is
Class A network has Default subnet needed for starting data transfer.
mask 255.0.0.0.
Class C network has Default subnet Q.4 (c)
mask 255.255.255.0.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 81
 Routing table for ‘V’ by using Dijktra's N
≤1/2
Algorithm. 1 + 2a
V to V = 0 Tp
V to W = 4(Via V) a=
Tt
V to X = 12(Via W) Tp (Propagation delay) = d / v =
V to Y = 10(Via V)
V to Z = 15(Via X) 1000*103
= 2.5 sec
4*106
Q.5 (d) Tt (Transmission delay) = 20 msec
Routing in classless addressing follows
“Longest Mask Matching”. This states a = 2.5 / 20*10-3 = 125
that the routing table is sorted from the N 1
≤
longest mask to the shortest mask. 1 + 2*125 2
So, here firstly start with /27
To find N/W address = (Window size) N ≤ 125.5
205.32.16.63 AND 255.255.255.224 In selective repeat protocol, the
= 205.32.16.32 {Matched with R3} window must be less than half the
So, interface chosen is R3. sequence number space.

Q.6 (107) 2m
≤ 125
2
2m-1 ≤ 125
2m-1 = 27
Propagation delay to travel from S to m=8
R1
Distance Q.8 (c)
= 105 / 108 = 1ms Subnet/Network mask is to check
Link Speed whether some other IP belongs to the
same network or some other network.
Total propagation delay to travel from Case-I:
S to D = 4 * 1ms = 4ms. (IP of C1) IPc1 = 197.168.76.24
Total transmission delay for 1 packet = (Network mask of C1) NMc1 =
Number of bits 255.255.192.0
4* = 4 * (1000) / 106
Bandwidth NIDc1c1 = 197.168.64.0
= 4 ms
So the first packet will take 8 ms to Network ID of C2 as per C1:
reach D. While first packet was (IP of C2) IPc2 = 197.168.114.28
reaching D, other packets must have NMc1 = 255.255.192.0
been processing in parallel. NIDc2c1 = 197.168.64.0
So D will receive remaining packets, 1
packet per 1 ms from R3. So remaining Since, NIDc1c1 = NIDc2c1
99 packets will take 99 ms and total So, C1 will assume C2 is on the same
time will be 99 + 8 = 107 ms. network.

Q.7 (8) Case-II:

Link utilization ≤ 50 % (IP of C2) = 197.168.114.28

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 82
 (Network mask of C2) = 255.255.224.0 IPv4 supports multicast addressing
NIDc2c2 = 197.168.96.0 mode along with unicast and broadcast
addressing mode.
Network ID of C1 as per C2: ARP is used to acquire the MAC
IPc1 = 197.168.76.24 address of a host whose IP address is
NMc2 = 255.255.224.0 known.
NIDc1c2 = 197.168.64.0
Q.3 (0.50)
Therefore, NIDc1c2 # NIDc2c2 C
t=
So, C2 will assume C1 is on a different M −ρ
network. Where, C: Capacity of token bucket
ρ : Token generation rate
Q.9 (312.5)
Formula for total time in packet switch M: Maximum data rate of token bucket
network = t = Time for which token bucket can
send the data with maximum data rate.
M d
(x-1) + (x-1) , where So, t=
B v
4*106 bytes
x = Number of hopes
M = Packet Size 16*106 bytes / sec− 8*106 bytes / sec
B = Bandwidth t = 4 / 8 = 0.50 sec.
d = distance between hopes
v = speed of signal Q.4 (d)
So, 10*25 + 10*6.25 Number of subnets = 1024
= 250 + 62.5 Bits required for subnet = 10
= 312.5 s Subnet mask = 255.255.255.192
[i.e. take bits from host portion]
Q.10 (b) Number of hosts = 26 – 2
Class B network has 16 host bits. Ranges are:
From which 2 hosts cannot be used. 172.89.0.0/26 to 172.89.0.63/26
So, (216 – 2) = 65,534. 1st subnet
172.89.0.64/26 to 172.89.0.127/26
2nd subnet
LEVEL 3 :
:
Q.1 (c) 172.89.255.192/26 to
For hierarchical routing, table size 172.89.255.255/26
should be reduced logically, so that 1024th subnet
searching time reduces. 5+48+20 = 73 So, first host address will be
3+80+20 = 103 172.89.0.1/26 and last host address will
15+16+20 = 51 be 172.89.255.254/26.
5+24+40 = 69
Here, 51 is minimum. So, option (c) is
Q.5 (d)
correct. Frame header is the header attached at
DLL. Since fragmentation is at
Q.2 (a) network layer. Hence MTU for A – R1
TTL field is used to prevent packet → 1010B
looping.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 83
 MTU for R1 – R2 → 504B
MTU for R2 – B → 500B
Total IP packet size → 940 B [920 B +
20 B]
Link A-R1:
Length = 940, DF = 0, MF = 0, Offset
=0 [920 B + 20 B]

Link R1 – R2:
1. Length = 500, DF = 0, MF = 1,
Offset = 0 [480 B + 20 B]
2. Length = 460, DF = 0, MF = 0,
Offset = 60 [440 B + 20 B]

Link R2 – B:
1. Length = 500, DF = 0, MF = 1,
Offset = 0 [480 B + 20 B]
2. Length = 460, DF = 0, MF = 0,
Offset = 60 [440 B + 20 B]

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 84
 Computer Networks

Topics: Transport Layer, Application Layer, Basics of Wi-Fi, Network

Security
-------------------------------------------------------------------------------------------
LEVEL 1 Q.6 Which of the following application
layer protocol uses UDP as a transport
layer protocol?
Q.1 Which of the following is an a) FTP
application layer service? b) DNS
a) Remote login c) SMTP
b) Mail Service d) HTTP
c) File transfer and access
d) All of above Q.7 SMTP uses which protocol at the
transport layer
Q.2 The transport layer protocols used for a) TCP
real time multimedia, FTP, DNS and b) UDP
email respectively are: c) IP
a) TCP, UDP, UDP and TCP d) None of these
b) UDP, TCP, TCP and UDP
c) UDP, TCP, UDP and TCP Q.8 In the checksum calculation at TCP,
d) TCP, UDP, TCP and UDP which of the following are used?
a) TCP header
Q.3 The transport layer protocols used for b) TCP data
TFTP, SNMP, SMTP, RIP? c) Pseudo header from IP
a) TCP, TCP, TCP, UDP d) All the above
b) UDP, TCP, TCP, UDP
c) UDP, UDP, TCP, UDP Q.9 What are the main responsibilities of
d) UDP, UDP, UDP, UDP transport layer?
a) End-to-end delivery
Q.4 What mechanism is used by TCP to b) Flow control
provide flow control as segments travel c) Segmentation
from source to destination? d) All the above
a) Sequence number
b) Session establishment Q.10 Wrap around time in TCP depends on
c) Window size a) Sequence number bits
d) Acknowledgement b) Bandwidth
c) Both (a) and (b)
Q.5 In TCP, the sequence number given to d) None of these
a segment is sequence number of
_______ byte
a) First byte LEVEL 2
b) Last byte
c) Middle byte Q.1 Match the following control commands
d) None of these and their respective port numbers:
Group-A Group-B
A. HTTPS 1. 110
B. POP 2. 443

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 85
 C. FTP control 3. 21 C. Connect ( ) 3. Associates a
D. FTP data 4. 20 socket with
Codes: socket address
A B C D structure.
(a) 1 2 3 4 D. Poll () 4. Check the
(b) 1 2 4 3 state of a socket
(c) 2 1 3 4 in set of sockets.
(d) 2 1 4 3 Codes:
A B C D
a) a (a) 4 2 1 3
b) b (b) 3 2 4 1
c) c (c) 3 2 1 4
d) d (d) 4 1 2 3

Q.2 Station A needs to send a message a) a

consisting of 10 packets to station B b) b
using a sliding window of size 3. All c) c
packets are ready and can be d) d
transferred immediately. Selective
repeat and GBN are used at 2 different Q.5 If the TCP round-trip time, RTT is
times and every 5th packet gets lost for currently 5.5 msec and the
both protocols. (ACK's from B never acknowledgement comes in after 1.5
gets lost). Let x and y be the number of ms. If α = 7/8 , then the new RTT
transmissions that A has to make in estimate is ____ msec.
selective repeat and GBN respectively
to ensure safe delivery to B. Then x + y Q.6 If the initial sequence number is 1 and
=? it increment the counter by 2, 56,000
for every 4 sec, how long does it take
Q.3 In a RSA cryptosystem, a participant A for the counter to wrap around?
uses two prime numbers p= 13 and q = a) 67,108 seconds
31 to generate public and private keys. b) 54,554 seconds
If the public key of A is 103, then the c) 33,455 seconds
private key of A is _____. d) 44,455 seconds

Q.4 Match the following functions: Q.7 In Go-Back-N protocol, if the

Group-A Group-B maximum window size is 127, what is
A. Bind () 1. Used at client the range of the Sequence number?
side to assign a a) 0 to 127
free local port b) 0 to 128
number to a c) 1 to 127
socket. d) 1 to 128
B. Listen () 2. Causes bound
TCP socket to Q.8 Consider a TCP connection in a state
enter listening where there are no outstanding Acks.
state. The sender sends two segments back to
back. The sequence numbers of first

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 86
 and second segments are 750 and 870 c) When a strict routing has been
respectively. The first segment was specified by the source but can't be
lost, but second was received correctly followed by the intermediate routers,
by the receiver. Let X be the amount of such events will not be reported by
data carried in first segment (in Bytes), ICMP.
Y be the Ack number sent by the d) Symmetric encryption algorithm is
receiver. The value of X and Y are: same as public key encryption
a) 120 and 870 algorithm.
b) 120 and 990
c) 750 and 990 Q.2 Consider the following statements:
d) 120 and 750 S1: Asymmetric key cryptography and
digital signature both use the private
Q.9 In public key, private key and public keys of the sender.
cryptography, if ‘A’ has PuA and PrA, S2: Public key cryptosystems require n
‘B’ has PuB and PrB as public and (n-1) / 2 keys for a set of n individual
private keys. Then if ‘A’ wants to send to be able to communicate with each
a message to ‘B’ securely ‘A’ will use other.
which key for encryption Which of the statements options are
a) PrA true?
b) PuA a) S1 and S2 are true
c) PuB b) S1 is true, S2 is false
d) PrB c) S1 is false, S2 is true
d) Both are false
Q.10 In a IP datagram, a TCP segment is
present. Total length of IP datagram is Q.3 Consider the following statements:
1000 bytes. Header length field in TCP S1: TCP is message oriented protocol
header is 7. Then what is size of TCP while UDP is stream oriented.
data present in the datagram. S2: TCP guarantees that No-out-of-
a) 988 order segment is delivered to the
b) 952 process.
c) 964 S3: In congestion avoidance algorithm,
d) 900 the size of congestion window
increases exponentially until it reaches
a threshold.
LEVEL 3 Which of the statements are NOT true?
a) Only S3
Q.1 Which of the following statements is b) Only S2 and S3
true? c) Only S1 and S3
a) The size of TCP receiver window d) Only S1 and S2
never changes throughout the duration
of connection. Q.4 Let the size of congestion window of a
b) Suppose host A is sending a large TCP connection be 38 KB when a
file to host B over a TCP connection, timeout occurs. The propagation time
the number of unacknowledged bytes of the connection is 100 msec and the
that host A send cannot exceed the size maximum segment size used is 2 KB.
of advertised receiver buffer. The time taken (in msec) by the TCP

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 87
 connection to get back to 36 KB
congestion window is ___.

Q.5 If size of a TCP segment is 1KB and

header length value is 6, the sequence
number = 3500. Given that URG flag =
1 and URG pointer = 45. Then what is
the total size of data. How many of
them are urgent, Give the sequence
numbers of urgent data.
a) 45 bytes of urgent data, sequence no.
3500 – 3544
b) 45 bytes of urgent data, sequence
no. 1024 – 1069
c) 46 bytes of urgent data, sequence no.
1024 – 1070
d) 46 bytes of urgent data, sequence
no. 3500 – 3545

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 88
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(d) (c) (c) (c) (a) (b) (a) (d) (d) (c)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(c) (30) (7) (c) (5) (a) (a) (d) (c) (b)

LEVEL 3
1 2 3 4 5
(b) (d) (c) (3000) (d)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 89
 SMTP is reliable protocol. It uses
SOLUTIONS connection oriented service. So, it uses
TCP as transport layer protocol.
LEVEL 1
Q.8 (d)
In TCP check sum calculation, it is
Q.1 (d) calculated on TCP header, TCP data
Remote login – Telnet, and pseudo header from IP.
Mail service – SMTP,
File transfer and access – FTP are Q.9 (d)
application layer protocols.
Q.10 (c)
Q.2 (c) If ‘n’ is the sequence number bits and
Real time multimedia- UDP B is the bandwidth in bytes/sec.
FTP- TCP
2n
DNS- UDP Then, wrap around time =
Email- TCP B
∴ WAT depends on sequence number
Q.3 (c) bits as well as BW.
• Trivial File Transfer Protocol (TFTP)
process includes flow and error control. LEVEL 2
It can easily use UDP.
• UDP is used for management Q.1 (c)
processes such as SNMP.
• SMTP uses TCP. Q.2 (30)
• UDP is used for some route updating With Go-back-N:
protocols such as Routing Information
Protocol (RIP).

Q.4 (c) So, total number of retransmissions is

Window size is used to provide flow 18.
control mechanism.
With selective repeat protocol:
Q.5 (a)
The sequence number given to a
segment is sequence number of first So, total number of retransmission is
byte. 12.
So, total x + y = 18 + 12 = 30
Q.6 (b)
FTP, SMTP and HTTP application Q.3 (7)
layer protocol uses TCP as transport p = 13, q = 31
layer protocol, whereas DNS uses n = p × q = 13 × 31 = 403
UDP. ɸ(n) = (p – 1) × (q – 1) = 12 × 30 = 360
(d × e) mod ɸ(n) = 1
Q.7 (a) (103 × e) mod 360 = 1
Here e =?
(103 × 7) mod 360 = 1

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 90
 Amount of data carried in first segment
Q.4 (c) = 870 – 750 = 120.
Bind () is used on the server side and Acknowledgement number field
associates a socket with socket address contains the next sequence number that
structure. is expected to receive. Here, 1st
Listen () is also used on the server side segment is lost. Receiver expects the
and bound TCP socket to enter 1st segment sequence number which is
listening state. 750. So, Ack number sent by the
Connect () is used on the client side receiver is 750.
and assigns a free local port number to
a socket. In case of TCP socket, it Q.9 (c)
causes an attempt to establish a new If the message is encrypted using PuB,
TCP connection. then it can be decrypted using only PrB,
Poll () is used to check the state of a which will be known only to ‘B’.
socket in a set of sockets. The set can
be tested to see if any socket can be Q.10 (b)
written to, read from, or if an error
occurred.

Q.5 (5)
IRTT = 5.5 msec
α = 7/8
NRTT = 1.5 msec Given IP header field = 5
Basic algorithm = ∴ IP header size= 5*4= 20 bytes
α ( IRTT ) + (1 − α ) ( NRTT ) Similarly, TCP header field = 7
RTT = 7/8 (5.5) + 1/8 (1.5) = 5 msec Therefore, TCP header size = 7*4 = 28
TCP data = 1000 – 20 – 28 = 952.
Q.6 (a)
Given for every 4 secs, counter is
incremented by 2, 56,000 LEVEL 3
So for every 1 sec, counter increments
by 2, 56, 000/4 = 64000 Q.1 (b)
The sequence no. is 32 bit long and it 1). Receiver window is used to give the
can hold only 232 – 1. sender an idea of how much free buffer
So it takes (232 – 1) / (64000) = space is available at receiver. So, as the
67,108.86 seconds receiver buffer changed receiver
window also change.
Q.7 (a) 2). TCP does not allow the host to send
Sequence no. will be given to sender more data than receiver buffer
window and receiver window. Sender requirement which restrict the data
window size is 127 and receiver overflow.
window size is 1. So, 127 + 1 = 128. 3). When a strict routing has been
Sequence no. will start from 0. So, specified by the source but can't be
answer is 0-127. followed by the intermediate routers,
such events will be reported by ICMP.
Q.8 (d)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 91
 4). Secret key encryption algorithm is 2 → 4 → 8 → 16 → 19 → 21 → 23 →
often referred to as symmetric 25 → 27 → 29 → 31 → 33 → 35 →
encryption algorithm as the same key 37 → 38
can be used in bidirectional Time taken to reach 38 KB = 15
communication between sender and segments * 200 msec = 3000 msec.
receiver.
Q.5 (d)
Q.2 (d) Given Size of TCP segment = 1KB =
S1: Asymmetric key cryptography uses 1024 Bytes
the private key for decryption and Header Length field is 6, so header size
public key for encryption of the = 6*4 = 24 bytes
receiver. While digital signature use Total data size
the private and public keys of the = size of Segment – Header size
sender. = 1024 – 24 = 1000 bytes of data
S2: For public key encryption, each Starting Sequence no. is 3500
individual needs to have a public and So the range of sequence no. of the
private key, so the total keys required data is 3500 to 4499
is 2 × n. URG pointer = 45 so data from 0th byte
till 45th byte are urgent so 46 bytes are
Q.3 (c) urgent data.
S1: UDP is a message oriented Therefore, the urgent data is 1000 to
protocol. In UDP, a process sends 1045 and its sequence no. range is
messages, with predefined boundaries, 3500 – 3545.
to UDP for delivery. TCP allows the
sending process to deliver data as a
stream of bytes and allows the
receiving process to obtain data as a
stream of bytes. So TCP is stream
oriented protocol.
S2: Data may arrive out of order and
temporarily stored by the receiving
TCP, but TCP guarantees that No-out-
of-order segment is delivered to the
process.
S3: In congestion avoidance algorithm,
the size of the congestion window
increases additively until congestion is
detected. The window size increases
exponentially in slow-start algorithm.

Q.4 (3000)
Congestion window = 38 KB
Threshold = 19 KB
Round trip time = 2 × propagation time
= 2 × 100 = 200 msec

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 92
 Computer Organization and Architecture

Topics: Memory hierarchy, Machine instructions, Addressing modes,

ALU, Control Unit
-------------------------------------------------------------------------------------------
LEVEL 1
Q.1 Consider the following µ -program.
I1: MAR ← IR[Addr]
I2: MBR ← M[MAR]
I3: ALU ← MBR
Which of the following operation is
performed by above µ -program?
(a) Instruction fetch Q.6 Consider a 2-way set associative cache
(b) Direct operand fetch with 8 cache blocks. If the memory
(c) Interrupt sub program initiation block requests are accessed 2 times in
(d) Indirect operand fetch the following order 0, 4, 8, 4, 0, 4, 8, 4,
3, 15, 19, 15, 3, 15, 19, 15. If LRU
Q.2 Suppose the cache memory is 10 times replacement policy is used, then the
faster than main memory and it is used total number of misses are ______.
40% of the time. The performance is
gained by introducing this cache is Q.7 Consider the main memory and cache
_____. (Upto 2 decimal places) memory of size 4 GB and 2 MB
respectively, block size of both main
Q.3 Consider a Direct cache with 64 blocks memory and cache memory is same.
and block size of 16 bits. The block The maximum size of tag bits if direct
number where the byte address mapped cache is used _____. (Assume
(1440)10 mapped in the cache is ____. memory is byte addressable)
Q.4 Which of the following is true? Q.8 The access time of cache memory is 45
(a) In write through protocol, cache nsec and that of main memory is 750
location and main memory location are nsec. It is found that 75% of memory
updated simultaneously. requests are for read and remaining for
(b) In write back protocol, cache write. If the hit access for read and
location and main memory location are write is 0.9 and 1 respectively and
updated simultaneously. write through protocol is used, then the
(c) Modified or dirty bits are used by average memory access time is ____.
write through protocol. (in nsec) (Upto 3 decimal places)
(d) None of these
Q.9 Consider a P-way set associative cache
Q.5 A CPU has 24 bit instructions and we consisting of 128 lines with line offset
have to calculate the sum of n number of 64 words. Then number of TAG, set
by using below code: and words offset fields are 9, 5 and 6
The value of X, if target address of bits respectively. If CPU generator 20
branch is loop, when instruction is uses bits address of a word then value of P
PC relative addressing mode is ____. is _____.
(Assume memory is byte addressable)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 93
 Q.10 A computer has 170 different
operations. Word size is 4 bytes one
word instruction requires two address
fields. One address for register and one
address for memory. If there are 37
registers then the memory size is_____
(in KB).
(a) 2, 6 (b) 2, 8
LEVEL 2 (c) 8, 2 (d) 6, 2

Q.4 Consider a micro program control unit

Q.1 Consider a machine with a byte
and list of corresponding properties in
addressable main memory of 256 MB,
control unit design:
block size of 128 bytes and 8-way set
associative cache of size 64 kB. If the
address of one of the memory location
AB01C23H accessed by the CPU.
What are the tag field of the
corresponding cache line is
(a) 101010110000000 Which of the following is the correct
(b) 101110100000 match between the Micro program
(c) 101010110000 control unit and their properties?
(d) 101101010100000 (a) (P - 1, 4, 6) and (Q - 2, 3, 5)
(b) (P - 1, 4, 5) and (Q - 2, 3, 6)
Q.2 Which of the following statements are (c) (P - 2, 4, 6) and (Q - 1, 3, 5)
True? (d) (P - 2, 3, 6) and (Q - 1, 4, 5)
S1: Update bit is used in the write back
cache to indicate the cache updation. Q.5 Consider the evaluation of the
S2: In hierarchical memory access following expression tree on a machine
organization, CPU perform read and in which memory can be accessed only
write operation on only level 1 through load and store instructions.
memory. The variables a, b, c, d, e, and f are
S3: In simultaneous memory access initially stored in memory. The binary
organization, CPU perform read and operators used in this tree can be
write operation on any level of evaluated by the machine only when
memory. the operands are in registers. The
(a) S1 and S2 only instructions produce result only in a
(b) S1 and S3 only register.
(c) S2 and S3 only
(d) S1, S2 and S3

Q.3 Consider the following assembly code:

What is the number of anti dependency
and output dependency present in the
below code?

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 94
 While designing the vertical µ -
If no intermediate results can be stored programmed control unit, single
in memory, what is the minimum address field format is used for branch
number of registers needed to evaluate control logic. The size of control
this expression? memory in byte is ____.
(a) 2 (b) 3
(c) 4 (d) 5 Q.9 Suppose a CPU contains 1000 memory
references there are 40 misses in L1
Q.6 Suppose that a 4-way set associative cache (First Level Cache) and 20
cache has 2048 cache lines and main misses in the L2 cache (Second Level
memory of size 225 bytes. Both cache Cache). Assume miss penalty from the
and main memory are partitioned into L2 cache to memory is 100 clock
64 words block, where word size is 1 cycles the hit time of L2 cache is 10
byte. Which of the following represents clock cycles, the hit time of L1 cache is
the size of tag memory? 1 clock cycle. What is the average
(a) 2800 bytes memory access time?
(b) 2048 bytes (a) 3.4 clock cycles
(c) 2560 bytes (b) 3.5 clock cycles
(d) 2304 bytes (c) 5.3 clock cycles
(d) 1.8 clock cycles
Q.7 Consider the following instruction uses
addition operation i.e. C ← [A] + [B]: Q.10 Consider the following µ -operation:
(i) Add A, B [[A] + [B] → B] I1: PC → MBR I2: IR[Add] → MAR
= I3: M[MAR] → MBR I4: MBR → ACC
MOV B, C [[B] → C], I5: PC → MAR I6: SP → MAR
Add A, C [[A] + [C] → C] I7: MBR → M[MAR] I8: ACC → MBR
Which of the following sequence is
(ii) Add A, B, C [[A] + [B] → C] correct for operand store?
= (a) I1, I4, I7
MOV A, C [[A] → C], (b) I8, I5, I7
Add B, C [[B] + [C] → C] (c) I8, I2, I7
(d) None of these
In which of the following option, result
of RHS execution is same as execution
of LHS? LEVEL 3
(a) Only (i)
(b) Only (ii) Q.1 Consider a 32 bit microprocessor that
(c) Both (i) and (ii) has an on chip 16 Kbyte four way set
(d) Neither (i) nor (ii) associative cache. Assume that cache
has a line size of four 32 bit words.
Q.8 Consider a CPU where all the Which of the following set number in
instruction requires 10 clock cycles to the cache to which the word from
complete execution. There are 258 memory location FFFAE8FA is
instructions in instruction set. It is mapped ?
found that 129 control signals are (a) 142
needed to be generated by control unit. (b) 143
(c) 135

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 95
 (d) 139

Q.2 A computer has a cache, main memory,

and a disk used for virtual memory. If a
referenced word is in the cache, 30 ns
are required to access it. If it is in main
memory but not in cache, 80 ns are
needed to load it into the cache, and
then the reference is started again. If
the word is not in main memory, 22 ms
are required to fetch the word from
disk, followed by 80 ns to copy it to the
cache and the start again. The cache hit Q.5 A computer system has a main memory
ratio is 0.8 and main memory hit ratio of 64 K words where a word is a byte
is 0.9. in length. It also has a 32 B cache
What is the approximate average time organized as 4 lines of 8 words. The
( µ s) required to access a referenced cache mapping function is the direct
word on this system? mapping function. The tags in the
(a) 450 cache are currently:
(b) 520
(c) 440
(d) 480

Q.3 Consider two cache organizations. The

first one is 32 kB 2-way associative Consider memory addresses 73B2,
with 32 byte block size. The second DF45, 8C7C and 07E9. For each
one is of the 32 kB direct mapped memory address, the cache lines in
cache. The size of an address is 32 bits which they will be mapped are
in both organization. A 2 to 1 respectively.
multiplexer has latency of 0.7 ns which (a) 00, 01, 10, 11
k bit comparator has latency of k / 5 (b) 10, 00, 11, 01
nsec. What is the difference between (c) 11, 00, 01, 10
the hit latencies of both cache (d) 01, 10, 00, 11
organization (i.e. associative hit latency
– direct mapped hit latency) (in nsec)?
(a) 1.2 (b) 0.9
(c) 0.6 (d) 1.4

Q.4 Consider the following cache A and B.

Let the average access times in cache
A and B is tA and tB respectively. Find
the value (in ns) of tA + tB.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 96
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(b) (1.56) (16) (a) (-12) (18) (11) (274.125) (4) (256)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(a) (d) (c) (a) (b) (c) (b) (6450) (a) (c)

LEVEL 3

1 2 3 4 5
(b) (c) (b) (20.9) (b)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 97
 Dirty bits are used by only write back
SOLUTIONS protocol to know which cache block is
updated.
LEVEL 1
Q.5 (-12)
Each instruction of 24 bits = 24 / 8 =3B
Q.1 (b)
In T1 cycle content register goes to
MAR i.e. address which contain actual
data.
In T2 cycle content of MAR i.e. actual
data transferred to MBR.
In T3 cycle content of MBR transferred
to ALU after the execution start.
So, this µ -program represent direct
operand fetch which is shown by below
diagram.
Q.6 (18)
Number of blocks
Number of sets =
2
=8/2=4

Q.2 (1.56)
Apply Amdhal’s law
S= 10
f = 40%
−1
 0.4 
Soverall= 1 − 0.4 + =[0.64]-1 = 1.56
 10 
So total number of miss = 10 + 8 = 18
Q.3 (16)
Block size is 16 bits = 2 bytes. Q.7 (11)
So, Block number with respect to Assume each block contains 2x words
(1440)10 cell is = 1440 / 2 = 720 or bytes
Location in cache memory = So, x bits as word offset.
720 mod 64 = 16

Q.4 (a)
Write through protocol update cache
and main memory simultaneously
221
where in write back first cache is Number of blocks in cache = = 221-x
updated and marked by dirty bit then 2x
main memory is updated.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 98
 LEVEL 2
Q.1 (a)
Cache = 64KB = 216
Block size = Cache line size = 128
Q.8 (274.125) bytes = 27
Taverage read = 0.90 x 45 + 0.10 x 750 216
= 40.5+75 = 115.5 nsec Number of cache lines = 7 = 29
2
In write-through technique, CPU 9
2
performs simultaneous WRITE Number of sets = 3 = 26 = 64
operation in both cache memory and 2
main memory. So,
Taverage read = Max [Tupdation (Cache),
Tupdation (Main memory)]
= Max [45, 750]
= 750 nsec
So, Taverage = 0.75(115.5) + 0.25 (750)
= 86.625 + 187.5
= 274.125 nsec Q.2 (d)
All the above statements are correct.
Q.9 (4) S1: Update bit is used in the write back
cache to indicate the cache updation.
S2: In hierarchical memory access,
CPU perform read and write operation
only on level 1 memory. If miss occur
#lines then data is first transferred to level 1
Number of set = [Since 128
P − way then CPU access data.
lines, so # bits = 7] S3: In simultaneous memory access,
27 CPU perform read and write operation
32 = on any level of memory i.e. not
P − way
necessary to take data first into level 1
27 memory than access it.
P-way = 5 = 22 = 4
2
Q.3 (c)
Q.10 (256) Anti dependency (WAR): I1 – I2, I1 –
I3, I2 – I3, I1 – I4, I3 – I4, I2 – I5, I4 – I5,
I3 – I5
Output dependency (WAW): I1-I5, I2-I4

Q.4 (a)
In Horizontal µ Control unit design
Control signals are in decoded binary
format, Longer Control Word and High
degree of parallelism.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 99
 In Vertical µ Control unit design Q.8 (6450)
Control signals are in encoded binary
format, Shorter Control Word and Low
degree of parallelism.

Q.5 (b) 12 + 8 + 0 + 0 = 20 bits.

Cache memory size = 2580 Control words

2580 x 20 bits
Size in bytes = = 6450.
8

Q.9 (a)
Average Memory Access time = Hit
Time L1 + Miss rate L1 x (Hit time L2
+ Miss rate L2 x Miss Penalty L2)
= 1 + 4% (10 + 50% x 100) [Miss rate
L1 = (40/1000) x 100= 4%]
Q.6 (c)
= 1 + 4% x 60 [Miss rate L2 = (20/40)
Number of cache lines = 2048 = 211
x100 = 50%]
Number of sets = 211 / 22 = 29 (9 bits
= 3.4 clock cycles
required)
Alternate method,
Block size = Cache line size = 26 (6
Tavg = Hit time L1 + (Miss rate L1 *
bits required)
Miss penalty L1)
Miss penalty L1 = Hit time L2 + (Miss
rate L2 * Miss penalty L2)
= 10 cycles + 50 cycles = 60 cycles
Tavg = (1 + ((40/1000) x 60))
= 3.4 cycles
Tag bits = 10
Tag memory size = Number of tag bits x Q.10 (c)
Number of set x Number of cache line in I1: ACC → MBR
each set I2: IR [Add] → MAR
= 10 x 29 x 22 I3: MBR → M[MAR]
= 10 x 211 bits
= 20480 / 8 bytes
= 2560 bytes LEVEL 3

Q.7 (b) Q.1 (b)

32 - (8 + 4) = 20 tag bits

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 100
 16 × 210 bytes H2)(L1 access time + L2 access time +
Number of lines = = 210 Main memory access time)]
16 bytes
tB = (0.70)(2ns) + (0.3) x [0.85 x 10 ns
lines + 0.15 x (110 ns)] = 8.9ns
Number of sets = 210 / 4 = 256 set = 8 tA + tB = 12 + 8.9 = 20.9
set bits
8 F = 10001111 Mapped to 143rd set. Q.5 (b)
The address division will be given as:
Q.2 (c)

Average access time = 0.8 x 30 ns +

0.18 x 110 ns + (0.02) x (30 + 80 +
22000000) ns = 440.046 µ s = 440 µ s

Q.3 (b)
Number of cache lines = 32KB / 32B
= 1K = 210
Number of sets = 210 / 2 = 29

H1 = 18 / 5 + 0.7 ns
= 3.6 + 0.7 ns = 4.3 ns
Direct mapped cache:

There is no need of multiplexer in

direct mapped cache.
So, H2 = 17 / 5 nsec = 3.4 nsec
Difference = H1 – H2
= 4.3 – 3.4 = 0.9 nsec

Q.4 (20.9)
‘A’ is slower than B.
Cache A:
Average access time = H(Access
time) + (1 – H) (Cache access time +
Memory access time)
tA = 0.92(4 ns) + 0.08 (4 + 100) = 12 ns
Cache B:
Average access time = H1(L1 access) +
(1-H1)[H2 (L2 access time) + (1 –

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 101
 Computer Organization and Architecture

Topics: Instruction Pipelining, I/O interface

-------------------------------------------------------------------------------------------
LEVEL 1 S1: In RISC type system, memory visit
is required for storing partial result
while executing a program.
Q.1 Consider the following statements S2: RISC type processor is designed
S1: Generally I/O mapped I/O with minimum number of processing
technique is used to connect more registers.
number of I/O devices than memory S3: RISC type processor design uses
mapped I/O technique micro-programmed type control unit.
S2: Processor generates memory read S4: Instruction execution cycle requires
(MRD) and memory write(MWR) more number of clock cycles in RISC
control signals while communicating type processor than CISC type
with I/O devices in memory mapped processor.
I/O technique. S5: RISC type processor is designed
S3: In Isolated I/O: the maximum for supporting 10 or more addressing
number of I/O devices that can be modes.
addressed by the processor is 220 when Correct statements are
port address size is 16 bit. (a) S1, S2, S3, S4 & S5
S4: While communicating the I/O (b) S1, S2, S4 & S5
devices in I/O mapped I/O, the (c) S1, S2, S3 & S5
valuable memory space address is (d) None of the above
wasted due to the utilization of this
memory space address to the I/O Q.4 A processor has 65 distinct instructions
devices. and 33 general purpose registers. Size
Correct statements is/are of the instruction code is 32 bit, it has
(a) S2 Only an opcode, three register operands and
(b) S2 and S3 one immediate operand. Assuming that
(c) S2 and S4 the immediate operand is an unsigned
(d) All integer, the maximum value of the
immediate operand is ______.
Q.2 A computer handles several interrupt
sources of which of the following are Q.5 The minimum number of stages in a
relevant for this question. pipelined processor that achieves a
Interrupt from the input device scanner speed of 9.2 for 200 operations is ___
Interrupt form Floppy
Interrupt from CD/DVD Q.6 In a CPU, the type of interrupt which
Interrupt from keyboard permits only one I/O device connection
Which one of these will be handled at is
LEAST priority? (a) Non-vectored interrupt
(a) Interrupt from Floppy (b) Vectored interrupt
(b) Interrupt from Keyboard (c) Maskable interrupt
(c) Interrupt from CD/DVD (d) Non-Maskable interrupt
(d) Interrupt from Scanner
Q.7 Consider the following statements
Q.3 Consider the following statements S1: Hardwired control unit is faster
than micro-programmed control unit

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 102
 S2: Number of control bits required in (a) Data dependency
the horizontal micro- programmed (b) Resource conflict
control unit is log2N bits for ‘N’ (c) Structural dependency
number of control signals. (d) Anti dependency
S3: Base register addressing mode is
used for program relocation
S4: PC relative addressing mode is LEVEL 2
used for reducing the opcode size of
the instruction. Q.1 Consider 4 stage instruction pipeline
S5: Memory indirect addressing mode executed on a system:
instruction execution requires less
number of clock cycles than Register
indirect addressing mode instruction
execution.
In the above, the correct statements are
(a) All
(b) S1, S2, S3 and S4 If all instructions are executed only
(c) S1, S3 and S4 once, what is the throughput of
(d) S1, S2, S4 and S5 system?
(a) 4/9 cycles
Q.8 Consider the following 5 stage (b) 4/10 cycles
pipelined processor. The number of (c) 4/12 cycles
clock cycles needed by the four (d) 4/11 cycles
instructions I1, I2, I3, l4 in stages S1,
S2, S3, S4, S5 is shown below. Q.2 Consider the following statements:
S1: Micro programmed control unit
uses fixed logic to interrupt instruction.
S2: Horizontal microprogramming
control unit requires an additional
hardware (like a decoder)
S3: The overall performance of a
system is directly proportional to the
memory accesses which can be
The minimum number of clock cycles satisfied by the cache.
required to execute the above is ____ Which of the following options is
correct?
Q.9 Which of the following addressing (a) Only S1 and S2 are true
mode is best to specify the target (b) Only S1 and S3 are true
address in branch instruction? (c) Only S3 is true
(a) Direct addressing mode (d) None of the above
(b) Relative addressing mode
(c) Auto decrement addressing mode Q.3 A DMA is transferring characters to
(d) Index addressing mode processor from a device transmitting at
8000 bits per sec. Assume DMA using
Q.10 Which of the following dependency cycle stealing mode. If processor needs
problem can be eliminated by using access to main memory once every
Register Renaming technique?

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 103
 micro second. The percentage Q.6 An instruction pipelined processor has
processor be slow down due to DMA 5 stages where each stage takes 4
activity is______ (in %). nanosec and these five stages are
(a) 0.1 required for executing all stages. Only
(b) 0.3 branch instructions are not overlapped
(c) 0.5 because there is no branch prediction.
(d) 0.7 The instruction after the branch
instruction is not fetched until the
Q.4 Consider a 6 stage instruction pipeline, branch instruction operation is
where all stages are perfectly balanced. completed in 4th stage.
When a program is executing on this Assume that 10% instructions are
pipeline, the speed up achieved with branch instructions; while executing
respect to non-pipeline execution, if the larger number of instructions the
30% instructions incur ‘3’ pipeline stall average instruction time in nano
cycles and 15% of instructions incur seconds is _____
‘2’ pipeline stall cycles is _____
Q.7 Instruction size of the system is 64 bits.
Q.5 Consider the following Instructions (I1 A program starts at address 250. All
to I5), each instruction is represented values are in decimal. Which of the
with three address fields following is a legal program counter
OPR Rd Rs2 Rs1 value?
Operation OPR is performed on the (a) 300
content of Rs1 and Rs2 and the result is (b) 350
placed in the register Rd. (c) 650
I1 ADD R3, R5, R6 (d) 950
I2 MUL R7, R1, R8
I3 SUB R7, R1, R3 Q.8 Assume that for a certain processor,
I4 ADD R9, R5, R6 TCM  ( cache access time ) = 10 ns,
I5 DIV R3, R4, R2
TMM ( main memory access time ) =
Consider the following statements 100 ns and Hit rate is 90% [Note:
S1: There is a true dependency always words to the processor are
between I3 and I5 supplied by the cache memory only].
S2: There is an anti-dependency The average read access time in nano
between I3 and I5 seconds is?
S3: Within an instruction pipeline, an (a) 20
anti-dependency always does not (b) 30
requires stalls (c) 40
Which one of the above statements (d) 60
is/are correct?
(a) S1 only Q.9 A system has 32 bits instruction that
(b) S1 and S3 support zero address, one address and
(c) S2 and S3 two address instructions. Assume each
(d) S2 only address field size is 12 bits, it is
designed for supporting 200 number of
two address instructions, 376 number
of one address instructions. The

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 104
 maximum number of zero address Which of the following is the correct
instructions that can be formulated is__ match between the List-1 and List-2?
P Q R S
Q.10 A DMA controller has two registers (a) (ii) (iii) (iv) (i)
known as address register and (b) (ii) (iv) (i) (iii)
Terminal Count Register (TCR). (c) (ii) (iii) (i) (iv)
Address register is used to hold the (d) (ii) (iv) (iii) (i)
address of the data to be transferred
and TCR is used to hold the size of the Q.3 Consider the two processors P1 and P2
data to be transferred (i.e. number of with intermediate register gateway is 0.
bytes to be transferred) for one DMA P1: Has four stage pipeline with stage
request. Assume that the size of the latency 1.5 nsec, 2 nsec, 1 nsec and 0.5
above registers is 32 bits each and nsec.
DMA is operated in burst transfer P2: Has five stage pipeline with stage
mode. Memory is byte addressable latency 1 nsec, 2.5 nsec, 1.5 nsec, 2
type. The maximum size of the data nsec and 1 nsec.
that can be transferred by the DMA If each processor has infinite number
controller if it generates the DMA of instruction to execute, then which of
request signal 100 times is the following is true?
(a) 400 GB (a) Processor 2 slower than processor 1
(b) 4 GB (b) Both processors have same clock
(c) 3.2 GB rate
(d) 400 MB (c) Processor 1 slower than processor 2
(d) None of the above
LEVEL 3 Q.4 A hard disk with a transfer rate of 1
KBps is constantly transferring data to
Q.1 Consider the following 8 bit memory using DMA cycle stealing
multiplication process (-121) x (-113). mode. The size of the data transfer is
What is the recorded multiplier in the 16 bytes. The processor runs at 400
multiplication? kHz clock frequency. The DMA
(a) -100+1000-1 controller requires 10 cycles for
(b) 100+1000-1 initialization of operation and transfer
(c) -100-1000+1 takes 2 cycles to transfer one byte of
(d) None of these data from the device to the memory.
What is the percentage of processor
Q.2 Consider the floating point Arithmetic time blocked during this DMA
for single precision, match the operation?
following pairs if E' = Biased exponent (a) 0.70 %
and M = Mantissa: (b) 0.65%
(c) 0.50 %
(d) 0.55%

Q.5 Consider a pipeline processor with 5

stages, Instruction Fetch (IF),
Instruction Decode and Operand Fetch

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 105
 (ID), Operation performed (OP), Data
memory access (MA) and Write back
(WB). The IF, ID, MA and WB stages
takes 1 clock cycle each for any
instruction. The OP stage takes 1 clock
cycle for ADD and SUB instructions
and takes 3 clock cycles for MUL
instruction. The minimum numbers of
clock cycles are needed to complete
following sequence of instruction if
operand forwarding is used ______.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 106
 ANSWER KEY:

LEVEL 1

1 2 3 4 5 6 7 8 9 10
(a) (b) (d) (127) (10) (b) (c) (15) (b) (d)

LEVEL 2

1 2 3 4 5 6 7 8 9 10
(d) (c) (a) (2.72) (c) (5.2) (c) (a) (937984000) (a)

LEVEL 3

1 2 3 4 5
(a) (c) (a) (b) (11)

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 107
 SOLUTIONS Type of operand is unsigned integer,
the longest integer that can be
produced with 7 bit is (27-1) = 127
LEVEL 1
Q.5 (10)
tn n×k
Q.1 (a) Speed up = =
‘S1’ is false because memory mapped t p n + k −1
I/O us used to connect more number of n = number of instruction = 200
I/O devices than I/O mapped I/O. k = number of segments = ?
‘S2’ is true. tn = non-pipeline time
‘S3’ is false because I/O mapped I/O is tp = pipeline time
also known as Isolated I/O, in this n×k 200 × k
= 9.2, = 9.2
technique processor can address n + k −1 199 + k
maximum 216 input devices and 216 200k = 9.2 ×199 + 9.2k
output devices when it’s port address
size is 16 bit. 190.8k = 1830.8
‘S4’ is false because valuable memory k = 9.59
space is wasted when the I/O devices Therefore minimum number of
are addressed in memory mapped I/O, segments (stages) is 10.
because in this mode one memory
address is reserved for one I/O device Q.6 (b)
connection. In all processors; Vectored interrupt
permits only one IO device but Non-
Q.2 (b) vectored interrupt permits one or more
Processor gives highest priority for IO devices.
high speed devices (like secondary
memories) and least priority for I/O Q.7 (c)
devices (like keyboard and mouse) S1 is true; S2 is false because
because I/O devices are slower than horizontal micro-programmed control
magnetic and optical memories unit word size is ‘N’ bit for generating
and keyboard is slower than scanner. ‘N’ control signals.
S3, S4 are True. S5 is false because
Q.3 (d) memory indirect addressing mode
All statements are false. instruction requires two memory visits
in execution cycle (except fetch cycle)
Q.4 (127) but register indirect addressing mode
Size of the opcode word = 32 bits instruction requires only one memory
Number of Registers = 33 visit.
Register field size = log2 33 ≅ 6
Number of instruction = 65 Q.8 (15)
Opcode field size = log2 65 ≅ 7

Size of operand is ‘7’ bit

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 108
 since it uses encoded scheme for the
instruction.
• Horizontal microprogramming
control unit does not require an
additional hardware (like a decoder)
because a fixed logic is associated with
the instructions.
• The performance of a system depends
on the direct proportion of memory
accesses satisfied by cache.

Q.3 (a)
8000 bits transmitting → 1seconds
8
1 character takes = seconds
8000
= 1 msec = 1000 µ sec
Processor access main memory in
Q.9 (b) every µ sec.
Relative addressing mode is best to
y
specify the target address in branch Slow % = * 100
instruction. x+ y
x = 1000 µ sec and y= 1 µ sec
Q.10 (d) 1 µ sec
So slow % = *100 = 0.1
Register Renaming is used to remove (1000 + 1) µ sec
output dependency and anti
dependency. Q.4 (2.72)
Non-Pipeline: tn = 6 clocks
LEVEL 2 Pipeline: tp = (0.30*(3+1)) +
(0.15*(2+1)) + (0.55*(1)) = 2.2 clocks
Q.1 (d) Speed up = tn / tp = 6 / 2.2 = 2.72
1 2 3 4 5 6 7 8 9 10 11
S1 I1 I2 I2 I3 I4 I4 Q.5 (c)
S2 I1 I1 I1 I2 I3 I3 I4 S1 is False and S2 is TRUE because
S3 I1 I2 I2 I3 I4 I4 there is an anti-dependency (write after
S4 I1 I2 I3 I3 I4 read) between I3 and I5.
Throughput = S3 is true because anti-dependency
Number of task completed requires stall cycles only some time but
Total time taken to process the tasks not always.
= 4 / 11 cycles
Q.6 (5.2)
While executing large number of
Q.2 (c)
instructions, only one segment time (4
Considering each statement:
ns) is sufficient to complete one
• Micro programmed control unit uses
instruction when there is no hazard.
variable logic to interrupt instruction
For non branch instruction one clock (4
ns) and for branch instruction 3 (stall

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 109
 cycles) clocks and one more clock is Now in these it uses only 376 and
required for execution 12+4=16ns. remaining 229000 free combinations
Average instruction time = (0.9 x 4 ns) are used for zero address instruction.
+ (0.1 xx 16 ns) = 3.6 + 1.6 = 5.2 ns i.e. maximum number of zero address
instruction to be formulated
Q.7 (c) = 229000 x 212 = 937984000
Legal PC value can be (250 + 8x)
where x is an integer variable. (64 bits Q.10 (a)
= 8 Bytes) Size of TCR is 32 bit, i.e. the
250 + 8x ≠ 300 maximum size of the data to be
250 + 8x ≠ 350 transferred for one DMA request is
232 B = 4 GB. Total number of requests
250 + 8x ≠ 950 made the processor is 100. Then
250 + 8x = 650 (x = 50) maximum size of the data that can be
So, ans is 650. transferred = 4 GB x 100 = 400 GB
Q.8 (a)
As given in note; always words are
supplied through CM only
LEVEL 3
Q.1 (a)

TCM = 10ns
TMM = 100ns
Tavg = (H × TCM ) + (1-H)(TMM + TCM )

= [0.9 × 10] ns + [0.1 × 110] ns

= 9 + 11 ns
= 20 ns

Q.9 (937984000)
Size of instruction = 32 bits
Maximum no of 2 address instructions
= 28 = 256 Q.2 (c)

Q.3 (a)
For P1:
Pipeline time = max (1.5, 2, 1 and 0.5)
= 2 nsec
But it uses only 200 instructions, 1
remaining 56 free combinations used Clock rate = = 500 MHz
for one address instruction, then 2 nsec
maximum number of one address For P2:
instructions to be formulated = 56 x 212 Pipeline time = max (1, 2.5, 1.5, 2, 1)
= 229376 nsec = 2.5 nsec

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 110
 1
Clock rate = = 400 MHz
2.5 nsec
So, processor-1 is faster than
processor-2

Q.4 (b)
% Time CPU is blocked =
Transfer Time
Transfer Time + Preparation Time
Preparation Time:
103 bytes → 1 sec
1 byte → 1 msec
So, 16 bytes → 16msec
Transfer time = 10 cycles for
initialization + (2 x 16) cycles for
transfer = 42cycles
1
1 cycle time =
400 KHz
= 0.0025 msec
So, 42cycles = 42 x 0.0025 msec
= 0.105msec
0.105
% of time CPU blocked =
16 + 0.105
= 0.0065 x 100
= 0.65%

Q.5 (11)

So, Total number of clock cycles needed for

given program is 11.

© Copyright Reserved by Gateflix.in No part of this material should be copied or reproduced without permission 111