You are on page 1of 4

​ NPTEL Online Certification Courses

​ Indian Institute of Technology Kharagpur

Cloud Computing
Assignment-Week 5
TYPE OF QUESTION: MCQ/MSQ
Number of questions: 10 Total mark: 10 X 1 = 10

QUESTION 1:
Which of the following is/are the objective(s) of Resource Management?
A) Increased overhead
B) Increased throughput
C) Increased latency
D) Scalability
Correct Answer: B, D
Detailed Solution: From definition of resource management. Reduced overhead and reduced latency are
objectives. Others are correct.

QUESTION 2:
Multiple KPIs are aggregated to SLA.
A) True
B) False

Correct Answer: B
Detailed Solution: KPIs are aggregated to SLO.

QUESTION 3:
In computing, Performance/Watt follows Moore’s law
A) True
B) False
Correct Answer: B
Detailed Solution: In computing, Performance/Watt does not follow Moore’s law. Refer slide 7 of Resource
Management-I.

QUESTION 4:
Which of the following is/are resource adaptation approaches?

A) Reinforcement learning guided control policy


B) Intelligent multi-agent model
C) Network queueing model
​ NPTEL Online Certification Courses

​ Indian Institute of Technology Kharagpur

D) Web-service based prototype

Correct Answer: A, D

Detailed Solution: Reinforcement learning guided control policy and web-service based prototype are
resource adaptation approaches. The Intelligent multi-agent model is a resource allocation approach.
Network queueing model is a resource provisioning approach.

QUESTION 5:
A third party application runs in the cloud for 18 hours/day. At the end of one month (30 days), it was found
that the cloud suffered outages totaling 6 hours and T hours, on different days over the service period. The
cloud guarantees service availability for 98% of the time. What are the value(s) of T among the given
options such that the SLA negotiation does NOT get honored in terms of service availability?
A) 3 hours
B) 5 hours
C) 6 hours
D) 4 hours

Correct Answer: B, C
Detailed Solution: Total Outage: (6+T) hours, application runs for 540 hours in a month. Availability = 1 -
(downtime/uptime). For availability: [1- {(6+T)/(534-T)}] >= 0.98, T<=4.59 hours.
Options B and C are correct as the SLA negotiation does not get honored.

QUESTION 6:
Consider that the peak computing demand for an organization is 200 units. The demand as a function of time
can be expressed as D(t) = 8t. Baseline (owned) unit cost is 120 and cloud unit cost is 150. The cloud is
owned for a period of T time units. Select the values of T for which cloud is cheaper than owning.
A) 25
B) 30
C) 35
D) 45

Correct Answer: A, B, C
Detailed Solution: Total baseline cost BT = P ⨯ B ⨯ T = 200*120*T = 24,000*T units.
𝑇 𝑇 𝑇
2
𝑡
Total cloud cost CT = ∫ 𝐶 * 𝐷(𝑡)𝑑𝑡 = ∫ 150 * 8𝑡 𝑑𝑡 = 1200* [ 2
] = 600*(T2) units
0 0 0
2
Utility function UT = (CT/BT) = 600*T /24,000*T = T/40.
​ NPTEL Online Certification Courses

​ Indian Institute of Technology Kharagpur

For T < 40 units, cloud is cheaper than owning. For the other cases, cloud is costlier than owning.

QUESTION 7:
Which of the following is/are true regarding penalty cost? (Here D(t) and R(t) are instantaneous demand and
resources at time t.)

A) If demand is flat, the penalty is equal to 0.

B) Penalty cost ∝ ∫|𝐷(𝑡)/𝑅(𝑡)|𝑑𝑡

C) If demand is exponential (D(t)=et), any fixed provisioning interval (tp) according to the current
demands will fall linearly behind.
D) The penalty cost for exponential demand is exponential.

Correct Answer: A, D

Detailed Solution: Penalty cost ∝ ∫|𝐷(𝑡) − 𝑅(𝑡)|𝑑𝑡. If demand is exponential (D(t)=et), any fixed

provisioning interval (tp) according to the current demands will fall exponentially, NOT linearly behind.

QUESTION 8:
In resource management, _____________ is determining when an activity should start or end, depending on
its duration, predecessor activities, predecessor relationships etc.

A) Resource allocation
B) Resource mapping
C) Resource scheduling
D) Resource modeling

Correct Answer: C

Detailed Solution: Resource scheduling is determining when an activity should start or end, depending on
its duration, predecessor activities, predecessor relationships and resources allocated, from slide-10 of
resource management-II. So, the correct option is C.
.

QUESTION 9:
Which of the following techniques can be used to conserve energy within a Cloud environment?

A) Schedule VMs to conserve energy.


B) Optimize data center design.
C) Maximize operating inefficiencies for non-essential tasks.
D) Management of both VMs and underlying infrastructure.

Correct Answer: A, B, D
​ NPTEL Online Certification Courses

​ Indian Institute of Technology Kharagpur

Detailed Solution: Refer to slide-8 of Resource Management-I. The operating inefficiencies for
non-essential tasks should be minimized.

QUESTION 10:

An app collects the ratings of restaurants in a particular location and logs it in the following form: <userId,
restaurantId, rating>. ‘restaurantId’ takes any of the values (A, B, C, D) and each user having a unique userId
can provide [0,5] rating. There are three files and each file consists of three columns: userId, restaurantId and
rating. Each file may consist of the data for the same user or same restaurant multiple times. Now, out of this
data, we need to calculate the number of times each restaurant gets a “5 star” rating from users.

The MapReduce framework will divide it into three map tasks and each map task will perform data functions
on one of the three files. The inputs are as shown below.

M1: {(u1,A,5); (u1,B,3); (u2,A,5); (u3,C,5); (u3,D,2); (u7,D,5) }

M2: {(u1,C,3); (u1,D,5); (u2,C,5); (u2,D,5); (u4,D,5); (u6,A,5) }

M3: {(u1,A,5); (u2,A,5); (u2, B,5); (u3,B,5); (u4,C,5); (u3,D,5) }

The tasks of all mappers (M1, M2 and M3) are passed to the reducer job. Reducer will output:

(a) <A, 5>; <B, 5>; <C, 5>; <D, 5>


(b) <A, 5>; <B, 2>; <C, 3>; <D, 5>
(c) A; B; C; D
(d) <A, 5>; <D,5>

Correct Answer: B
Detailed solution: The reducer function aggregates the number of times each restaurant gets the highest
rating (5) from all the mappers’ output. It is observed that option B is correct after counting the number of
highest ratings each restaurant received.

You might also like