BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE, Pilani

Pilani Campus
AUGS/ AGSR Division

FIRST SEMESTER 2019-20

COURSE HANDOUT
Date: 02.08.2019

In addition to part-I (General Handout for all courses appended to the time table) this portion gives
further specific details regarding the course

Course No. : CS / SS G527

Course Title : Cloud Computing
Instructor: Shan Sundar Balasubramaniam (email: sundarb)
Hari Babu K ([email protected])

1. Scope and Objective:

The primary objective of the course is to introduce the student to cloud computing from architectural and
design perspectives. As such, the emphasis of the course would be on the underlying infrastructure and
architecture of clouds, techniques for enabling services and the quality of such services, as well as issues in
designing clouds. Select research issues in performance, security, and management would also be
addressed. Programming on the cloud would be encouraged but not taught in class. Students are expected
to learn and understand tools and techniques for using, designing, and implementing clouds and services via
assignments and a term project.

2. Text and References:

a.Text Book: NONE

b.References:
R1. Kai Hwang , Jack Dongarra , Geoffrey C. Fox Distributed and Cloud Computing: From Parallel
Processing to the Internet of Things. Morgan Kauffman 2011.

R2. Jim Smith, Ravi Nair. Virtual Machines: Versatile Platforms for Systems and Processes. Morgan
Kaufmann. 2005

AR. Additional references (papers) to be posted online.

3. Course Plan:

3a. Modules

Mo Theme Learning Objectives

d-
ule

I Cloud Computing –  To understand the motivation for Cloud Computing.

Introduction,  To understand the underlying (distributed) computing model.
Principles, and Issues.

1
 BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE, Pilani
Pilani Campus
AUGS/ AGSR Division

II Cloud Architecture –  To understand how to leverage and provision computing

Resources and resources at different levels of abstraction.
Virtualization  To understand virtualization techniques at different levels of
abstraction.
 To understand how to architect a cloud to suit different
requirements
III Programming for the  To understand the execution of applications on the cloud
Cloud and Application  To understand how to develop & deploy applications for the
Models cloud and the relevant tools & technologies

IV Services, Service  To understand how to use the cloud to deliver Software as a

Models, and QoS. Service.
 To understand how to deliver computing Infrastructure (e.g.
processors, storage, network) as a Services
 To understand Quality of Service issues and QoS support
mechanisms for Services on the Cloud
V Big Data  To understand techniques and tools for handling large volumes
of data;
 To understand the structure and issues of file systems and
databases for large volumes of data.
VI Cloud Management,  To understand how to manage a cloud platform and a services
Performance and ecosystem
Security Issues  To understand performance issues and techniques to enable
performance of a cloud at different levels of abstraction
 To understand security issues specific to cloud computing and
solutions to address them.

[Note: Module III will not be covered in detail in class. It is to be learnt primarily through assignments
and project. End of Note.]

3.b Lecture Schedule:

Lecture Module Topics Readings

1 Evolution of Computing Systems – Mainframes to PCs to R1 Sec. 1.2

Networked Systems to Clouds. Cloud Computing and Services – and 1.4.1
User’s Perspective: Economic / Business Motivation.
I
2 Spectrum of Parallel and Distributed Computers – Programmer’s R1 Sec. 1.3.4
Perspective. Forms of Parallelism: Data Parallelism vs. Task
Parallelism vs. Request Parallelism.

2
 BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE, Pilani
Pilani Campus
AUGS/ AGSR Division

3 Distributed Computing Models – Message Passing, Failures, and

NONE
Buffering.

4-6 Introduction to Computer Clusters: Structure and Components of a R1 Sec. 2.1 to

cluster – nodes, interconnect, and middleware. 2.3, AR
Design of a cluster: Single System Image, Design for availability,
Fault-tolerance and Fault Recovery - Reliability and Availability.
7 Computer Clusters: Task Management and Resource Management; R1 Sec. 2.4,
Scheduling and Load Balancing. AR

8 High Availability Clusters: Design Issues. Check-pointing and R1. Sec. 2.3.2
Recovery. to 2.3.4
9-12 Programming on Clusters- Speedup and Scalability; Scale-out R1 Sec. 1.4.3
Clusters: Map-reduce Architectures; In-memory Distributed and 6.2.2, AR
III Programming; Platform Architectures and Programming Models
13 Building Microservices – Programming Models and Platform
Example
14 Virtualization: Introduction, Different Levels of Abstraction, and R2 Ch. 1, AR
Mechanisms for Virtualization. Process vs. System VMs, HLL VMs.

15 Process VMs: Structure and Behavior. R2 Sec. 3.1,

AR
16 System VMs : Applications and Implementation Models. R2 Sec. 8.1

17 System VMs: Resource Virtualization and ISA Virtualizability R2 Sec. 8.2,

AR
18 II System VMs: Virtualization of Memory and I/O R2 Sec. 8.3-
8.4
19 Case Studies: VMWare/Xen AR

20-21 Storage Systems and Storage Virtualization: Storage Devices, File AR

Systems and Volumes, Storage Networks – NAS and SAN, NAS
Internals, Virtual Storage and Storage on the Cloud. (Case Study:
DropBox / Google Drive)
22-23 Virtualized Networks and Virtual Clusters. Process Migration and R1 Sec. 3.4,
VM Migration. Live Migration – Suspend and Resume. R2 Sec. 10.2,
AR

24 SDN in the Data Center AR

25-26 Containers – Light-Weight VMs, Comparison with VMs; Case Study AR

(Dockers/LXC). Scheduling and Load Balancing of VMs and
Containers

27 Elasticity AR

3
 BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE, Pilani
Pilani Campus
AUGS/ AGSR Division

28 IV Services on the Cloud. Software as a Service; SOA; REST; Web R1 Sec. 5.1,
Services; AR
29- 30 Infrastructure as a Service (IaaS), Storage as a Service (StaaS) AR
Platform as a Service – Features, Programming Models (e.g.
Serverless Architectures, AWS MapReduce, Amazon EBS, S3,
SimpleDB)

31 Resource Scaling and Capacity Management: Managing VMs AR

Resource Provisioning. Quality of Service – Models, Parameters,
and Metrics. SLAs.

32-33 V Big Data: Consistency Protocols- 2-Phase and 2-Phase Commit AR

Protocol, Paxos Protocol. Case Study: Chubby locking

34-35 Big Data: Consistency Protocols – Implementation and Practical AR

Platforms. Case Studies: Zookeeper, , Raft

36 Big Data File Systems- Case study: Google FS / Hadoop Distributed AR

FS;

37-38 Big Data: CAP Theorem; PACELC theorem; NoSQL Databases. AR

39 Handling Large Data: Big Data Issues; Programming and Storage

Models: GFS/HDFS, NoSQL, Google BigTable, Google Spanner, /
AWS DynamoDB etc.

40-41 VI Cloud Security AR

42 Cloud Economics – Service Pricing AR

3.c. Practicum

Practical work by students is expected to carried out at the students’ own convenience via
assignments and term project. The assignments are primarily meant for students to learn
implementation skills with some design. On the other hand, term project is meant for the student to
explore a problem and/or solution through readings, carry out a new design, and implement as well.

Both assignments and term project will require students to work in a team, set up
developmental / experimental platform(s) and conduct performance studies on their
implementations. The term project will also require students to articulate their solutions and results
through interactions with the instructor and through design / implementation documents.

4. Evaluation
4. a. Evaluation Scheme:

4
 BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE, Pilani
Pilani Campus
AUGS/ AGSR Division

Component Weight Date & Time Remarks

Mid-Term Test 20% <TEST_1> Open Book

Comprehensive Exam 30% <TEST_C> Open Book

Class Tests / Quizzes 15% TBA Small exercises either take-

home or In-class

Assignments / Term 35% TBA Programming, Take Home

Project (3)

4. b. Make-up Policy:
- Assignments / Term Project:
 No Make-up will be available for assignments or term project. Late submissions will be evaluated at
25% less weight for that component for a delay of up to 24% hours after which no submissions will
be accepted.
- Test:
 Prior Permission of the Instructor-in-Charge is usually required to get a make-up for a test/quiz.
 A make-up test shall be granted only in genuine cases where - in the Instructor’s judgment - the
student would be physically unable to appear for the test. Instructor’s decision in this matter would
be final.
- Comprehensive Exam:
 Make-up for the comprehensive exam may be applied only with the Associate Dean of Under-
Graduate Studies or Associate Dean of Graduate Studies / Research.

4.c. Fairness Policy:

- Students are expected to work within their team on assignments / project expect where explicitly
instructed / permitted otherwise.
- When students are allowed to consult/discuss with other students/teams such
consultation/discussion should be explicitly acknowledged and reported to the instructor prior to
evaluation.
- When students are expected to collaborate within a team:
o Individual contributions should be identified and documented in qualitative and
quantitative terms by each team member.
o Instructor may assess and mark each individual in a team separately.
o The instructor’s assessment of the contributions in this matter would be final.
-

5. Consultation Hours: To be notified (via the course website on nalanda)

6. Notices: All notices concerning this course will be displayed online only (i.e on the course website). If there is a
need, email would be used on short notice (12 hours) – only BITS Pilani email would be used.

5
BIRLA INSTITUTE OF TECHNOLOGY AND SCIENCE, Pilani
Pilani Campus
AUGS/ AGSR Division

Instructor –In- Charge

CS / SS G527