Birla Institute of Technology & Science, Pilani
Hyderabad Campus
Computer Science and Information Systems Department
First Semester 2020-2021 Course Handout (Part II)
Date: 02.11.2020
In addition to Part-I (General handout for all courses appended to the timetable) this portion gives further specific details
regarding the course:
COURSE NO.: CS G623
Advanced Operating Systems
Instructor In-Charge: Prof. G Geethakumari
Scope:
Over the last few decades, considerable amount of research has been done in “Distributed OS”. The aim of this course is
to introduce the design and implementation issues of Distributed OS. Distributed OS’s work in an environment where we
have independent machines (both hardware and software) connected with each other over a computer network.
Distributed OSs have at their center the reasoning that you should use faster machines for more tasks that need speed, and
slower ones for the tasks that don't. Also, central to the design of distributed OS's is making this design transparent to the
user. A Distributed OS makes a Distributed System a virtual uniprocessor system.
Objectives:
The distributed OS to be studied in this course is microkernel based. It's just that the user level processes that are
separated from the kernel can run on remote machines.
Few case studies like Sun NFS, HDFS, GFS, MapReduce, Vector Clocks, Causal ordering, Agreement protocols,
Mutual Exclusion, distributed file systems etc. shall be discussed and also implemented as part of the coding
practices in the laboratories.
Text Book:
T1 M. Singhal & N. Shivaratri, “Advanced Concepts in Operating Systems: Distributed, Database and Multiprocessor
Operating Systems”, Tata McGraw Hill, 2001.
Reference Books:
R1 Distributed Operating Systems – The Logical Design by A. Goscinski, AW
R2 Modern Operating Systems by A. S. Tanenbaum, PHI
R3 Distributed Systems-Concepts and Design by G. Coulouris, AW
Plan of Study
S.No. Learning Objectives TOPIC CHAPTER Lect
REF (Text)
1. Why do we develop distributed What is a distributed system, Characteristics of Chapter 1 2
operating systems? a distributed system, Challenges in building
distributed systems, Examples: Planetlab,
Aadhaar, GARUDA, SETI@home etc.
2. Why do we need to have a Traditional Vs Virtualized architecture, Class notes 2
virtualized environment, and what Different types of hypervisors, Need for
are its’ characteristics? virtualization, Uses of Virtual machines
3. Why do we need logical clocks in Limitations of a distributed system, Lamport’s Chapter 5 4
building a distributed system? How logical clock, Fidge-Mattern’s vector clock,
does causal ordering of messages causal ordering of messages, global state of a
help? distributed system, Cuts.
1/2
4. How to solve the problem of critical Lamport’s DME, Ricart-agrawala, and Chapter 6 4
section in a distributed Maekawa’s algorithms; Suzuki-kasami
environment? broadcast algorithm, and Raymond’s tree based
algorithm.
5. Why deadlocks are seen as Resource Vs. Communication deadlock, Chapter 7 4
challenges in building distributed Strategies to handle deadlock, Ho-
systems? How to handle those Ramamoorthy algorithm, Path-Pushing, Edge-
without using avoidance algorithms Chasing, Diffusion Computation-based
like Bankers? algorithms.
6. What is the importance of consensus System model, Classification of agreement Chapter 8 4
in distributed systems? problems, Solutions to Byzantine agreement
problems.
7. Distributed transparent storage: how Bigdata characteristics, Structured vs. Chapter 9 6
to access data stored at different unstructured data, Mechanisms for building and Class
machines? What are the DFSs, Design Issues, Case studies: Sun DFS, notes.
characteristics of Bigdata? Sprite DFS, Hadoop Distributed File System
(HDFS): Hadoop cluster, HDFS
communications, HDFS read/write operations.
8. Why do we need to run a single Issues in Load Distribution, Components of a Chapter 11 6
computation at multiple places? load distribution algorithm, Load Distribution and Class
Algorithms, Case studies: System-V, Sprite, notes.
and Condor, HDFS MapReduce: Programming
on Hadoop.
9. Research paper discussions. Recent research on DFS and Distributed IEEE/ACM 2
schedulers.
9. How to build a logical memory Algorithms for implementing Distributed Chapter 10 4
consisting of RAM contents of Shared Memories (DSMs), Memory Coherence
different machines? models, and Coherence Protocols, Case study:
Integrated Virtual Memory at Yale (IVY).
10. How to recover from failures in a Classification of failures, Synchronous and Chapter 12 4
distributed computation? Asynchronous Check Pointing and Recovery
algorithms.
Evaluation Scheme:
Sl No. Component & Nature Duration Weightage Date and Time
1. Mid Sem Test 90 mins 35% As announced in the
Timetable
2. Lab Tests/Assignments (2 nos.) 35% To be announced in CMS
3. Comprehensive Exam 120 mins 30% As announced in the
Timetable
Note: Coding labs for the course will be based on the design aspects of various components of Distributed Operating
Systems like distributed middleware, thread synchronizations, logical clock implementations, distributed file systems, and
distributed scheduling or load balancing etc. using HDFS/ MapReduce platforms.
Notices: Will be put up in CMS
Consultation Hour: Will be announced in the class.
Academic Honesty and Integrity Policy: Academic honesty and integrity are to be maintained by all the students
throughout the semester and no type of academic dishonesty is acceptable.
Instructor-in-charge, CS G623
2/2
