MSc:DistributedSystems

Distributed Systems

• Course name: Distributed Systems
• Course number: XYZ

Course Characteristics

Key concepts of the class

• Principles of designing and implementing distributed systems
• Structure of distributed system, and algorithms used to implement distributed system

What is the purpose of this course?

This is an introductory course in distributed systems. During the course, students will learn the fundamental principles and techniques that can be applied to design and develop distributed systems, and will cover architectures of distributed systems, communication, naming, fault tolerance, consistency replication, virtualization and security. In addition, we will discuss different aspects of design and implementation of popular distributed systems (such as bittorrent, google file system, HDFS, etc.), programming models (MapReduce, MapReduce2/YARN) and consensus algorithms (Raft and Paxos). The course will not only cover computer science related topics, but will also include a substantial part of software engineering activities.

- What should a student remember at the end of the course?

By the end of the course, the students should be able to understand distributed systems

• Key principles involved in designing and implementing distributed systems
• Distributed Architectures and their management
• Canonical problems and solutions for distributed systems
• Resource sharing, replication and consistency algorithms in distributed environments
• Virtualization, Orchestration and cloud management

- What should a student be able to understand at the end of the course?

By the end of the course, the students should be able to understand the key components of distributed systems

• Role of different components in distributed systems
• Role of virtualization in distributed systems
• Management and Orchestration of resources in distributed systems
• Fault tolerant approaches in distributed systems

- What should a student be able to apply at the end of the course?

By the end of the course, the students should be able to develop and implement different components in a distributed environment

• Implementation of access and location transparency in distributed systems
• Designing and implementing consensus algorithms for different distributed environments
• Management of concurrent tasks in distributed settings
• Designing fault tolerant distributed systems

Course evaluation

Course grade breakdown

		Proposed points
Labs/seminar classes	20	20
Interim performance assessment	30	25
Exams	50	55

Grades range

Course grading range

		Proposed range
A. Excellent	90-100	90-100
B. Good	75-89	75-89
C. Satisfactory	60-74	60-74
D. Poor	0-59	0-59

Course Sections

The main sections of the course and approximate hour distribution between them is as follows:

Course Sections

Section	Section Title	Teaching Hours
1	Introduction to distributed systems	16
2	Virtualization and cloud computing	16
3	Canonical Problems and Solutions	24

Section 1

Section title:

Introduction to distributed systems

Topics covered in this section:

• Distributed architectures
• Types of distributed systems
• Processes & Threads
• Multiprocessor and distributed scheduling
• Communication in distributed systems
• Naming in distributed systems

What forms of evaluation were used to test students’ performance in this section?

|a|c| & Yes/No
Development of individual parts of software product code & 1
Homework and group projects & 1
Midterm evaluation & 1
Testing (written or computer based) & 1
Reports & 0
Essays & 0
Oral polls & 0
Discussions & 1

Typical questions for ongoing performance evaluation within this section

1. State advantages of object based architectures over layer based architectures in distributed system?
2. State advantages of the random walk approach over the flooding approach in unstructured P2P networks for locating the data?
3. Why global variables are not allowed in a RPC?
4. Describe approaches for a server to handle incoming socket connections (at least two).
5. What are pros and cons of different approaches of sockets? When would you choose which one?

Typical questions for seminar classes (labs) within this section

1. Find the process ID (PID) of the running program and list of threads associated to that process?
2. Implementation of new threads
3. Performance analysis of given task
4. Implementation of Process and Global Interpreter Lock (GIL)

Test questions for final assessment in this section

1. Explain difference between threads and process
2. Define Global Interpreter Lock (GIL) and its role
3. What are remote method invocation (RMI). Please state the benefits and challenges of RMI.
4. Define the role of Marshaling and Unmarshaling in RPC
5. State at least two benefits of both Iterative and recursive naming resolution schemes.

Section 2

Section title:

Virtualization and cloud computing

Topics covered in this section:

• Foundations of virtualization
• OS-level virtualization
• System level virtualization
• Memory virtualization
• Cloud and data centres

What forms of evaluation were used to test students’ performance in this section?

|a|c| & Yes/No
Development of individual parts of software product code & 1
Homework and group projects & 1
Midterm evaluation & 1
Testing (written or computer based) & 1
Reports & 0
Essays & 0
Oral polls & 0
Discussions & 1

Typical questions for ongoing performance evaluation within this section

1. What are containers? State at least three benefits of using them?
2. Why memory reclamation approaches are needed for virtual machines?
3. What are “cgroup” and “namespace” subsystems. How can we use the “cgroup” subsystem?
4. Briefly explain about the “Ballooning” approach and its working to reclaim VM memory?
5. What is a Unikernel? State at least two benefits and one drawback.

Typical questions for seminar classes (labs) within this section

1. Hands on Amazon Elastic Compute Cloud (EC2)
2. Set up a public address in Elastic IPs tab and assign it to the web server on which WordPress is hosted.
3. Configure WebServer so that your blog page is accessible only through the public IP address (without specifying the /wordpress path). For example, you specify a public IP address in the address bar of the browser and gain access to the blog.
4. There may be problems with CSS. Search the web for how to fix broken CSS after changing the site URL.

Test questions for final assessment in this section

1. State two benefits of immersion cooling over traditional air cooling in data centers?
2. State one main difference of a container from a System Virtual machine?
3. Briefly explain about Copy-on-write storage.
4. What are the roles of “ISA” and “ABI” in operating systems?
5. What is the biggest drawback of Google File system over CephFS?

Section 3

Section title:

Canonical Problems and Solutions

Topics covered in this section:

• Mutual exclusion
• Leader election,
• Clock synchronization
• Consistency issues
• Caching and replication
• Fault Tolerance

What forms of evaluation were used to test students’ performance in this section?

|a|c| & Yes/No
Development of individual parts of software product code & 1
Homework and group projects & 1
Midterm evaluation & 1
Testing (written or computer based) & 1
Reports & 0
Essays & 0
Oral polls & 0
Discussions & 1

Typical questions for ongoing performance evaluation within this section

1. Why is Anti-Entropy Protocol considered better than the Gossiping protocol for achieving consistency.
2. What is the role of Heartbeats in RAFT?
3. Explain PAXOS to achieve consistency?
4. What is the benefit of three phase commit over two phase commit? Please specify in terms of coordinator failure?
5. How many replicas are required to identify the fault in Byzantine failure scenarios?

Typical questions for seminar classes (labs) within this section

1. In the replication set of databases, what problem may appear if we have an even number of nodes?
2. ICRUD operations in MongoDB.
3. Create simple chat web-application which uses replica set.
4. Explain the steps to shutdown all VPS instances.

Test questions for final assessment in this section

1. What is a schema-less data model
2. Why do we need NoSQL? What are its benefits over SQL in terms of ACID and BASE properties?
3. Why is version control recommended? Name at least two version control systems?
4. Role of recovery line and check pointing in distributed snapshot?
5. Explain the difference between Lamport and vector clocks?
6. How the use of buffers at receivers enhance the QoS in networks for streaming media applications? Briefly explain using an example.