Scribd
Upload
18 views3 pages

Unit1 Aos

The document outlines various architectures of distributed systems, including client-server, peer-to-peer, and service-oriented architectures. It discusses system architecture types, issues in distributed operating systems, communication networks, and theoretical foundations such as Lamport's logical clocks and vector clocks. Additionally, it covers distributed mutual exclusion algorithms and their comparative performance analysis.

Uploaded by

Sandhya
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
18 views3 pages

Unit1 Aos

The document outlines various architectures of distributed systems, including client-server, peer-to-peer, and service-oriented architectures. It discusses system architecture types, issues in distributed operating systems, communication networks, and theoretical foundations such as Lamport's logical clocks and vector clocks. Additionally, it covers distributed mutual exclusion algorithms and their comparative performance analysis.

Uploaded by

Sandhya
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Unit-1

Architectures of Distributed Systems

Distributed systems consist of multiple independent computers that work together to provide a
seamless computing experience. The architectures include:

1. Client-Server Model – Clients request services from a centralized server.

2. Peer-to-Peer Model – Nodes communicate directly without a central authority.

3. Three-Tier Architecture – Separates the presentation, application logic, and data storage
layers.

4. N-Tier Architecture – Extends the three-tier model for scalability and modularity.

5. Service-Oriented Architecture (SOA) – Uses web services for communication between


distributed components.

System Architecture Types

1. Monolithic Systems – Single large application with all functionalities.

2. Layered Systems – Divides functionality into hierarchical layers.

3. Microservices Architecture – Breaks applications into independent services communicating


via APIs.

4. Event-Driven Architecture – Components communicate through event messages.

5. Shared-Memory and Message-Passing Models – Defines how nodes interact within a


system.

Issues in Distributed Operating Systems

 Concurrency and Synchronization – Managing multiple processes executing simultaneously.

 Fault Tolerance – Ensuring system reliability despite node failures.

 Security – Protecting data and communications from unauthorized access.

 Scalability – Handling increasing workloads effectively.

 Resource Management – Efficient scheduling and allocation of resources.

Communication Networks in Distributed Systems

 Local Area Network (LAN) – High-speed networks connecting nearby devices.

 Wide Area Network (WAN) – Covers large geographical areas (e.g., the internet).
 Wireless Networks – Wi-Fi, Bluetooth, and cellular networks.

 Interprocess Communication (IPC) – Includes message passing and shared memory.

Communication Primitives

 Send/Receive – Basic message-passing functions.

 Remote Procedure Call (RPC) – Invokes procedures on remote machines as if local.

 Message Queues – Stores messages asynchronously for decoupled communication.

 Publish-Subscribe Model – Components publish events consumed by subscribers.

Theoretical Foundations of Distributed Systems

Covers fundamental principles like:

 Concurrency and Synchronization – Managing shared resources among distributed


processes.

 Distributed Consensus – Ensuring agreement among nodes (e.g., Paxos, Raft).

 Leader Election – Selecting a leader node dynamically.

Inherent Limitations of a Distributed System

 Network Delays – Unpredictable transmission times.

 Lack of Global Clock – Nodes operate asynchronously.

 Partial Failures – Some nodes may fail while others continue running.

 CAP Theorem – Trade-offs between Consistency, Availability, and Partition tolerance.

Lamport’s Logical Clocks

A logical clock mechanism where each process maintains a counter. It follows two rules:

1. Increment the counter before sending a message.

2. Set the counter to the maximum of the local and received timestamp plus one.

This helps order events in a distributed system.

Vector Clocks

An extension of Lamport clocks where each process maintains a vector of timestamps, allowing for
detection of causality violations.
Causal Ordering of Messages

Ensures that if event A causally influences event B, then A is delivered before B. Achieved using
vector clocks.

Global State & Cuts of a Distributed Computation

 Global State – The collective state of all processes at a given instant.

 Cuts – A snapshot of process states taken at different times to analyze system behavior.

Termination Detection

Mechanisms to determine if a distributed computation has completed execution. Algorithms include:

 Dijkstra-Scholten Algorithm

 Chandy-Lamport Algorithm

Distributed Mutual Exclusion

Ensures that only one process accesses a shared resource at a time.

Classification of Mutual Exclusion Algorithms

1. Centralized Algorithm – A coordinator grants access.

2. Token-Based Algorithms – A unique token allows access.

3. Timestamp-Based Algorithms – Requests are ordered based on timestamps.

Comparative Performance Analysis of Mutual Exclusion Algorithms

Message
Algorithm Fault Tolerance Performance
Complexity

2 messages per Coordinator failure leads to


Centralized Simple and fast
request system failure

Ricart-Agrawala 2(N-1) messages per


More resilient but complex Higher overhead
(Timestamp-Based) request

1 message per Token loss requires Efficient but needs token


Token-Based
request regeneration management

You might also like