(SEM VII) THEORY EXAMINATION 2018-19 DISTRIBUTED SYSTEMS

SECTION A

(Attempt all questions in brief – 2 × 10 = 20 marks)

(a) Web challenges involved in Distributed Systems

Web-based distributed systems face challenges such as heterogeneity, because different hardware, software, and platforms must work together. Scalability is another challenge, as systems must support a growing number of users. Security is critical because data travels over networks. Latency and network failures also affect performance and reliability.

(b) System Model in Distributed Systems

A system model describes how a distributed system is structured and how it behaves. It includes the physical model (hardware and network), architectural model (client-server, peer-to-peer), and fundamental model (interaction, failure, and security assumptions).

(c) Distributed Deadlock

A distributed deadlock occurs when processes running on different computers wait indefinitely for resources held by each other. Since no global clock or shared memory exists, detecting deadlock in distributed systems is more complex than in centralized systems.

(d) Commit Protocol

A commit protocol ensures that all participating nodes in a distributed transaction either commit or abort together. It maintains data consistency across sites. The most common example is the Two-Phase Commit (2PC) protocol.

(e) Transparency in Distributed Systems

Transparency hides system complexity from users. Types include access transparency, location transparency, replication transparency, and failure transparency, allowing users to see the system as a single unified resource.

(f) Scalability

Scalability refers to a system’s ability to handle increased load by adding resources. Distributed systems achieve scalability through replication, load balancing, and decentralization.

(g) Heterogeneity

Heterogeneity means a distributed system consists of different hardware, operating systems, networks, and programming languages. Middleware helps manage this diversity.

(h) Fault Tolerance

Fault tolerance is the ability of a system to continue functioning even if some components fail. This is achieved using redundancy, replication, and failure detection mechanisms.

(i) Consistency

Consistency ensures all users see the same data at the same time. Distributed systems use different consistency models such as strong consistency and eventual consistency.

(j) Middleware

Middleware is software that lies between applications and operating systems. It provides services such as communication, security, and transaction management in distributed systems.

SECTION B

(Attempt any one – 10 marks)

(a) Distributed Mutual Exclusion vs Single System Mutual Exclusion

In a single computer system, mutual exclusion is achieved using shared memory and locks. In distributed systems, there is no shared memory or global clock, so processes communicate using messages.

Distributed mutual exclusion algorithms use message passing to ensure only one process enters the critical section at a time. Performance is measured by message complexity, synchronization delay, and system throughput. Examples include Lamport’s algorithm and Ricart–Agrawala algorithm.

(b) Sequential Consistency and Highly Available Services

Sequential consistency ensures that operations appear to execute in some sequential order that is consistent across all nodes. It provides a strong consistency guarantee but may reduce performance.

Highly available services ensure that systems remain operational even during failures. This is achieved using replication, failover mechanisms, and load balancing to minimize downtime.

SECTION C

(Attempt any one – 10 marks)

(a) Distributed File System Architecture & Distributed Shared Memory

A Distributed File System (DFS) allows files to be stored on multiple machines while appearing as a single file system to users. Its architecture includes clients, file servers, and naming services. The DFS manages file access, consistency, and security.

Distributed Shared Memory (DSM) allows processes on different machines to share memory as if it were local. The DSM algorithm manages memory consistency, page replacement, and synchronization using message passing.

(b) Byzantine Agreement Problem

The Byzantine Agreement Problem deals with achieving consensus in a distributed system where some nodes may fail or act maliciously. The challenge is to ensure all non-faulty nodes agree on the same value.

The solution involves Byzantine fault-tolerant algorithms, which require at least 3f + 1 nodes to tolerate f faulty nodes. These algorithms use majority voting and message verification to reach agreement.