THEORY EXAMINATION (SEM–VIII) 2016-17 DISTRIBUTED SYSTEM

SECTION – A

Explain the following (Short Answer Type)

Lamport’s logical clock helps order events in distributed systems, but it cannot capture the exact cause-and-effect relationship between events. It only provides partial ordering and cannot detect concurrent events accurately.

There is no global clock in distributed systems because each computer operates independently with its own local clock, and network delays make it impossible to synchronize perfectly.

Distributed systems have limitations such as network failure, security risks, data inconsistency, and complex system management.

Termination detection refers to identifying when all processes in a distributed system have completed their tasks and no further activity remains.

Distributed deadlock detection algorithms include centralized, distributed, and hierarchical detection methods.

A process is a program in execution, while a thread is a smaller unit of execution within a process that shares resources.

Phantom deadlock occurs when a system incorrectly detects a deadlock due to outdated or incomplete information.

A digital signature is an electronic authentication method that verifies the identity of the sender and ensures data integrity.

A fault is an error in a system component, while failure occurs when the system cannot perform its intended function.

The transport layer provides a security handshake to initiate TCP/IP connections.

Questions (Section A)

What are the shortcomings of Lamport’s logical clock?

Why is there no global clock in distributed systems?

What is termination detection?

Define digital signature.

Differentiate between fault and failure.

SECTION – B

Attempt any five (Long Answer Type)

Lamport’s logical clock assigns timestamps to events to maintain their order in distributed systems. It ensures that if event A happens before event B, then the timestamp of A is smaller than that of B.

The Bully Algorithm is used for leader election in distributed systems. When a coordinator fails, the process with the highest ID becomes the new leader.

A distributed system consists of multiple independent computers that communicate over a network to achieve a common goal. Examples include cloud computing and online banking systems.

Atomic commit in distributed databases ensures that all participating systems either complete a transaction successfully or none do. Communication deadlock occurs when processes wait indefinitely for messages from each other.

Byzantine agreement refers to reaching consensus among distributed processors even when some are faulty. It becomes difficult when too many processors fail.

The Ricart-Agrawala algorithm is used to achieve mutual exclusion without a central coordinator by exchanging request and reply messages.

Memory coherence ensures consistency of shared data across multiple processors in distributed systems.

Fault tolerance is achieved by detecting errors, isolating faulty components, and recovering from failures to maintain system reliability.

Questions (Section B)

Explain Lamport’s logical clock.

What is Bully Algorithm?

Define distributed system with examples.

What is atomic commit?

Explain Ricart-Agrawala algorithm.

SECTION – C

Attempt any two (Detailed Explanation)

Design issues of distributed systems include scalability, reliability, security, communication delays, and synchronization challenges. These issues make system design complex.

Concurrency control ensures multiple processes access shared resources without conflict. Time stamping avoids deadlock because transactions are ordered based on timestamps, while locking methods may cause waiting cycles.

ARP (Address Resolution Protocol) converts IP addresses into physical MAC addresses. RARP performs the opposite function. Deadlock-free packet switching ensures smooth data transmission without blocking communication channels.

Questions (Section C)

What are the design issues of distributed systems?

Explain time stamping vs locking methods.

What is ARP and RARP?