Free Practice Quiz Question List

Parallel and Distributed Computing Online Quiz Questions

Use this free practice quiz with 20 questions to review Parallel and Distributed Computing, test your knowledge, and prepare for your next test or exam.

20 questions

True or false

1 point

True or false: Several activities can be concurrent even when they are interleaved on only one processor.

A
True
B
False

Choose one

1 point

A word-counting system assigns a different file partition to each worker, with every worker performing the same counting operation. Which decomposition strategy is this?

A
Task decomposition
B
Data decomposition
C
Pipeline decomposition
D
Consensus

Fill in the blank

1 point

In the word-counting MapReduce example, workers first produce intermediate pairs such as (word, 1). Before the values for each word can be added, the system must .

Choose one

1 point

A client receives no response before its timeout expires. What is the most accurate conclusion?

A
It proves that the remote process has crashed.
B
It indicates that progress is taking too long but does not usually prove remote failure.
C
It proves that the network has permanently lost the message.
D
It guarantees that retrying will be safe.

True or false

1 point

True or false: A logical clock is intended to measure the actual elapsed time between distributed events.

A
True
B
False

Choose all

1 point

Select all statements that accurately describe coordination or synchronization mechanisms.

A
Locks or mutexes can restrict a critical section to one participant at a time.
B
Queues require every producer and consumer to stop at the same point.
C
Barriers require participants to reach a point before any proceed.
D
Leases guarantee that all replicas have identical state at every instant.

Written response

1 point

A computation takes 120 seconds with one worker and 30 seconds with multiple workers. What is its speedup, calculated as sequential time divided by parallel time? Enter the exact numeric value.

Fill in the blank

1 point

A replicated system guarantees that two leaders never commit conflicting values. This guarantee is an example of .

Choose one

1 point

A team keeps the total data set fixed and measures how execution time changes as it adds workers. Which scaling experiment is being performed?

A
Strong scaling
B
Weak scaling
C
Data decomposition
D
Tail-latency scaling

Choose all

1 point

Select all examples of partial failure in a distributed system.

A
One machine crashes while other machines continue operating.
B
All components stop simultaneously as one indivisible unit.
C
A network partition prevents groups of processes from communicating.
D
A faster processor completes every task without any component becoming unavailable.

Open ended

1 point

Design a high-level architecture for a distributed image-processing service that must continue making progress when individual workers fail or become slow. Explain how the design should divide work, regulate traffic, communicate results, detect and recover from failures, preserve metadata and outputs, prevent conflicting task ownership, and measure performance.

Written response

1 point

What is the name of the causal ordering relation used to describe whether one distributed event could have influenced another?

Choose one

1 point

Which set of conditions matches the assumptions of the FLP impossibility result described in the material?

A
Synchronous communication with known bounded delays and no failures
B
A shared-memory system in which every process observes one global clock
C
A completely asynchronous message-passing system with reliable delivery, unbounded delays, and the possibility of one crash failure
D
A Byzantine system in which every message is guaranteed to arrive immediately

Choose one

1 point

A single-core server switches among several network requests so that all of them make progress during the same period. Which concept does this scenario primarily illustrate?

A
It demonstrates parallelism because one processor is executing several activities simultaneously.
B
It demonstrates concurrency because one processor interleaves work on several activities.
C
It demonstrates neither concurrency nor parallelism because the activities share a processor.
D
It demonstrates distributed computing because several requests are being handled.

Choose one

1 point

In a distributed word-count computation using the MapReduce structure, what should happen after workers produce intermediate pairs such as (word, 1)?

A
Each worker directly writes the final global count without communication.
B
Workers sort documents by file size and discard duplicate documents.
C
Intermediate pairs are grouped by word and their associated values are added.
D
A coordinator sends the same complete document collection to every worker.

Choose one

1 point

A distributed client times out while waiting for a reply. What conclusion is justified solely by that timeout?

A
It cannot by itself prove that the remote process has failed.
B
It proves that the remote process has crashed.
C
It proves that the network permanently lost the request.
D
It proves that the remote process completed the request successfully.

Choose one

1 point

What is the primary purpose of a logical clock in a distributed system?

A
They measure the exact elapsed time between events on different machines.
B
They synchronize all machines to one universally visible wall clock.
C
They determine the physical distance between communicating processes.
D
They capture causal ordering without measuring elapsed real time.

True or false

1 point

True or false: A system can preserve safety by refusing to act during uncertainty, even though that choice may reduce availability.

A
True
B
False

Written response

1 point

A computation takes 120 seconds with one worker and 30 seconds with multiple workers. Using the speedup definition S(p) = T(1) / T(p), what is the speedup?

Written response

1 point

What failure type describes a situation in which groups of processes cannot communicate with one another while some processes may continue running?