Why Spark replaced MapReduce for big data in Apache Spark - Quick Recap

Recall & Review

beginner

What is the main limitation of MapReduce that Spark addresses?

MapReduce writes data to disk after each step, which slows down processing. Spark keeps data in memory, making it much faster.

beginner

How does Spark improve speed compared to MapReduce?

Spark uses in-memory computing, which means it stores data in RAM during processing instead of writing to disk after every step.

intermediate

What kind of tasks is Spark better suited for compared to MapReduce?

Spark is better for iterative tasks like machine learning and interactive data analysis because it can reuse data in memory.

beginner

Why is Spark considered more user-friendly than MapReduce?

Spark provides easy-to-use APIs in multiple languages like Python, Java, and Scala, while MapReduce requires writing complex Java code.

intermediate

What is a key architectural difference between Spark and MapReduce?

MapReduce follows a strict two-stage process (map and reduce) with disk writes in between, while Spark uses a directed acyclic graph (DAG) to optimize execution and reduce disk I/O.

What does Spark use to speed up data processing compared to MapReduce?

A In-memory computing

B More disk writes

C Slower network

D Single-threaded processing

Spark speeds up processing by keeping data in memory, avoiding slow disk writes after each step.

Which programming languages does Spark support with easy APIs?

A HTML and CSS

B Only Java

C C++ and Ruby

D Python, Java, Scala

Spark offers APIs in Python, Java, and Scala, making it accessible to many developers.

Why is Spark better for machine learning tasks than MapReduce?

A It writes data to disk more often

B It can reuse data in memory for multiple iterations

C It uses only one CPU core

D It does not support iterative tasks

Machine learning often requires repeating calculations; Spark keeps data in memory to speed this up.

What is a major drawback of MapReduce compared to Spark?

A It uses a DAG for execution

B It processes data only in memory

C It writes intermediate data to disk after each step

D It supports too many programming languages

MapReduce writes data to disk after each map and reduce step, which slows down processing.

What execution model does Spark use to optimize tasks?

A Directed Acyclic Graph (DAG)

B Linear pipeline

C Single map-reduce step

D Random execution

Spark uses a DAG to plan and optimize task execution, reducing unnecessary disk I/O.

Explain why Spark replaced MapReduce for big data processing.

Think about speed, usability, and how data is handled during processing.

Describe the architectural differences between Spark and MapReduce.

Focus on how tasks are planned and how data moves during processing.