Hadoop's HDFS is a distributed file system, which of the following application scenarios is suitable for data storage and management? (Multiple choice)

Question

Options

• ALarge number of small file storage
• BHigh capacity, high throughput
• CLow latency read
• DStreaming data access

• A
  11% (7)
• B
  73% (46)
• C
  16% (10)

Explanation

HDFS is architected for batch-oriented, large-scale data workloads, making B and D the correct fit. It achieves high capacity and high throughput (B) by distributing large files across many DataNodes and reading them in parallel, and it is explicitly designed for streaming data access (D) - a sequential, read-once pattern that maximizes throughput over responsiveness.

A is wrong because HDFS stores metadata for every file in the NameNode's RAM. Millions of small files rapidly exhaust that memory, causing severe performance degradation - HDFS strongly favors fewer, very large files (GB to TB range).

C is wrong because HDFS deliberately trades low latency for high throughput; it buffers and pipelines data in large blocks (default 128 MB), so the time-to-first-byte is high. Systems like HBase or Redis are used when low-latency reads are required.

Memory tip: Think of HDFS as a freight train - it moves enormous loads efficiently (B, D) but is slow to start and terrible at delivering individual small packages (not A) quickly (not C).

No community discussion yet for this question.