Bitzu

SAA-C03 ยท Data ingestion, lakes and analytics

Kinesis Data Streams vs Amazon Data Firehose

Kinesis Data Streams (shards, retention, custom consumers) versus Amazon Data Firehose (managed delivery to S3, Redshift and others) and choosing by latency and effort.

Key points

  • Kinesis Data Streams: shards hold records 24 hours by default, up to 365 days.
  • Own consumers, several read the same data, replay possible; provisioned shards or on-demand.
  • Amazon Data Firehose: managed delivery, no shards, no consumer code.
  • Buffers and loads into S3, Redshift, OpenSearch, Splunk, HTTP; Lambda transforms, JSON to Parquet/ORC.
  • Data Streams: custom real-time processing, replay; Firehose: least-effort loading.

Source: https://docs.aws.amazon.com/streams/latest/dev/key-concepts.html

Practice questions

Single choice

How long does a new Kinesis data stream keep records by default?

  1. 1 hour
  2. 24 hours
  3. 7 days
  4. 365 days
Show answer

Answer: B. 24 hours

A new stream keeps records for 24 hours. You can raise retention up to 365 days (8,760 hours) for an extra charge, but that is not the default.

Select all that apply

Which statements about Amazon Data Firehose are correct?

  1. You do not size shards or write consumer applications
  2. It keeps records for up to 365 days so consumers can replay them
  3. It can load data into Amazon Redshift through an S3 staging step
  4. It can convert incoming JSON records to Parquet or ORC
  5. Each downstream team attaches its own KCL consumer to it
Show answer

Answer: A. You do not size shards or write consumer applications, C. It can load data into Amazon Redshift through an S3 staging step, D. It can convert incoming JSON records to Parquet or ORC

Firehose has no shards to size and needs no consumer code; it can deliver to Redshift (it stages data in S3 and issues a COPY); and it can convert JSON to Parquet or ORC. Replaying stored records for days and multiple custom consumers are features of Kinesis Data Streams, not Firehose.

True or false

Amazon Data Firehose lets several independent applications read the same records and replay them from an earlier point, just like Kinesis Data Streams.

Show answer

Answer: False

False. Firehose delivers buffered records to a destination; it is not a stream that applications read from or rewind. Multiple independent consumers and replay within the retention period are what Kinesis Data Streams offers.

Practice 8 questions on this topic in Bitzu

Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.

Download on the App StoreGet it on Google Play

More SAA-C03 topics in this module