0

I have been struggling with finding accurate information for concurrency capabilities of AWS Kinesis. So far, my understanding of intake concurrency and output concurrency of AWS Kinesis differ from one another (DataStream, Firehose, and Analytics).

Here is my current understanding of the differences:

AWS Kinesis DataStream: Can intake data from only "One source", but can have "Multiple source" as an output. (Meaning, AWS Kinesis DataStream is NOT CAPABLE of "intake concurrency" but is CAPABLE of "output concurrency").

AWS Kinesis Firehose: Can intake data from "Multiple sources", but can only have "One source" as an output. (Meaning, AWS Kinesis Firehose is CAPABLE of "intake concurrency" but is NOT CAPABLE of "output concurrency").

AWS Kinesis Analytics: Can intake and output data from only "One source". (Meaning AWS Kinesis Analytics is NOT CAPABLE of "intake concurrency" and "output concurrency").

Please let me know if my understanding of AWS Kinesis is correct or not. Thank you.

I tried to search for an answer on the AWS docs and FAQ, google search, Bard, and ChatGPT. I was not able to find a definite answer. (A.I. searches were conflicting).

Shin Yoo
  • 1

2 Answers2

0

AWS Kinesis DataStream. Can intake data from only "One source", but can have "Multiple source" as an output. (Meaning, AWS Kinesis DataStream is NOT CAPABLE of "intake concurrency" but is CAPABLE of "output concurrency").

That's incorrect. There can be many different surfaces injecting data to the same steam, as long as the stream has capacity to support large data injections.

AWS Kinesis Firehose: Can intake data from "Multiple sources", but can only have "One source" as an output. (Meaning, AWS Kinesis Firehose is CAPABLE of "intake concurrency" but is NOT CAPABLE of "output concurrency").

That's correct.

AWS Kinesis Analytics: Can intake and output data from only "One source". (Meaning AWS Kinesis Analytics is NOT CAPABLE of "intake concurrency" and "output concurrency").

That's incorrect. You can have multiple input streams for AWS Kinesis Analytics application.

Marcin
  • 215,873
  • 14
  • 235
  • 294
  • Thanks for your quick reply. So, both DataStream and Analytics are able to intake data from multiple sources. Could you clarify if DataStream and Analytics are capable of delivering output to multiple consumers (concurrency) at the same time as well? – Shin Yoo Mar 27 '23 at 00:56
0

I looked further into my own questions to make sure. So far, my new understanding of Kinesis concurrency is as follows:

DataStreams: Can intake data from multiple sources at the same time. Can output data to multiple sources.

Firehose: Can intake data from only one source. Can output data to only one source.

Analytics: Can intake data from multiple sources at the same time. Can output data to multiple sources at the same time.

I appreciate Marcin for answering my question, but I have found some conflicting information based on some practice exams for Solutions Architect Associate. (Practice exams are from Udemy course made by "Stéphane Maarek")

If there are others with more accurate knowledge of AWS Kinesis concurrency abilities, please leave an answer or a comment. Thank you.

Shin Yoo
  • 1