Exam AWS Certified Data Analytics - Specialty topic 1 question 1 discussion

complex, analytic queries running with minimal latency = REDSHIFT KDF load data into redshift Answer = C

the keywords "complex, analytic queries running with minimal latency"

C. Use Amazon Kinesis Data Firehose to stream data to Amazon Redshift. Use Amazon Redshift as a data source for Amazon to create a business intelligence dashboard. C is valid

the keywords "complex, analytic queries running with minimal latency"

"but also occasionally allows data to be modified using SQL". This is only possible in redshift and redshift does not support kinesis stream directly , so firehouse and then redshift will work as source for quick sight

Answer is now B as you can alter data in the materialized view when ingesting from stream. "You can now connect to and access the data from the stream using SQL and simplify your data pipelines by creating materialized views directly on top of the stream. The materialized views can also include SQL transforms as part of your ELT (extract, load and transform) pipeline."

Can Kinesis Data Streams directly write to S3?

D is a best. KFH provides near-real time data, so it's latency is little higher than that of KDS. Athena is the best data visualizer for complex and raw data. Since the data are in s3, it already appears in athena. Redshift is a data warehouse and can't reciev raw data.

can't use kds to deliver to s3 so must use kdf. use athena or redshift

KDS cant connect datahouse directly. This is reason why We have to use KDF.

The key is "ntegrate complex, analytic queries running with minimal latency."

Definitely C. While you can stream data from KDS into Redshift, you cannot modify the materialised view using SQL, you can only run selects on a materialised view. S3 on the other hand doesn't support data modification using Athena.

C is a good answer, but it is slower compared to data stream (B). Kinesis data firehose has to buffer stream data (thus near-realtime only), and then copy to s3 bucket as a staging area, then issue COPY command to Insert into Redshift table. Whereas Kinesis data stream just transfer data directly into the data warehouse via materialized view(s). Second, the question requires the stream be transform-able via SQL, Data firehose do support transformation of data yes, but only via Lambda blueprints; Data stream transfer into Materialized View, which is written in SQL to convert Json body of the stream data, you can add more SQL here to transform it as you need - so Data Stream satisfies the requirement - albeit with lots of limitations. Link to back my claim up: https://docs.aws.amazon.com/redshift/latest/dg/materialized-view-streaming-ingestion.html

corect answer is D is wrong ?

it's a C

There is a new feature of Redshift, supporting read KDS directly into a materialized view, which support lowest latency and make B quite promising. But it might not meet the modify by SQL requirements. So, this make C be the best choice.

C: I passed the test