Exam AWS Certified Data Analytics - Specialty topic 1 question 24 discussion

Answer = D

It's C. Your confusing the "immediately" and ignoring that prior to that it's an event that happens as a sequence within 10 minutes, so firehose is a viable option. AWS auto scaling does not support Amazon Kinesis so that rules out D https://docs.aws.amazon.com/autoscaling/application/userguide/integrated-services-list.html

Bing Option D is correct because it provides a highly available, scalable, and real-time solution. Amazon Kinesis Data Streams can capture and process large streams of data records in real time. AWS Application Auto Scaling can automatically adjust capacity to maintain steady, predictable performance at the lowest possible cost. Amazon Kinesis Data Analytics can process and analyze streaming data using standard SQL, and AWS Lambda can run your code in response to events and automatically manage the compute resources for you. Other options are not the best solutions for this scenario. For example, Options A and B involve using technologies that may not provide the real-time processing required for this use case. Option C does not provide a mechanism for detecting the known event sequence in real time.

I will go with D as KDS is designed to handle massive streams of real-time data and can scale automatically to match the volume of data input and KDA processes of streaming data using SQL and analyse the incoming data stream for patterns, such as the sequence of voltage drops within 10 minutes of a voltage increase. KDA can scale based on demand, supporting the expansion to other cities. AWS Lambda to poll a notification stream and then publish to the Amazon SNS topic for immediate action upon detecting the specified event. Option A involves managing an Apache Kafka cluster and an EMR cluster, which adds complexity and operational overhead. B uses a REST-based web service and RDS, might not scale as seamlessly and could introduce latency in detecting and notifying about the events. C serverless, does not offer the same level of real-time processing and pattern detection capabilities needed for this specific use case as KDA.

kinesis data stream don't support aws auto-scaling, but like everything that happens in the cloud, there exists a way to scale your kds based on lambda and aws auto-scaling... then the D question is not completely wrong... emr is not highly available, then the A question is killed in this fact... this problem announces real-time data streaming, but we discard kdf... not because kdf, but because lambda function... in near-real-time data processing using kdf and lambda with a consumer, just the records streamed in 1MB or 60s are processed by lambda... b makes no sense... this is a poor question and they should be removed from this dump...

The answer is D. It has a requirement of " All notifications must be delivered as quickly as possible.". That requirement rules out Firehose as it can have a delay for buffering and Kinesis data stream is real time.

C. Option D mention "Set up AWS Application Auto Scaling on both streams" and this is not possible. Here is not mention that you enable on demand capacity... no, it mentions "AWS Application Auto Scaling". And what about this part? "Create another stream for notifications." It is not needed! Option C is the correct one

Vote C. A: No, as EMR is not HA. Single region only. B: No, RDS is not ideal for streaming. C: Yes. All systems are fully managed. D: No. Although there is a solution in 2018 (https://aws.amazon.com/blogs/big-data/scaling-amazon-kinesis-data-streams-with-aws-application-auto-scaling/) but it can’t scale more 10 times per 24-hrs nor scale double current shard.

C can't be right, it asks for "real-time", Firehouse collects before it pushes the data out which makes it "near real-time". Answer should be D

C, because of non-HA EMR

EMR is an incorrect answer because it is not high-available

EMR clusters dy default is not highly available.

Very good question. Both A and C are correct. But A can be ruled out because it uses EMR and out of box its not a highly available service, which is a critical requirement here.

C: I passed the test

Answer = C According to https://aws.amazon.com/kinesis/data-firehose/faqs/ It is a fully managed service that automatically scales to match the throughput of your data and requires no ongoing administration.

C is correct

Its A By Elimination. B is not ideal for streams. C. Is not immediate. We need real time thing. D. Way too complex. Not sure if we would need additional Stream, Application auto scalling will scale Lambda. Where did Java come from? I believe there is a hard limit for KDS too. MSK wins on the scalability front. Spark streaming Application can call SNS SDK. Unbelievable, but it is A.