Google Professional Data Engineer
Get started today
Ultimate access to all questions.
A popular live TV show engages its audience by allowing viewers to cast votes using their mobile phones within a 3-minute window during the program. This activity generates a significant volume of data in a short period. As the person responsible for the 'Voting infrastructure,' your job is to ensure that the platform can manage the high load and that all the cast votes are accurately processed. During the voting period, it is essential to display partial results in real-time. Once the voting period closes, your task is to count all votes precisely once, while also optimizing for cost efficiency. What approach should you take to achieve these objectives?
A popular live TV show engages its audience by allowing viewers to cast votes using their mobile phones within a 3-minute window during the program. This activity generates a significant volume of data in a short period. As the person responsible for the 'Voting infrastructure,' your job is to ensure that the platform can manage the high load and that all the cast votes are accurately processed. During the voting period, it is essential to display partial results in real-time. Once the voting period closes, your task is to count all votes precisely once, while also optimizing for cost efficiency. What approach should you take to achieve these objectives?
Explanation:
Option D is the correct answer. This is because using Pub/Sub to ingest votes provides scalable and reliable transport. Loading into both Bigtable and BigQuery via a Dataflow pipeline allows for low-latency reads from Bigtable for real-time results and cost-effective storage in BigQuery for long-term analysis. Additionally, shutting down the Bigtable instance after voting concludes helps in optimizing costs. Therefore, option D combines the best of both real-time performance and cost-optimized storage.