nerdexam
Google

PROFESSIONAL-DATA-ENGINEER · Question #378

Your infrastructure team has set up an interconnect link between Google Cloud and the on-premises network. You are designing a high-throughput streaming pipeline to ingest data in streaming from an…

The correct answer is C. Use Dataflow, write a pipeline that reads the data from Kafka, and writes the data to BigQuery. Option C is correct because Dataflow (Apache Beam) has a native KafkaIO connector that can read directly from on-premises Kafka over the existing interconnect link, then write to BigQuery - giving you the fewest hops and thus the lowest latency possible. Why the distractors…

Submitted by jian89· Mar 30, 2026Designing data processing systems

Question

Your infrastructure team has set up an interconnect link between Google Cloud and the on-premises network. You are designing a high-throughput streaming pipeline to ingest data in streaming from an Apache Kafka cluster hosted on- premises. You want to store the data in BigQuery, with as minimal latency as possible. What should you do?

Options

  • ASetup a Kafka Connect bridge between Kafka and Pub/Sub. Use a Google-provided Dataflow template to read the data from Pub/Sub, and write the data to
  • BUse a proxy host in the VPC in Google Cloud connecting to Kafka. Write a Dataflow pipeline, read data from the proxy host, and write the data to BigQuery.
  • CUse Dataflow, write a pipeline that reads the data from Kafka, and writes the data to BigQuery.
  • DSetup a Kafka Connect bridge between Kafka and Pub/Sub. Write a Dataflow pipeline, read the data from Pub/Sub, and write the data to BigQuery.

How the community answered

(31 responses)
  • A
    10% (3)
  • B
    16% (5)
  • C
    71% (22)
  • D
    3% (1)

Explanation

Option C is correct because Dataflow (Apache Beam) has a native KafkaIO connector that can read directly from on-premises Kafka over the existing interconnect link, then write to BigQuery - giving you the fewest hops and thus the lowest latency possible.

Why the distractors fail:

  • A and D both route data through Pub/Sub as an intermediary, which adds an unnecessary buffering layer that increases latency - directly contradicting the "minimal latency" requirement. The difference between A and D is only whether you use a Google-provided template (A) or a custom pipeline (D); neither is optimal because Pub/Sub itself is the problem.
  • B introduces a custom proxy host in the VPC, which is unnecessary complexity - Dataflow can reach Kafka directly over the interconnect without an extra relay server, and the proxy becomes an additional latency and failure point.

Memory tip: When you see "interconnect + Kafka + minimal latency," think direct path wins - the interconnect already gives Dataflow private network access to Kafka, so any solution that adds Pub/Sub or a proxy in the middle is introducing latency for no benefit. Pub/Sub shines when you need decoupling or don't have direct connectivity; here, you already have it.

Topics

#Streaming ingestion#Dataflow#Kafka#BigQuery

Community Discussion

No community discussion yet for this question.

Full PROFESSIONAL-DATA-ENGINEER Practice