Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Developer practice bank has an additional 608 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your team is designing a Cloud Bigtable schema for an IoT telemetry platform that ingests sensor readings (temperature, pressure, and vibration) every second from 500,000 industrial machines. Each machine is assigned a sequential numeric machine_id upon onboarding, and newly onboarded machines generate the highest write traffic. Each machine produces about 5 MB of data per week.
Your schema must meet these requirements:
Which row key strategy and table structure should you implement?
machine_id followed by a millisecond timestamp (e.g., 000499999#1746100800000), and use a tall, narrow structure with one row per reading.This design prefixes each row key with a zero-padded sequential numeric ID followed by a timestamp suffix, writing each single-second telemetry event as its own narrow row.
machine_id values are assigned sequentially upon onboarding and newly onboarded machines generate the highest write volume, unreversed sequential IDs cluster active writes at the upper end of the key space, overloading a small subset of nodes.Your machine IDs were already uniformly active regardless of age (avoiding sequential skew) and each single-second event contained large amounts of unserialized column data that would exceed 100 MB per row if grouped into time buckets.
machine_id using SHA-256 as the sole row key without a time bucket, and use a wide structure that writes all historical readings for a machine as new cells in that single row.This approach uses a cryptographic hash of the machine_id as the entire row key and appends every measurement indefinitely into cells of a single unbounded row per machine.
You only needed point lookups for a small, bounded entity state where garbage collection strictly retains only the latest 1–3 cell versions (preventing unbounded row growth) and you had no need for Key Visualizer readability or ordered range scans.
machine_id (e.g., 2025-05-01T12:00:00#499999), and use a tall, narrow structure that creates a new row for every single-second reading.A timestamp-prefixed tall table places a monotonically increasing timestamp at the start of every row key and creates a separate, narrow row for each individual event or measurement.
There is almost never a valid reason to start a high-write Bigtable row key with a raw timestamp. A tall, single-timestamp layout (with the timestamp as a suffix, never a prefix) would only be preferred if you needed to store an unbounded number of large events per entity that would exceed the 100 MB row size limit under time bucketing.
machine_id followed by a week identifier (e.g., reversed_id#week20), and use a wide time-bucket structure that writes each reading as a new timestamped cell in metric columns.A reversed-ID time-bucket schema combines a reversed numeric identifier as the row key prefix with a time-period suffix (such as a day or week), grouping multiple chronological measurements into a single wide row using timestamped cells. In Bigtable, each intersection of a row key and a column can store multiple cell versions distinguished by their timestamps.
machine_id scatters newly onboarded, high-activity machines across the entire lexicographical key space, ensuring writes are balanced across all tablets and cluster nodes rather than hitting a single node.reversed_id first and the time bucket second stores all weekly rows for a given machine in contiguous row ranges, enabling efficient prefix or range scans via ReadRows.Unlike timestamp-prefixed keys, hashed keys, or tall single-timestamp rows, this design simultaneously eliminates tablet hotspotting, supports contiguous range scans, and maximizes batch read/write throughput and compression without risking oversized rows.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated