Google Professional Cloud Architect on Google Cloud Platform Exam Dumps and Practice Test Questions Set 9 Q121-135
Visit here for our full Google Professional Cloud Architect exam dumps and practice test questions.
Question 121:
A global e-commerce company needs a database to store user profiles, shopping carts, and order histories with strong consistency, global distribution, and SQL support. Which service should they choose?
A) Cloud SQL
B) Cloud Spanner
C) Firestore
D) Bigtable
Answer: B)
Explanation:
A global e-commerce company handling user profiles, shopping carts, and order histories requires a database that supports ACID transactions, relational queries, and global distribution. These workloads include millions of users performing concurrent operations from different regions. Consistency is critical; a user must see the correct cart contents regardless of location, and inventory must reflect purchases accurately. Cloud Spanner provides horizontal scalability with SQL support, globally consistent transactions, and automated multi-region replication. It allows the company to serve customers worldwide without worrying about regional inconsistencies or latency issues.
While traditional architectures often require choosing between strong consistency, high availability, or global scale, Cloud Spanner is engineered to overcome these limitations through its unique distributed SQL design. By combining horizontal scaling with relational capabilities, it closes the gap between operational database requirements and global traffic distribution. This makes it particularly advantageous for e-commerce ecosystems that demand real-time accuracy across geographically dispersed users, such as synchronised inventory updates, consistent product catalogue views, and unified pricing logic. Unlike Cloud SQL—which must scale vertically and often requires extensive sharding strategies for global workloads—Spanner distributes data seamlessly across nodes while maintaining external consistency. This ensures that customers in different regions never encounter issues like double-selling inventory or inconsistent order statuses.
Additionally, Cloud Spanner’s strongly consistent reads and writes help eliminate race conditions and data conflicts that commonly arise in multi-region NoSQL deployments. Systems relying on Firestore, for example, must often rely on eventual consistency or custom logic to handle conflicting updates under heavy load. Firestore’s real-time syncing is powerful for mobile and lightweight applications, but its architecture is not optimised for complex relational joins or multi-record ACID transactions at scale. E-commerce systems frequently require complex transactional guarantees, such as linking orders, payments, inventory reservations, and shipment records. Spanner’s support for ANSI SQL, schema management, and multi-row transactions ensures that these operations remain secure, consistent, and audit-friendly across global nodes.
Similarly, while Bigtable excels at large-scale analytical and time-series workloads—such as user activity logs, recommendation system inputs, or real-time metrics—it is not designed to enforce relational constraints or manage interconnected datasets. As a result, businesses using Bigtable for primary transactional backing stores must rely on external systems to maintain integrity, increasing architectural overhead. Bigtable’s strength lies in throughput, not structured consistency or relational semantics. This further emphasises Spanner’s unique ability to blend scalability with strict data correctness.
Another key advantage of Cloud Spanner is its automated management of replication, failover, and storage optimisation. Multi-region configurations support synchronous replication across geographically distributed nodes, enabling high availability levels that significantly reduce downtime risk. For e-commerce platforms where even minutes of downtime can cause substantial financial loss and customer dissatisfaction, Spanner’s built-in resilience is a crucial differentiator. Features such as automatic leader election, continuous backups, and point-in-time recovery further strengthen the database’s operational reliability.
Integration with Google Cloud’s broader ecosystem enhances Spanner’s flexibility. It can easily connect with Dataflow, BigQuery, Pub/Sub, and Vertex AI to enable advanced analytics, machine learning–driven personalisation, and predictive inventory management. This seamless interoperability allows organisations to process transactional data in real time, feed it into analytics pipelines, and generate actionable insights without impacting performance. In environments where user behaviour insight drives competitive advantage, this unified operational-analytical approach is essential.
Ultimately, Cloud Spanner’s combination of global consistency, massive horizontal scalability, and relational SQL capabilities positions it as the optimal solution for enterprise-grade e-commerce platforms. Its ability to handle large, unpredictable workloads while maintaining ACID guarantees ensures that businesses can operate confidently at a global scale without sacrificing performance or data integrity. By delivering high availability, seamless integration, and robust consistency, Cloud Spanner outperforms alternatives and provides the architectural foundation necessary for modern, high-traffic e-commerce systems.
Question 122:
A mobile gaming company wants to track player actions in real time and update leaderboards with millisecond latency. Which architecture should they use?
A) Pub/Sub → Dataflow → Firestore
B) Cloud SQL → Cloud Functions → Cloud Storage
C) Dataproc → Cloud Storage → BigQuery
D) Memorystore → Compute Engine → Cloud SQL
Answer: A)
Explanation:
This architecture also ensures that the gaming platform can gracefully scale during traffic spikes, which are common in multiplayer environments—such as during tournaments, seasonal events, or major feature launches. Pub/Sub’s ability to automatically scale to millions of messages per second prevents event ingestion delays, ensuring that every action—from a player movement to a score update—is captured instantly. Without this kind of elastic scaling, bursts in player activity could overwhelm traditional messaging systems or relational databases, leading to dropped events or outdated leaderboards. Pub/Sub’s at-least-once delivery guarantees also ensure reliability, preventing data loss even during unexpected failures or spikes in throughput.
Dataflow further enhances this resilience by providing a fully managed, autoscaling stream processing environment. Instead of pre-provisioning compute resources or manually managing clusters, the system grows and shrinks based on event volume. This is especially important for games with variable traffic patterns, where off-peak hours may require minimal processing, while peak hours demand massive parallelisation. Dataflow’s unified batch and stream processing engine, built on Apache Beam, supports advanced event-time processing features such as watermarking and late data handling. These capabilities ensure accurate real-time analytics even when network delays cause events to arrive out of order.
Another critical advantage of using Dataflow is its ability to enrich incoming event streams before writing results to Firestore or other downstream systems. Player actions can be joined with reference data—such as user profiles, historical performance, or matchmaking ratings—to create contextual insights. This enables game developers to implement advanced mechanics like dynamic difficulty adjustments, fraud detection, or real-time matchmaking optimisations. Dataflow’s integration with external services such as BigQuery, Cloud Storage, or Vertex AI allows for seamless incorporation of ML models into the processing pipeline. For example, a machine learning model hosted on Vertex AI could analyse player behaviour to predict churn risk or trigger personalised in-game incentives.
Firestore plays a crucial role by serving as the low-latency operational data store for player-facing features. Its millisecond-level reads and hierarchical document structure make it ideal for storing nested user attributes such as equipment loadouts, recent achievements, or social connections. Since Firestore supports real-time listeners, game clients can receive live updates whenever data changes—ensuring that leaderboards, matchmaking results, and game states update instantly for all connected players. This real-time synchronisation is a core requirement for multiplayer games that aim to deliver an immersive, responsive user experience.
Additionally, Firestore’s automatic multi-region replication and strong consistency make it reliable for global player bases. Players located in different continents should always see up-to-date leaderboards and consistent profile data. Traditional databases—especially those relying on manual sharding or asynchronous replication—struggle to deliver this level of consistency at scale. Firestore removes this operational burden and provides a serverless environment with built-in reliability and security.
The alternative services fail to meet these requirements for specific reasons. Cloud SQL lacks horizontal write scalability and performs poorly under heavy concurrent insert loads typical of player event streams. Cloud Functions are excellent for lightweight event triggers but not for continuous, stateful streaming pipelines. Dataproc is optimised for batch analytics with tools like Spark or Hadoop; its inherent latency makes it unsuitable for real-time gaming mechanics. Memorystore is ideal for caching, but cannot replace a durable, globally distributed data store with flexible query capabilities.
Ultimately, the Pub/Sub → Dataflow → Firestore architecture delivers the high throughput, low latency, fault tolerance, and global consistency needed for modern multiplayer gaming. This combination provides the foundation for real-time leaderboards, instant event handling, and dynamic gameplay experiences—allowing developers to innovate rapidly while maintaining operational simplicity.
Question 123:
A healthcare company needs a relational database to store sensitive patient data with automated backups, point-in-time recovery, and HIPAA compliance. Which service should they use?
A) Cloud SQL
B) Firestore
C) Bigtable
D) Cloud Spanner
Answer: A)
Explanation:
Healthcare organisations must meet strict regulatory standards for storing patient data, including HIPAA. Cloud SQL provides a fully managed relational database with automated backups, point-in-time recovery, encryption at rest and in transit, and IAM integration. It supports structured relational schemas for storing patient records, lab results, and billing information.
Firestore is a NoSQL document store, suitable for hierarchical application data, but it does not offer the traditional relational transactions required for healthcare records. Bigtable is optimised for high-throughput analytical workloads and is unsuitable for transactional patient data. Cloud Spanner is globally distributed and supports relational workloads, but its operational complexity and cost may be overkill for regional healthcare data storage.
Cloud SQL also provides predictable performance for healthcare applications that rely heavily on relational queries, such as electronic health record (EHR) systems, patient appointment scheduling, laboratory information systems, and billing workflows. Healthcare environments frequently require complex joins, referential integrity, and strict transactional consistency to ensure that patient information is accurate and up to date. Cloud SQL’s mature SQL capabilities and compatibility with widely supported engines such as PostgreSQL and MySQL make it easier for healthcare organisations to migrate existing applications without rewriting large portions of their data layer.
Another key advantage is Cloud SQL’s ability to integrate with identity and access management controls, enabling organisations to apply least-privilege principles across databases. Administrators can configure IAM permissions, enforce strong authentication, and apply fine-grained role management to comply with regulatory standards like HIPAA and HITRUST. Cloud SQL’s automatic storage encryption—combined with customer-managed encryption keys (CMEK) for added security—ensures that both data at rest and data in transit remain protected against unauthorised access. These capabilities are essential for maintaining trust and ensuring that sensitive patient information cannot be compromised.
From an operational standpoint, Cloud SQL reduces the burden of database maintenance, allowing healthcare IT teams to focus on clinical innovation instead of infrastructure management. Automated patching ensures the timely application of security updates without service disruption. The built-in monitoring and performance insights help administrators optimise queries, detect anomalies, and maintain stable application behaviour even during peak usage periods, such as appointment surges or telemedicine events.
Cloud SQL’s seamless integration with other Google Cloud services also enhances the overall ecosystem for healthcare providers. For example, data can be securely exported to BigQuery for population health analytics, predictive modelling, or clinical research, while maintaining strict separation and access controls. This creates a reliable, compliant, and high-performance data environment tailored for modern healthcare applications.
Overall, Cloud SQL delivers the right combination of relational consistency, regulatory alignment, operational automation, and security protections—making it the best fit for healthcare systems that prioritise data integrity, compliance, and ease of management.
Question 124:
A biotech company wants to run genomics pipelines on large datasets using preemptible VMs to reduce costs. Which service should they use?
A) Cloud Run
B) Cloud Batch
C) Cloud Functions
D) App Engine
Answer: B)
Explanation:
Genomics pipelines involve long-running, containerised compute tasks with large datasets. Cloud Batch orchestrates these jobs across multiple preemptible VMs, managing retries, dependencies, and scaling automatically. It integrates with Cloud Storage for input/output datasets and supports containerised workloads, making it ideal for genomics pipelines like sequence alignment, variant calling, and data normalisation.
Cloud Run is for lightweight stateless containerised services and cannot handle long-running HPC workloads. Cloud Functions is event-driven and short-lived, unsuitable for multi-hour genomics processing. App Engine is for web applications and does not support distributed containerised pipelines.
Cloud Batch also simplifies the execution of genomics pipelines by eliminating the overhead associated with cluster provisioning, scaling, and lifecycle management. Traditional HPC clusters require manual allocation of compute nodes, queue management, and resource planning—tasks that can slow down research workflows and introduce operational bottlenecks. Cloud Batch removes these complexities by dynamically allocating resources based on job requirements, ensuring that large batches of genomic analyses execute efficiently without human intervention. This is especially important for workflows such as whole genome sequencing (WGS), variant calling, RNA-seq alignment, and methylation analysis, which often involve thousands of parallelizable tasks.
Because Cloud Batch supports granular job definitions and dependency graphs, researchers can model complex, multi-step genomic pipelines with precision. Steps such as read alignment, sorting, marking duplicates, base recalibration, and variant calling can be expressed as separate tasks with clearly defined input-output relationships. Cloud Batch ensures that each stage executes only when all required upstream tasks have successfully completed. This orchestration guarantees data integrity throughout the pipeline and reduces the likelihood of downstream errors caused by incomplete or corrupted intermediate files.
The seamless integration with Cloud Storage further enhances Cloud Batch’s suitability for large-scale genomics work. Researchers can store raw FASTQ files, reference genomes, and intermediate BAM/VCF files in highly durable, cost-effective buckets. Cloud Batch jobs can read directly from these buckets and write back results without moving data unnecessarily, minimising I/O overhead and reducing pipeline complexity. For extremely large cohorts or longitudinal studies, this tight coupling between compute and storage becomes essential for maintaining performance and cost efficiency.
Cloud Batch’s compatibility with workflow managers such as Nextflow, Cromwell, and Snakemake also enhances flexibility and portability. Many genomics pipelines are already written in these frameworks, and Cloud Batch allows teams to execute them at scale without modifying workflow definitions. This accelerates reproducibility, supports collaborative research, and enables institutions to standardise computational pipelines across teams and projects. Containerization ensures uniform software environments, preventing version conflicts and guaranteeing consistent results—even when processing hundreds or thousands of samples concurrently.
Security and compliance features are equally important for genomics research, particularly when handling clinical samples or human genetic data. Cloud Batch supports VPC Service Controls, private networking, encrypted storage, and IAM-based access control, allowing researchers to maintain strict compliance with standards such as HIPAA, GDPR, and institutional data governance requirements. This ensures that computational scalability does not come at the cost of data privacy or regulatory risk.
Ultimately, Cloud Batch empowers genomics teams to run high-throughput pipelines with minimal operational burden, exceptional scalability, and significant cost savings. By combining automated job management, preemptible computing, container support, and tight integration with the broader Google Cloud ecosystem, it provides an ideal platform for modern genomics research—accelerating discovery, improving reproducibility, and enabling analyses at population sc.
Question 125:
A streaming media company wants to process millions of events per second for real-time personalisation. Which architecture is best?
A) Pub/Sub → Dataflow → BigQuery
B) Cloud SQL → Cloud Functions → Cloud Storage
C) Dataproc → Cloud Storage → Cloud SQL
D) Memorystore → Compute Engine → BigQuery
Answer: A)
Explanation:
Streaming media platforms generate high-velocity events, including play, pause, search, and interaction events. Pub/Sub ingests these events reliably at scale, ensuring durability and ordering when needed. Dataflow performs real-time transformations, aggregations, filtering, and enrichment. It also allows integration with ML models for personalised recommendations. BigQuery stores aggregated data for analytical queries, batch processing, and historical trend analysis.
The Pub/Sub → Dataflow → BigQuery pipeline is particularly powerful for media platforms because it supports both operational and analytical use cases in real time. Modern media consumption—such as video streaming, music playback, podcast listening, and personalised content feeds—generates massive volumes of granular event data. These include play events, pauses, skips, search queries, recommendations interactions, ad impressions, click-through actions, and device telemetry. Handling this volume requires an ingestion system that can scale elastically without bottlenecks, something Cloud SQL, Cloud Functions, and other traditional components are not built to do.
Pub/Sub excels at serving as the high-throughput ingestion layer, capable of processing millions of events per second with extremely low latency. It provides decoupled, asynchronous communication between producers and consumers, ensuring that spikes in user activity—such as major live events, new content drops, or peak evening hours—do not overload downstream systems. Its durability guarantees that no data is lost, and publishers do not need to manage backpressure manually.
Dataflow then transforms these streams in real time, performing tasks such as sessionization, enrichment with metadata, handling late-arriving data, and applying event-time processing. This makes it ideal for real-time content personalisation, where the system must react immediately to user behaviour. For instance, when a user skips a certain genre, Dataflow can update their preference profile within milliseconds and feed that information into downstream personalisation services. It can also join streaming data with external datasets—such as user histories, device information, or ML-driven feature embeddings—to produce richer insights. Additionally, Dataflow’s autoscaling ensures consistent performance without the need for manual resource management.
BigQuery serves as the analytical backbone of the architecture. It allows teams to run complex queries over fresh streaming data, enabling dashboards, real-time KPIs, and large-scale analytical workloads without impacting ingestion performance. With BigQuery’s streaming inserts, insights such as trending videos, real-time ad performance, churn estimates, and engagement patterns can be computed on demand. Integrated machine learning capabilities (BigQuery ML) allow organisations to build and train predictive models directly on hot data, accelerating time-to-insight for recommendation engines, fraud detection systems, and advertising optimisation.
This architecture is also highly cost-effective. Pub/Sub charges based on data volume, Dataflow charges by actual processing usage with autoscaling, and BigQuery operates on a pay-as-you-go model for queries and storage. Together, they eliminate the need for managing clusters, provisioning fixed-capacity systems, or worrying about hardware limitations.
Overall, the Pub/Sub → Dataflow → BigQuery pipeline provides a unified, scalable, and real-time streaming analytics architecture that empowers media platforms to deliver personalised, engaging, and data-driven user experiences at a global scale.
Question 126:
A logistics company wants to store vehicle telemetry data for millions of vehicles and query it efficiently by time ranges. Which database should they use?
A) Bigtable
B) Cloud SQL
C) Firestore
D) Cloud Spanner
Answer: A)
Explanation:
Vehicle telemetry systems generate massive amounts of time-series data at extremely high velocity. Each vehicle produces hundreds of metrics per second—GPS coordinates, engine RPM, fuel level, tyre pressure, battery voltage, braking patterns, temperature readings, and diagnostic fault codes. When multiplied across millions of vehicles, the workload easily reaches billions of writes per day. A database for this use case must support horizontally scalable ingestion, very high write throughput, and efficient retrieval of data by timestamp. Bigtable is designed specifically for this category of workloads.
Bigtable supports wide-column, sparse schema design, enabling developers to model each vehicle as a row and store timestamped events as columns, or vice versa, depending on query patterns. This flexibility allows efficient time-range queries, which are essential for telemetry dashboards, anomaly detection, geospatial analysis, fleet optimisation, predictive maintenance, and regulatory reports. Queries such as “fetch engine temperature over the last 30 minutes,” “retrieve GPS path for the past 24 hours,” or “analyse acceleration events from the last trip” run efficiently due to Bigtable’s optimised key ordering and sorted storage format (SSTables).
Another strength is horizontal scalability. Bigtable can scale to millions of writes per second simply by adding more nodes, without downtime or schema redesign. This enables the logistics company to onboard additional vehicles, expand internationally, or increase telemetry resolution with no architectural constraints. Bigtable’s low-latency read/write performance—typically a few milliseconds—ensures real-time dashboards, alerting systems, and streaming analytics remain responsive even during peak load.
Bigtable also integrates seamlessly with Dataflow for streaming transformations such as data enrichment, filtering, feature extraction, and anomaly detection. It works with BigQuery for analytical workloads, enabling long-term trend analysis, predictive modelling, and fleet-wide operational intelligence. Together, these integrations allow businesses to build complete telemetry pipelines that support both real-time monitoring and large-scale analytics.
By contrast, Cloud SQL cannot ingest telemetry at this scale due to vertical scaling limitations and relational overhead. Firestore offers strong consistency but cannot handle sustained high-frequency writes or large time-series datasets efficiently. Cloud Spanner supports global consistency but introduces higher costs and is optimised for relational workloads, not time-series ingestion.
Overall, Bigtable’s architecture is purpose-built for massive time-series workloads, making it the ideal choice for storing, querying, and analysing vehicle telemetry across millions of data-producing devices.
Question 127:
A retail company wants to analyse petabytes of sales and inventory data using SQL without managing infrastructure. Which service should they choose?
A) BigQuery
B) Cloud SQL
C) Dataproc
D) Firestore
Answer: A)
Explanation:
BigQuery is the most suitable choice for a retail company that wants to analyse petabytes of sales, customer behaviour, inventory movement, and operational data without managing any infrastructure. Retail environments typically generate massive datasets from point-of-sale systems, online transactions, supply chain systems, mobile apps, loyalty programs, and pricing engines. These datasets must often be queried together to produce insights such as demand forecasting, inventory replenishment, product performance trends, and regional sales comparisons. BigQuery is built to handle this level of analytical complexity and scale while requiring no servers, hardware planning, database tuning, or cluster management.
BigQuery separates compute and storage, allowing queries to run efficiently even on extremely large tables. This architecture also enables on-demand scaling, so queries that require scanning terabytes of data complete within seconds by automatically allocating the necessary resources. Retail analysts can run complex joins, window functions, aggregations, or machine learning models directly using SQL. This is critical for use cases such as forecasting seasonal demand, identifying slow-moving products, analysing promotions, or studying pricing elasticity across different regions.
Another advantage is its tight integration with upstream and downstream services. Data can be ingested through Pub/Sub, Dataflow, Datastream, or Cloud Storage, enabling real-time or batch-based analytics pipelines. Retail companies often rely on dashboards for operations and decision-making, and BigQuery connects easily with Looker, Data Studio, and BI tools. This ensures that executives and analysts always have access to up-to-date metrics, whether they are monitoring live sales performance or analysing long-term business trends.
BigQuery ML also allows data scientists to build and train models such as customer segmentation, product recommendations, churn prediction, or sales forecasting directly within the warehouse. This eliminates the overhead of exporting data to external ML platforms.
Other options do not meet the scale or operational requirements. Cloud SQL cannot handle petabyte-scale analytics. Dataproc requires cluster administration and is better suited for batch workloads. Firestore is a document database and not designed for analytical SQL. BigQuery offers high performance, flexibility, and zero operational burden, making it ideal for retail analytics.
Question 128:
A financial company wants ultra-low latency storage for tick-level trading data. Which database should they choose?
A) Bigtable
B) Cloud SQL
C) Firestore
D) Cloud Spanner
Answer: A)
Explanation:
Tick-level trading data is characterised by extremely high-frequency updates across millions of financial instruments, with sub-second granularity. Each trade, quote, bid, or ask creates a continuous stream of events that must be captured, stored, and queried in near real time for risk management, algorithmic trading, market surveillance, and reporting. Bigtable is well-suited for this type of time-series workload because it provides predictable low-latency writes, high throughput, and horizontal scalability. Composite row keys can be structured to optimise queries by instrument and timestamp, allowing traders and analysts to efficiently scan historical price movements, aggregate trades, or detect anomalies within milliseconds.
Bigtable’s wide-column architecture allows financial firms to store multiple metrics, such as trade price, volume, bid-ask spread, and market depth, in a single row with column families that can be organised by type or instrument. This enables fast sequential inserts and efficient range scans, which are essential for backtesting trading strategies, generating live market feeds, or calculating intraday indicators. The database can handle millions of concurrent writes per second, ensuring that no market data is lost even during periods of extreme volatility.
In addition, Bigtable integrates with Dataflow for real-time stream processing, enabling preprocessing, enrichment, and aggregation of financial events before they are stored or forwarded to downstream analytics tools. For example, Dataflow pipelines can calculate moving averages, detect outliers, or generate trading signals in real time. Bigtable can also feed BigQuery for historical analysis, regulatory reporting, and machine learning models without introducing significant latency.
Other database options are less suitable for tick-level data. Cloud SQL lacks horizontal scalability and cannot efficiently manage millions of concurrent writes. Firestore is optimised for hierarchical document storage and cannot handle high-throughput time-series ingestion. Cloud Spanner provides consistency and relational capabilities but introduces higher latency and operational complexity, making it impractical for ultra-low-latency trading requirements.
Bigtable’s high write performance, time-range query efficiency, and integration with analytics pipelines make it the ideal database for storing and processing tick-level financial data in real time.
Question 129:
A gaming company wants to store player achievements, session data, and leaderboards with strong consistency and low latency. Which database should they choose?
A) Firestore
B) Cloud SQL
C) Bigtable
D) Cloud Spanner
Answer: A)
Explanation:
Real-time gaming platforms require databases that provide low-latency access, strong consistency, and support for highly concurrent users. Player achievements, session states, and leaderboards must be updated immediately and reflected accurately across all connected devices. Firestore is well-suited for this use case because it provides document-level strong consistency, real-time updates, offline support, and a flexible hierarchical schema that can model nested data such as achievements, inventories, session statistics, and social connections.
Firestore’s architecture allows millions of concurrent players to read and write simultaneously without performance degradation. Real-time listeners ensure that leaderboard changes, in-game rewards, or status updates propagate instantly to all connected clients, maintaining fairness and engagement. Offline support allows players to continue interacting with the game even when connectivity fluctuates, and changes are synchronised automatically when the connection is restored. These features are critical for ensuring a seamless, responsive gaming experience across mobile devices, consoles, and web clients.
Cloud SQL is limited in handling extremely high concurrency because relational databases typically require vertical scaling and can become bottlenecked under heavy write loads. Bigtable excels in time-series or telemetry workloads but does not provide transactional consistency across multiple documents or entities, which is essential for maintaining accurate player progress and leaderboard rankings. Cloud Spanner, while scalable and strongly consistent, is more complex and costly than necessary for gaming session data, making it less practical for typical real-time gaming applications.
Firestore also offers seamless integration with other Google Cloud services such as Cloud Functions, Pub/Sub, and Dataflow. This enables serverless event handling, real-time notifications, and analytics pipelines without manual infrastructure management. It allows developers to implement features such as personalised achievements, dynamic matchmaking, and reward systems with minimal operational overhead.
Overall, Firestore’s combination of strong consistency, low-latency reads and writes, automatic scaling, and hierarchical document support makes it the ideal database for storing player achievements, session data, and leaderboards in real-time gaming environments.
Question 130:
A healthcare provider wants a relational database to store patient records with automated backups, point-in-time recovery, and HIPAA compliance. Which service should they choose?
A) Cloud SQL
B) Firestore
C) Bigtable
D) Cloud Spanner
Answer: A)
Explanation:
Healthcare workloads demand databases that provide strong transactional integrity, structured relational schemas, automated backup and recovery, and strict compliance with regulations such as HIPAA. Patient records, appointment schedules, lab results, prescriptions, and billing information require ACID-compliant operations to ensure data accuracy and consistency. Cloud SQL meets these requirements by providing a fully managed relational database with features such as automated backups, point-in-time recovery, encryption at rest and in transit, and fine-grained access control through IAM integration.
Cloud SQL’s relational structure allows healthcare organisations to maintain normalised tables with relationships, constraints, and indexes, which are essential for complex queries, joins, and reporting. For example, querying a patient’s medical history, medication interactions, and recent lab results requires consistent and accurate relational data. Cloud SQL ensures that these operations execute reliably, without risk of partial updates or inconsistent state, which is critical for patient safety and operational efficiency.
The managed nature of Cloud SQL reduces administrative overhead for healthcare IT teams. Automated patching and updates ensure that the database remains secure and compliant without downtime or manual intervention. Point-in-time recovery enables rapid restoration of data in case of accidental deletion or corruption, minimising the risk of service disruption. Additionally, Cloud SQL integrates seamlessly with other Google Cloud services, allowing secure connections to analytics platforms, reporting tools, or machine learning workflows for population health analysis and predictive modelling.
Other database options are less suitable for healthcare workloads. Firestore is a NoSQL document store that does not provide the relational integrity or complex query capabilities required for structured patient records. Bigtable is designed for analytical and time-series data, not transactional healthcare data. Cloud Spanner, while relational and globally distributed, introduces additional complexity and cost that may be unnecessary for most healthcare applications.
Cloud SQL provides a combination of operational simplicity, relational integrity, compliance features, and reliability, making it the ideal choice for storing, managing, and querying patient records in healthcare environments.
Question 131:
A biotech lab wants to run containerised ML pipelines for genomics datasets using preemptible VMs. Which service should they use?
A) Cloud Run
B) Cloud Batch
C) Cloud Functions
D) App Engine
Answer: B)
Explanation:
Genomics pipelines are highly compute-intensive and involve multiple interdependent steps, including sequence alignment, variant calling, quality control, and annotation. Each step often requires different software tools and significant CPU and memory resources. To reduce cost while maintaining efficiency, biotech labs can leverage preemptible virtual machines, which are significantly cheaper than standard instances but may be terminated unexpectedly. Cloud Batch is specifically designed to handle this type of workload, providing automated orchestration, dependency management, retries, and scaling across thousands of containerised tasks.
Cloud Batch allows researchers to package each step of their genomics pipeline in containers, including tools like BWA, Samtools, DeepVariant, or GATK, ensuring that software dependencies are consistent and reproducible across executions. It automatically schedules jobs based on resource requirements, manages retries if preemptible VMs are terminated, and optimises parallel execution to make full use of available compute resources. This ensures that large-scale genomics analyses, which can involve processing thousands of samples, are executed efficiently without manual intervention.
Integration with Cloud Storage enables seamless access to input and output datasets, allowing raw FASTQ files, reference genomes, and intermediate BAM/VCF files to be stored durably and accessed by multiple pipeline stages. Logging, monitoring, and job tracking are built in, providing transparency and reproducibility, which is crucial for scientific research and regulatory compliance. Cloud Batch also supports flexible scaling, automatically allocating more compute resources as needed, and shutting down idle resources to minimise cost.
Other Google Cloud services are less suitable for these workloads. Cloud Run is designed for stateless web services and cannot efficiently handle long-running pipelines. Cloud Functions is event-driven and short-lived, making it impractical for multi-hour genomic analyses. App Engine is a platform for deploying web applications rather than high-performance scientific computing. Cloud Batch is specifically engineered for high-throughput, containerised, HPC-style workloads, providing both scalability and cost efficiency.
By using Cloud Batch, biotech labs can run complex genomics pipelines reliably, reproducibly, and at a lower cost, while leveraging the flexibility of containerised environments and preemptible computing.
Question 132:
A media company wants to analyse user interactions in real time for personalisation using SQL. Which architecture is best?
A) Pub/Sub → Dataflow → BigQuery
B) Cloud SQL → Cloud Functions → Cloud Storage
C) Dataproc → Cloud Storage → Cloud SQL
D) Memorystore → Compute Engine → BigQuery
Answer: A)
Explanation:
Real-time personalisation for a media platform requires a system that can ingest, process, and analyse massive volumes of user interaction events with minimal latency. Events such as video plays, song skips, ad impressions, likes, searches, and clicks must be captured in real time to drive recommendations, user engagement metrics, and targeted advertising. The architecture consisting of Pub/Sub → Dataflow → BigQuery is well-suited for these requirements because it combines scalable ingestion, real-time processing, and high-performance analytics in a fully managed environment.
Pub/Sub acts as the messaging layer, capable of handling millions of events per second without bottlenecks. It decouples event producers and consumers, allowing event ingestion to scale elastically according to traffic patterns. Its reliability guarantees, including at-least-once delivery and ordering options, ensure that all user interactions are captured accurately for downstream processing. This is critical for media platforms where missing even small amounts of data can affect personalisation algorithms and analytics.
Dataflow serves as the processing engine, applying transformations, aggregations, filtering, enrichment, and windowed computations in real time. It supports both streaming and batch workloads, allowing historical and live data to be processed using the same pipelines. Integration with machine learning models enables immediate scoring of user behaviour for content recommendations, dynamic personalisation, fraud detection, or engagement prediction. Dataflow’s autoscaling and serverless operation remove the need to manage clusters, ensuring consistent performance during peak usage periods.
BigQuery provides a high-performance, serverless analytics layer where aggregated and processed data can be stored and queried using SQL. Analysts and product teams can run ad hoc queries, generate dashboards, and perform predictive analytics without managing infrastructure. Its columnar storage, partitioning, and clustering features optimise both performance and cost for petabyte-scale datasets, while integrations with tools like Looker or Data Studio enable visualisation and reporting for business decisions.
Alternative architectures such as Cloud SQL, Cloud Functions, Dataproc, or Memorystore cannot provide the combination of scalable ingestion, real-time transformations, and low-latency analytics required. The Pub/Sub → Dataflow → BigQuery pipeline offers a complete, resilient, and fully managed solution for real-time streaming analytics and personalised media experiences.
Question 133:
A logistics company wants to store vehicle telemetry data for real-time dashboards. Which database should they use?
A) Bigtable
B) Cloud SQL
C) Firestore
D) Cloud Spanner
Answer: A)
Explanation:
Vehicle telemetry for logistics companies generates high-volume, time-series data from GPS sensors, engine diagnostics, fuel levels, tyre pressure, and other vehicle metrics. Capturing this data in real time and making it available for dashboards, monitoring, and analytics requires a database that can handle rapid, sequential writes, large-scale storage, and efficient time-range queries. Bigtable is well-suited for this use case because it is designed for wide-column, high-throughput workloads that scale horizontally with minimal latency.
Bigtable allows each vehicle to be represented as a row, while timestamps and telemetry metrics can be organised as columns or column families. This structure supports fast sequential writes and efficient scans over time ranges, which are essential for applications such as trip tracking, route optimisation, predictive maintenance, and fleet performance analysis. Its low-latency reads and writes ensure that dashboards can display near real-time vehicle positions, status updates, and alerts without delay, enabling operational teams to react quickly to issues.
Scalability is another key advantage. As fleets grow and additional vehicles are added, Bigtable can scale horizontally by adding nodes, allowing ingestion of millions of events per second without degrading performance. Integration with Dataflow enables preprocessing, enrichment, or aggregation of raw telemetry before storage, while BigQuery can be used for historical trend analysis, operational reporting, and predictive modelling. This combination provides a complete pipeline for both operational and analytical needs.
Other database options do not meet these requirements. Cloud SQL cannot handle billions of time-series writes efficiently due to vertical scaling limitations. Firestore is optimised for hierarchical documents and low-latency access to individual records, but it is not ideal for high-frequency sequential telemetry. Cloud Spanner provides global distribution and relational consistency, but its capabilities are unnecessary for most IoT telemetry workloads and would increase complexity and cost.
Bigtable delivers high-throughput, low-latency, horizontally scalable storage with efficient time-range queries, making it the ideal database for storing vehicle telemetry and supporting real-time dashboards in logistics and fleet management applications.
Question 134:
A gaming company wants low-latency storage for player session data and leaderboards with strong consistency. Which database should they use?
A) Firestore
B) Cloud SQL
C) Bigtable
D) Cloud Spanner
Answer: A)
Explanation:
Real-time gaming platforms require databases that provide low-latency reads and writes, strong consistency, and the ability to handle millions of concurrent users. Player session data, inventories, achievements, and leaderboards must be synchronised in near real time to ensure a fair and engaging experience. Firestore is well-suited for this type of workload because it offers hierarchical document storage, real-time updates, offline support, and automatic scaling to accommodate sudden spikes in traffic.
Firestore’s document-oriented architecture allows developers to model complex player data, including nested fields for achievements, equipment, session statistics, and social interactions, without the constraints of a relational schema. Real-time listeners propagate changes instantly to all connected clients, ensuring that leaderboard rankings, in-game rewards, and session progress are immediately reflected across devices. Offline support allows players to continue interacting with the game even when connectivity fluctuates, with changes synchronised automatically once the connection is restored.
Cloud SQL, while relational, cannot scale efficiently to handle millions of concurrent writes, making it unsuitable for fast-moving gaming workloads. Bigtable excels at time-series and telemetry storage but lacks transactional consistency at the document or multi-entity level, which is critical for player session synchronisation and leaderboard accuracy. Cloud Spanner provides strong consistency and global scalability but introduces unnecessary operational complexity and higher cost for gaming workloads that do not require cross-region replication.
Firestore also integrates seamlessly with Google Cloud services such as Cloud Functions, Pub/Sub, and Dataflow, allowing developers to trigger serverless workflows, process analytics events, or update leaderboards in real time without managing infrastructure. This supports a responsive, interactive user experience, enabling features such as personalised achievements, matchmaking, rewards, and event-driven notifications.
Overall, Firestore’s combination of strong consistency, real-time updates, hierarchical document modelling, offline support, and automatic scalability makes it the best choice for storing player session data and leaderboards in modern, high-concurrency gaming applications.
Question 135:
A biotech company wants to run genomics pipelines using containers and preemptible VMs. Which service should they use?
A) Cloud Run
B) Cloud Batch
C) Cloud Functions
D) App Engine
Answer: B)
Explanation:
Genomics pipelines involve computationally intensive workloads that process large volumes of sequencing data, including FASTQ files, reference genomes, and intermediate BAM or VCF files. Each step in a pipeline, such as alignment, variant calling, quality control, and annotation, may require different software tools, significant CPU and memory resources, and coordination across multiple compute nodes. Cloud Batch is specifically designed to handle these high-throughput, containerised workloads at scale while minimising operational complexity and cost.
Cloud Batch allows researchers to package each step of the genomics pipeline into containers, ensuring consistent and reproducible execution environments across all compute nodes. It handles automatic scheduling, retries, and dependency management, allowing complex pipelines to run efficiently without manual oversight. By supporting preemptible virtual machines, Cloud Batch significantly reduces costs for large-scale genomic analyses, while still guaranteeing job completion through automated retries and rescheduling of preempted tasks.
Integration with Cloud Storage enables direct access to input datasets and secure storage of intermediate and output files. Cloud Batch can execute thousands of parallel tasks simultaneously, scaling automatically based on the computational requirements of the pipeline. Detailed logging, monitoring, and job tracking provide transparency, reproducibility, and auditability, which are crucial for scientific research, publication, and regulatory compliance.
Other Google Cloud services are less suitable for genomics workloads. Cloud Run is optimised for stateless web services and cannot efficiently handle long-running or resource-intensive computations. Cloud Functions is event-driven and short-lived, making it impractical for multi-hour genome analysis tasks. App Engine is designed for deploying web applications rather than orchestrating high-performance scientific pipelines. Cloud Batch is purpose-built for high-throughput, HPC-style workloads, providing both scalability and cost efficiency.
By using Cloud Batch, biotech companies can execute complex genomics pipelines reliably, reproducibly, and efficiently, leveraging containerization, preemptible VMs, and automated orchestration. This allows research teams to focus on scientific insights rather than infrastructure management, enabling faster, cost-effective analyses at scale.