Google Professional Cloud Architect on Google Cloud Platform Exam Dumps and Practice Test Questions Set 10 Q136-150
Visit here for our full Google Professional Cloud Architect exam dumps and practice test questions.
Question 136:
A global e-commerce company wants a database to store user orders, shopping carts, and profiles with strong consistency and global availability. Which service should they use?
A) Cloud SQL
B) Cloud Spanner
C) Firestore
D) Bigtable
Answer: B)
Explanation:
Global e-commerce platforms operate under extremely complex and dynamic environments. Every interaction, from browsing a product to placing an order, depends on real-time, consistent data across multiple regions. Users expect immediate feedback, accurate inventory, and up-to-date shopping cart contents regardless of where they log in. Achieving this requires a database that combines global distribution with relational structure, strong consistency, and seamless scalability — all features offered by Cloud Spanner.
Cloud Spanner is unique in providing horizontally scalable, strongly consistent relational storage. Unlike traditional relational databases that often require manual sharding and complex replication strategies for multi-region deployments, Spanner abstracts these complexities. It automatically partitions tables, replicates data across multiple zones, and ensures ACID transactions globally. This ensures that if a user updates a shopping cart in Paris, the change is immediately visible to the same user or others querying inventory from Tokyo or New York, maintaining transactional integrity.
A critical aspect for e-commerce is inventory accuracy and prevention of overselling, especially during high-traffic events like Black Friday or flash sales. Cloud Spanner’s global consistency prevents scenarios where simultaneous transactions in different regions result in conflicting stock counts. For example, if two customers on different continents try to buy the last item simultaneously, Spanner guarantees that only one transaction succeeds while the other receives an appropriate notification, avoiding overselling or inconsistent cart states.
Cloud SQL, while relational and fully managed, is regionally confined. To achieve multi-region availability, teams must implement replication manually or use external tools, which increases operational overhead, latency, and risk of conflicts. Firestore, although scalable and real-time, is document-oriented. Its hierarchical structure suits semi-structured data like user profiles or activity logs, but cannot efficiently handle complex joins, multi-table transactions, or global consistency across orders and inventory. Bigtable is excellent for high-throughput analytical workloads or time-series data, but lacks the transactional guarantees essential for relational e-commerce operations.
Moreover, Cloud Spanner integrates seamlessly with other Google Cloud services to enhance operational and analytical capabilities. For instance, it can feed real-time analytics dashboards, support recommendation engines, or provide operational metrics for fraud detection and supply chain monitoring. During peak traffic periods, Spanner’s horizontal scalability ensures consistent performance without manual provisioning or downtime. Automated failover mechanisms further guarantee business continuity in case of regional outages, a critical factor for a global platform where downtime translates directly into lost revenue.
From a security and compliance perspective, Cloud Spanner supports encryption at rest and in transit, integrates with IAM for access control, and can be configured to meet regulatory requirements across jurisdictions. This is crucial for protecting sensitive user data such as payment details, personal addresses, and purchase history.
In essence, Cloud Spanner offers a rare combination: relational SQL support, strong global consistency, automatic scaling, and high availability, all managed under one service. For global e-commerce applications requiring accurate, real-time, transactional operations across multiple continents, it provides a robust, operationally efficient, and cost-effective solution. It ensures that shopping carts, orders, and profiles remain consistent, reliable, and performant regardless of global traffic patterns, user concurrency, or regional failures, making it the optimal choice.
Question 137:
A mobile gaming company wants to track player actions in real time and update leaderboards globally 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:
Real-time gaming applications require the ability to ingest, process, and store player actions with extremely low latency. Player events such as scoring points, completing levels, or achieving milestones happen concurrently for millions of users worldwide. Pub/Sub acts as a highly scalable messaging system capable of handling massive event streams. Each user action generates a message published to a topic, ensuring no data loss even during peak traffic.
Dataflow then processes these events in real time. It can perform aggregations, filtering, joins with player metadata, or calculations for scoring and achievements. For example, Dataflow can compute global leaderboard positions by continuously updating cumulative scores while considering player rankings across different regions. Its streaming capabilities, including windowing and stateful processing, allow the system to manage late-arriving data and ensure accuracy in leaderboards. Additionally, Dataflow integrates easily with machine learning models to detect cheating, predict player behaviour, or personalise gameplay in real time.
Firestore acts as the storage layer, providing globally consistent, low-latency reads and writes. Its document-oriented model suits hierarchical player data, enabling fast retrieval of individual player stats, session information, and leaderboard positions. Offline support ensures that players experience uninterrupted gameplay even with intermittent connectivity, and synchronisation happens automatically once the connection is restored. Firestore’s scalability allows it to handle sudden spikes in player activity without manual intervention, which is critical during major game events or tournaments.
Other architectures are less suitable. Cloud SQL cannot efficiently process millions of concurrent writes, and Cloud Functions are stateless, limiting long-running event processing. Dataproc is designed for batch workloads, introducing latency that is incompatible with real-time leaderboards. Memorystore is a fast in-memory cache, but it does not provide durable storage for persistent player states.
Using Pub/Sub → Dataflow → Firestore ensures that all player events are processed reliably and reflected instantly in leaderboards across regions. The system supports massive scale, strong consistency, low latency, and real-time analytics, providing a seamless and engaging gaming experience for players worldwide.
Question 138:
A healthcare provider wants to store patient records with relational integrity, encryption, automated backups, and HIPAA compliance. Which service should they use?
A) Cloud SQL
B) Firestore
C) Bigtable
D) Cloud Spanner
Answer: A)
Explanation:
Healthcare data requires strict transactional integrity and strong security measures because patient records contain sensitive personal and medical information. Cloud SQL provides a fully managed relational database with ACID transactions, ensuring that data is consistent, reliable, and protected against corruption. For instance, if a patient’s medication information is updated in one transaction, Cloud SQL guarantees that related updates, such as lab results or treatment schedules, remain synchronised and complete.
Cloud SQL also provides automated backups, point-in-time recovery, and failover capabilities. Automated backups protect against accidental deletions or data corruption, while point-in-time recovery allows the provider to restore the database to a specific moment, which is critical in healthcare environments for maintaining audit trails and meeting regulatory requirements. Its built-in high availability and replication features reduce downtime risks, ensuring that healthcare providers can access records without interruptions, even during maintenance or regional outages.
Encryption is another key feature. Cloud SQL encrypts data at rest and in transit, safeguarding sensitive information from unauthorised access. Combined with Identity and Access Management (IAM), organisations can control which users or applications can access patient data, supporting HIPAA compliance. Logging and auditing capabilities provide detailed records of database activity, which are often required for regulatory audits.
Firestore, while flexible and scalable, is a document-oriented database and lacks the full relational capabilities necessary for maintaining complex patient relationships and performing multi-table transactions reliably. Bigtable is optimised for large-scale analytical or time-series workloads, not for transactional healthcare data that requires strong consistency. Cloud Spanner, though relational and globally consistent, introduces unnecessary complexity and cost for regional healthcare scenarios where global distribution may not be needed.
Using Cloud SQL allows the healthcare provider to focus on developing clinical applications, analytics, and patient-facing tools without worrying about the underlying database infrastructure. By offloading routine administrative tasks—such as provisioning, scaling, patching, and backups—Cloud SQL reduces the operational burden on IT teams, allowing them to dedicate more time to improving patient care and supporting medical research. Its fully managed nature ensures that updates are applied automatically, security patches are installed promptly, and underlying hardware and software are maintained to meet the highest standards of reliability and performance.
Cloud SQL also provides strong integration with monitoring and alerting services, such as Cloud Monitoring and Cloud Logging, enabling IT teams to proactively track performance, detect anomalies, and resolve potential issues before they impact healthcare operations. Metrics such as CPU usage, query performance, connection counts, and replication lag can be continuously monitored, providing actionable insights and ensuring that the system remains responsive under high demand. Automated failover and replication features further guarantee continuous availability, which is critical in clinical environments where downtime can directly affect patient care, treatment scheduling, and access to medical records.
Security and compliance are core advantages of Cloud SQL. Data is encrypted both at rest and in transit, and role-based access controls enforce strict permissions for medical staff, administrators, and applications. Cloud SQL integrates with Identity and Access Management (IAM) policies, enabling granular control over who can access or modify sensitive patient information. When combined with Business Associate Agreements (BAAs) and other regulatory safeguards, Cloud SQL helps healthcare providers maintain HIPAA compliance while delivering high-quality services.
Furthermore, Cloud SQL’s support for relational databases makes it straightforward to enforce relational integrity across patient records, prescriptions, lab results, and appointment schedules. This relational consistency is crucial in healthcare, as it prevents data corruption, ensures accurate reporting, and enables advanced analytics for patient outcomes, predictive modelling, and population health management. By leveraging a reliable, secure, and fully managed relational database, healthcare providers can confidently build and deploy applications that improve patient experiences, enhance operational efficiency, and drive better clinical outcomes.
Question 139:
A biotech company wants to run genomics pipelines using containerised workloads on 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 are computationally intensive workflows that involve multiple stages such as sequence alignment, variant calling, data normalisation, and annotation. These workflows often process terabytes of genomic data and require parallel execution across many compute nodes to finish in a reasonable time. Cloud Batch is designed to orchestrate containerised workloads for large-scale batch processing. It supports preemptible VMs, allowing companies to run expensive compute tasks at a fraction of the cost without compromising the ability to complete the workload.
Cloud Batch manages job scheduling, retries, dependencies, and automatic scaling. For example, if one stage of a genomics pipeline depends on the completion of another, Cloud Batch ensures that tasks are executed in the correct order and retries any failed tasks automatically. This orchestration reduces operational overhead compared to manually managing individual VM instances or complex scripts. Integration with Cloud Storage enables efficient input and output data handling, so raw sequencing data and processed results can be accessed seamlessly.
Alternative options are less suitable for this type of workload. Cloud Run is ideal for stateless, short-lived microservices, but cannot handle long-running high-performance computing (HPC) tasks or manage distributed container workloads efficiently. Cloud Functions are event-driven and ephemeral, designed for lightweight functions rather than multi-hour batch processes. App Engine is a platform for web applications and lacks the flexibility and parallel execution capabilities needed for containerised genomics pipelines.
By leveraging Cloud Batch with preemptible VMs, biotech companies achieve significant cost savings. Preemptible VMs can be reclaimed by Google Cloud at any time, but Cloud Batch handles retries and checkpointing automatically, ensuring workflow completion without manual intervention. Cloud Batch also provides logging, monitoring, and integration with container registries, enabling reproducibility, transparency, and operational control.
This architecture allows teams to focus on pipeline logic and scientific analysis rather than infrastructure management. It scales to meet large dataset requirements, executes tasks in parallel, reduces costs, and ensures reliable, automated orchestration for genomics workloads, making Cloud Batch the optimal choice for high-throughput bioinformatics pipelines.
Question 140:
A media streaming company wants to process millions of user interactions 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:
Media streaming platforms generate vast amounts of real-time user interaction data, including play events, pauses, skips, likes, searches, and browsing behaviour. To deliver personalised experiences—such as recommended content, tailored playlists, and targeted advertisements—the system must process this data continuously with low latency and high reliability. Pub/Sub serves as the ingestion layer, offering a highly scalable, durable, and low-latency messaging system capable of handling millions of events per second. Each user action is published as a message, guaranteeing delivery and enabling real-time analytics pipelines.
Dataflow processes the streaming events in real time, performing operations like filtering, aggregation, sessionization, and windowed computations. For instance, Dataflow can calculate engagement metrics such as average watch time per user or dynamically update recommendation models based on current viewing patterns. Its support for stateful processing, event-time windowing, and late data handling ensures accuracy even when data arrives out of order. Additionally, Dataflow can integrate with machine learning models for personalisation, anomaly detection, or content ranking, enabling intelligent recommendations that adapt instantly to user behaviour.
BigQuery serves as the analytics and storage layer, offering petabyte-scale SQL-based analysis without requiring infrastructure management. Aggregated or historical data can be queried to generate dashboards, conduct cohort analysis, retrain ML models, or produce reports on content popularity. The combination of real-time ingestion and processing with Pub/Sub and Dataflow, followed by storage and analytics in BigQuery, enables a seamless pipeline that delivers actionable insights immediately while maintaining historical data for strategic decisions.
Other architectures are less suitable. Cloud SQL cannot handle the high velocity and volume of millions of concurrent writes, leading to potential bottlenecks. Cloud Functions, being stateless and ephemeral, are insufficient for continuous, stateful processing at this scale. Dataproc is primarily designed for batch processing and introduces unacceptable latency for real-time personalisation. Memorystore is an in-memory store that provides fast but ephemeral storage and cannot serve as a reliable long-term analytics platform.
By combining Pub/Sub, Dataflow, and BigQuery, media streaming companies gain several benefits: reliable ingestion, near real-time processing, automated scaling, integration with ML models, and durable analytics storage. This architecture allows platforms to respond immediately to user interactions, update personalised content continuously, and support operational dashboards and analytics without managing infrastructure manually. It also ensures that the system remains resilient, consistent, and performant under high load, such as during viral content events or peak streaming hours, providing a smooth, engaging user experience globally.
Question 141:
A logistics company wants to store vehicle telemetry for millions of vehicles and query by time ranges efficiently. Which database should they use?
A) Bigtable
B) Cloud SQL
C) Firestore
D) Cloud Spanner
Answer: A)
Explanation:
Vehicle telemetry generates massive volumes of time-series data, including GPS locations, speed, engine metrics, fuel levels, and sensor readings. These datasets grow continuously as millions of vehicles report data multiple times per second. Bigtable is a wide-column NoSQL database specifically optimised for high-throughput writes, low-latency reads, and efficient queries over sequential time-series data. Its row-key design allows data to be stored in chronological order, enabling fast range queries to retrieve data from specific time intervals.
Cloud SQL, while relational, cannot efficiently handle billions of rows and high-frequency writes without complex sharding or scaling, which increases operational complexity and latency. Firestore is document-oriented and suitable for structured application data, but is not optimised for high-volume time-series telemetry. Cloud Spanner provides strong consistency and global distribution but may be overkill for vehicle telemetry, as these workloads do not always require cross-region ACID transactions.
Bigtable’s horizontal scalability allows it to handle sudden spikes in vehicle reporting, ensuring reliable ingestion without performance degradation. It integrates seamlessly with Dataflow for streaming transformations, aggregations, and anomaly detection. Combined with BigQuery, organisations can perform analytics, generate real-time dashboards, and identify trends like vehicle utilisation patterns or maintenance needs.
Moreover, Bigtable supports automated replication and failover, ensuring high availability and durability of telemetry data. Its design allows storage of both current and historical data efficiently, which is critical for fleet management, predictive maintenance, and operational optimisation. This architecture ensures that logistics companies can track vehicles in real time, analyse historical routes, and optimise delivery schedules with precision, providing operational efficiency and cost savings.
Question 142:
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:
Retail analytics often involves querying massive datasets that include sales transactions, customer behaviour, inventory levels, and promotions. BigQuery is a fully managed, serverless data warehouse designed to handle petabyte-scale datasets with SQL-based queries. It automatically scales compute resources, allowing companies to run complex analytical queries without provisioning or managing clusters.
Cloud SQL cannot scale efficiently for petabyte-scale data, leading to bottlenecks and high management overhead. Dataproc is cluster-based and optimised for batch processing with Hadoop or Spark, but it requires manual cluster management, making it less convenient for ad hoc analytics. Firestore is a NoSQL document database, unsuitable for large-scale SQL queries and analytics workloads.
BigQuery enables near real-time analysis of transactional and inventory data. Retailers can detect trends, forecast demand, and optimise stock levels using aggregate functions, joins, and advanced SQL analytics. Integration with ML workflows allows predictive modelling for customer segmentation, recommendation engines, and dynamic pricing. BigQuery’s separation of storage and compute also ensures cost efficiency and performance optimisation, enabling retailers to analyse massive datasets without the operational burden of maintaining infrastructure.
Question 143:
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 involves capturing market movements for millions of financial instruments with sub-second precision. Low-latency ingestion and storage are crucial because delays can result in financial loss or missed trading opportunities. Bigtable is ideal for this use case due to its ability to handle millions of writes per second with minimal latency. Its wide-column design allows efficient retrieval of time-series data for historical analysis, risk assessment, and algorithmic trading.
Cloud SQL cannot manage this high-frequency write load efficiently, and Firestore is not optimised for ultra-low-latency, high-throughput workloads. Cloud Spanner provides strong consistency but introduces slightly higher latency and unnecessary complexity for localised trading datasets.
Bigtable’s horizontal scalability ensures that trading platforms can ingest tick data from multiple exchanges simultaneously. Its integration with Dataflow and BigQuery allows real-time analytics, trend detection, and visualisation, enabling traders and automated systems to react instantly to market changes. Its performance, reliability, and ability to handle massive time-series data make Bigtable the optimal choice for financial trading environments.
Question 144:
A gaming company wants to store player achievements, session data, and leaderboards with strong consistency and low latency. Which database should they use?
A) Firestore
B) Cloud SQL
C) Bigtable
D) Cloud Spanner
Answer: A)
Explanation:
Real-time gaming applications require databases that can handle large numbers of concurrent users while maintaining strong consistency and extremely low latency. Players expect that their progress, achievements, and leaderboard rankings update immediately across devices and regions. Firestore, a fully managed NoSQL document database, provides millisecond-level read and write latency with strong document-level consistency. Its hierarchical data model is ideal for representing user profiles, session states, achievements, and leaderboard data, allowing efficient retrieval and updates.
Cloud SQL, while relational and fully managed, cannot easily scale to millions of concurrent users without sharding or complex replication strategies, which introduces operational overhead and latency. Bigtable excels at time-series or telemetry data but lacks transactional consistency for individual player data, making it less suitable for applications where leaderboards and achievements must be accurate in real time. Cloud Spanner is globally consistent and relational, but introduces complexity and cost that may not be necessary for gaming workloads that are primarily document-centric.
Firestore offers offline support, enabling players to continue interacting with the game even when temporarily disconnected. Changes are automatically synchronised once the connection is restored, ensuring that session data, scores, and achievements remain consistent across all devices. Firestore also scales horizontally without manual intervention, allowing game developers to focus on gameplay features instead of database management.
For example, in a multiplayer battle royale game with millions of players worldwide, Firestore can manage individual match results, update cumulative player statistics, and refresh global leaderboards in real time. Its automatic scaling ensures that sudden spikes, such as during live tournaments or special events, do not cause performance bottlenecks. Firestore also integrates with Firebase Authentication and other Google Cloud services, providing security, access control, and monitoring for game data.
By combining strong consistency, low latency, global scalability, hierarchical data modelling, and offline synchronisation, Firestore provides an optimal solution for real-time gaming workloads. It ensures that players see accurate scores and progress across sessions and devices, enhancing engagement, satisfaction, and competitive fairness.
Question 145:
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 use?
A) Cloud SQL
B) Firestore
C) Bigtable
D) Cloud Spanner
Answer: A)
Explanation:
Healthcare systems handle highly sensitive patient data that requires transactional integrity, security, and regulatory compliance. Cloud SQL is a fully managed relational database providing ACID transactions, which ensure that all operations on patient records—such as updates to prescriptions, lab results, or appointments—are consistent and reliable. This is critical for avoiding data corruption, duplicate entries, or inconsistencies that could affect patient care.
Cloud SQL offers automated backups and point-in-time recovery, reducing the risk of data loss from user errors, corruption, or system failures. For example, if a clinician accidentally updates or deletes a record, the database can be restored to a precise moment before the change. Its high availability configuration provides automated failover and replication across zones, ensuring continuous access to records even during outages.
Security and compliance are key in healthcare. Cloud SQL encrypts data at rest and in transit, supports fine-grained access controls via IAM, and integrates with logging and monitoring tools for auditing. These features help meet HIPAA and other regulatory requirements, ensuring that patient data remains private, secure, and auditable.
Firestore, while flexible and scalable, does not support multi-table ACID transactions at the scale required for relational healthcare applications. Bigtable is optimised for high-throughput analytical workloads but is unsuitable for transactional systems that require relational integrity. Cloud Spanner offers global distribution but may be unnecessarily complex and costly for regional healthcare operations where patient records are primarily accessed within a country or region.
Using Cloud SQL, healthcare providers can focus on clinical applications rather than infrastructure management. Automated patching, updates, monitoring, and security reduce operational overhead while maintaining compliance. Integration with network security, VPC Service Controls, and detailed auditing ensures a secure, reliable environment for managing critical patient information. Cloud SQL provides a balance of simplicity, reliability, compliance, and relational integrity, making it the optimal choice for storing sensitive healthcare data.
Question 146:
A biotech lab wants to run genomics pipelines using containerised workloads on 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 complex, multi-step workflows that process large volumes of genomic data. Typical tasks include sequence alignment, variant calling, normalisation, and data annotation, which require extensive computing resources and often run for hours or even days. Cloud Batch orchestrates containerised workloads across multiple compute nodes, making it ideal for these high-throughput, parallelizable pipelines.
One of the key advantages of Cloud Batch is support for preemptible VMs. Preemptible VMs are short-lived but significantly cheaper than standard instances, providing substantial cost savings for long-running, resource-intensive workflows. Cloud Batch automatically manages retries for preempted tasks, handles job dependencies, and ensures that pipelines execute reliably without manual intervention.
Cloud Run is designed for stateless microservices and short-lived requests, making it unsuitable for multi-hour HPC tasks. Cloud Functions are event-driven and ephemeral, incapable of handling the compute-intensive requirements of genomics workflows. App Engine is a PaaS for web applications and lacks the flexibility to orchestrate distributed containerised tasks.
Cloud Batch also integrates seamlessly with Cloud Storage for input and output data management, enabling efficient handling of terabyte-scale datasets. Logging, monitoring, and metrics provide operational visibility, ensuring reproducibility and reliability of scientific workflows. Its ability to parallelise tasks across many nodes accelerates processing, while dependency management ensures that tasks execute in the correct order.
For example, a lab running a whole-genome sequencing analysis can split the genome into segments, process each segment in parallel on preemptible VMs, and combine the results automatically. This setup reduces runtime, lowers costs, and frees scientists from managing infrastructure. Cloud Batch ensures reproducibility, reliability, and scalability, making it the optimal solution for genomics pipelines requiring container orchestration and cost efficiency.
Question 147:
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 in media streaming platforms relies on continuously ingesting and processing massive streams of user interaction data. This includes events such as video plays, pauses, skips, searches, likes, and ratings. The volume of these events is immense, often millions per second, and must be processed with low latency to provide immediate, personalised content recommendations.
Pub/Sub serves as a high-throughput, reliable messaging system that can handle this volume of events globally. Each user action is published as a message to a topic, ensuring durability and ordered delivery for subsequent processing. Pub/Sub decouples event ingestion from processing, allowing Dataflow to scale independently and handle bursts in user activity without dropping messages or creating bottlenecks.
Dataflow acts as the streaming computation layer. It transforms and aggregates data, handles late-arriving events through windowing, and performs stateful computations. For instance, it can calculate user engagement metrics, session lengths, or the most popular content in real time. Dataflow can also integrate machine learning models to recommend content, detect anomalies, or identify trends instantly. Its serverless nature allows it to scale automatically in response to the volume of incoming events, providing elasticity without manual provisioning.
BigQuery is the analytics and storage layer, designed for massive-scale SQL queries. Processed streams and aggregated datasets are stored in BigQuery, enabling ad hoc queries, historical trend analysis, and dashboard visualisations. Marketing teams can analyse content performance, while recommendation engines can train models on aggregated data for continuous improvement. BigQuery’s separation of storage and compute allows cost-efficient analytics without compromising performance.
Alternative architectures are less suitable for real-time personalisation. Cloud SQL cannot handle high-frequency writes efficiently, and Cloud Functions are stateless and ephemeral, limiting their ability to process continuous streams at scale. Dataproc is optimised for batch workloads and introduces latency incompatible with near-real-time requirements. Memorystore is an in-memory cache and is ephemeral, making it unsuitable for durable analytics or long-term personalisation data storage.
The combination of Pub/Sub → Dataflow → BigQuery provides a complete, end-to-end solution for real-time ingestion, processing, and analytics. It ensures scalability, low latency, global availability, and integration with machine learning workflows. This architecture enables immediate personalisation, content recommendations, and analytics for a media company, creating an engaging, responsive, and adaptive user experience worldwide.
Question 148:
A logistics company wants to store vehicle telemetry for real-time dashboards. Which database should they use?
A) Bigtable
B) Cloud SQL
C) Firestore
D) Cloud Spanner
Answer: A)
Explanation:
Vehicle telemetry data includes time-series information such as GPS coordinates, speed, engine diagnostics, fuel levels, and sensor readings. For a logistics company managing millions of vehicles, this translates into billions of data points per day. The company needs to query this data efficiently by time ranges for real-time dashboards, analytics, and operational decision-making. Bigtable, a wide-column NoSQL database, is specifically optimised for high-throughput writes and low-latency reads over large-scale time-series datasets.
Bigtable’s row-key design allows data to be stored chronologically, making range queries for specific time intervals extremely fast. For example, a dashboard showing vehicle location and performance metrics over the past hour can query billions of rows efficiently. Its horizontal scalability allows the system to accommodate spikes in telemetry data, such as during rush hours or large-scale operations, without performance degradation.
Cloud SQL, while relational, cannot handle billions of rows and high-frequency writes efficiently. Firestore is document-oriented, not optimised for sequential time-series data, and can introduce latency when performing large-scale queries. Cloud Spanner is designed for globally consistent relational workloads, but adds unnecessary complexity for regional telemetry needs.
Bigtable integrates seamlessly with Dataflow for streaming processing and BigQuery for analytics. Dataflow can clean, transform, and enrich telemetry streams, while BigQuery provides historical analysis, predictive maintenance models, and operational insights. Automated replication and failover ensure data durability and high availability, critical for fleet monitoring systems where downtime can result in operational inefficiency or safety risks.
The combination of low-latency ingestion, horizontal scalability, and integration with analytics pipelines makes Bigtable ideal for real-time dashboards. Logistics managers can monitor vehicle performance, optimise routing, predict maintenance needs, and improve fuel efficiency in near real time. Its architecture also supports long-term storage of historical data for trend analysis, regulatory reporting, and predictive modelling. Bigtable ensures that a logistics company can scale operations efficiently while maintaining high performance and operational visibility, making it the optimal choice for telemetry workloads.
Question 149:
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 requires millisecond latency and a consistent player state. Firestore provides strong consistency, offline support, hierarchical document schemas, and automatic scaling. Leaderboards and session data remain synchronised across devices.
Cloud SQL cannot scale efficiently for millions of concurrent users. While it is a fully managed relational database with ACID guarantees, relational databases typically rely on vertical scaling and may require complex sharding or multi-region replication to handle extremely high concurrency. In a gaming environment, where millions of players can simultaneously generate session updates, achievements, and leaderboard changes, Cloud SQL would quickly become a bottleneck. The overhead of managing connections, handling locks, and ensuring transaction consistency across many users introduces latency that is unacceptable in real-time applications. Even with connection pooling and read replicas, relational databases struggle with workloads that involve thousands of rapid write operations per second.
Bigtable, on the other hand, is optimised for telemetry or time-series workloads. It excels at ingesting massive streams of sequential data, such as vehicle sensors or financial ticks, but it does not provide transactional consistency at the per-document or per-user level. In gaming, session data often requires atomic updates—such as updating a player’s current score, achievements, and inventory simultaneously. Bigtable’s design focuses on throughput and scale rather than relational or transactional guarantees, so operations that require strong consistency across multiple attributes can be cumbersome and error-prone. This makes it less suited for player session management or real-time leaderboards, where a consistent state across multiple attributes is critical.
Cloud Spanner is a globally distributed, relational database that offers strong consistency and horizontal scalability. While it can handle very large workloads and multiple regions with transactional integrity, it introduces unnecessary complexity and cost for gaming applications that do not require global relational distribution. Spanner’s infrastructure and management requirements may be overkill for most gaming scenarios, where the primary needs are low-latency reads/writes, offline synchronisation, and hierarchical storage of player-specific data rather than complex multi-table relational joins across the globe.
Firestore is specifically designed to address the challenges of real-time, high-concurrency applications. It offers document-level strong consistency, meaning that when a player’s session data or leaderboard score is updated, all subsequent reads will reflect the latest state instantly. Firestore’s hierarchical document model allows developers to organise player data efficiently, storing session details, inventory, achievements, and rankings within nested documents and collections. This structure simplifies application logic and enables fast queries for frequently accessed data such as leaderboard rankings or current session stats.
Additionally, Firestore provides automatic scaling to accommodate fluctuating workloads. Gaming applications often experience spikes in traffic during peak hours, events, or new game releases. Firestore scales horizontally without requiring manual sharding or infrastructure provisioning, ensuring that the database maintains consistent performance under high load. Offline support further enhances the player experience, allowing updates to be cached locally on a player’s device and synchronised once connectivity is restored. This ensures uninterrupted gameplay even in areas with unstable internet connections.
Firestore integrates seamlessly with other Firebase and Google Cloud services, including Firebase Authentication, Cloud Functions, and Firebase Analytics. This ecosystem enables secure user authentication, server-side event triggers, and analytics-driven personalisation or anti-cheat mechanisms. By providing strong consistency, low latency, automatic scaling, offline support, and deep integration with real-time services, Firestore addresses the full spectrum of requirements for modern gaming workloads. It ensures players see accurate scores, progress, and leaderboard updates instantly, fostering engagement, fairness, and competitive integrity, all while simplifying backend management for game developers.
Question 150:
A biotech company wants to run genomics pipelines using containers on 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 inherently complex and involve multiple stages of data processing, each with specific computational and storage requirements. Typical workflows include sequence alignment, variant calling, annotation, normalisation, and quality control, often operating on terabyte-scale datasets. These pipelines are both compute- and memory-intensive, and their tasks can run for hours or even days. Orchestrating these pipelines manually or using traditional server-based approaches can be highly inefficient, prone to errors, and cost-prohibitive. Cloud Batch is designed to address these challenges by providing managed orchestration for containerised workloads, allowing biotech teams to focus on scientific analysis rather than infrastructure management.
Cloud Batch automatically handles task scheduling, dependency management, retries, and scaling across a large number of compute nodes. Each stage of a genomics pipeline can be encapsulated in a container, ensuring that dependencies, software versions, and runtime environments remain consistent and reproducible across multiple runs. For example, a sequence alignment step can run on multiple preemptible VMs in parallel, while subsequent variant calling tasks wait for completion, reducing overall processing time. Cloud Batch monitors the execution of every task, automatically retrying failed tasks, which is especially useful when using preemptible VMs that may be terminated unexpectedly. This ensures the pipeline continues seamlessly without manual intervention.
Preemptible VMs provide a significant cost advantage for large-scale genomics workloads. They are short-lived but extremely affordable instances, allowing labs to perform massive parallel computations at a fraction of the standard cost. Cloud Batch intelligently schedules tasks on these preemptible resources while managing retries in case of VM preemption, ensuring high reliability without sacrificing efficiency. This is particularly valuable in genomics, where processing thousands of whole-genome sequences can otherwise require enormous budgets and dedicated HPC clusters.
Alternative services are less suitable for genomics workflows. Cloud Run is optimised for stateless microservices with short execution times, making it unsuitable for long-running, multi-step batch computations. Cloud Functions are also event-driven and ephemeral, unable to handle high-throughput containerised workloads or maintain persistent execution for hours. App Engine is a Platform-as-a-Service designed for web applications and does not provide the control, scheduling, or parallelisation capabilities needed for distributed HPC tasks like genomics analysis.
Cloud Batch integrates seamlessly with Cloud Storage for input and output data, enabling efficient access to large sequencing datasets and intermediate results. Logs, metrics, and monitoring are built in, providing operational transparency and the ability to debug pipelines if needed. Furthermore, its container-based approach ensures that pipelines are portable and reproducible across environments, which is critical for scientific research where reproducibility and traceability of results are mandatory.
By combining automated orchestration, preemptible VM cost optimisation, retry mechanisms, dependency management, and seamless integration with cloud storage, Cloud Batch provides a highly scalable and reliable solution for genomics pipelines. It allows biotech teams to accelerate research timelines, reduce costs, and focus on deriving insights from data rather than managing infrastructure. Its ability to scale horizontally, manage complex dependencies, and handle high-throughput workloads makes Cloud Batch the optimal choice for genomics and other computationally intensive bioinformatics workflows.