Skip to content

Gcp General

← Back to all decks

45 cards — 🟢 15 easy | 🟡 14 medium | 🔴 5 hard

🟢 Easy (15)

1. What is Cloud Datastore?

Show answer Cloud Datastore is a schemaless NoSQL datastore in Google's cloud. Applications can use Datastore to query your data with SQL-like queries that support filtering and sorting. Datastore replicates data across multiple datacenters, which provides a high level of read/write availability.

2. What is private service connect ?

Show answer Private Service Connect enables secure and private connectivity between a customer's Virtual Private Cloud (VPC) network and a service provider's network. It allows organizations to consume managed services while keeping the traffic private.

3. What is spectrum access system ?

Show answer Spectrum Access System refers to the system managing shared access to spectrum frequencies. It enables dynamic spectrum sharing for communication services, optimizing the utilization of available spectrum resources.

Under the hood: SAS manages CBRS (Citizens Broadband Radio Service) spectrum in the 3.5 GHz band. More relevant to telecom than typical DevOps.

Gotcha: This is a niche GCP offering — most engineers will never use it directly.

4. What is TensorFlow Extended (TFX)?

Show answer TFX is an end-to-end platform for deploying production ML pipelines. It's designed to enable the orchestration of ML workflows, ensuring scalability and reproducibility. TFX provides:
* Data Ingestion and Validation: Integrates data from various sources and validates its quality.

5. What is capacity planner in GCP ?

Show answer Capacity Planner in GCP is a tool that assists in estimating the necessary resources for running workloads on Google Cloud. It helps in planning the required capacity of compute, storage, and other resources based on expected usage.

6. What is Google Transfer Appliance?

Show answer It's a physical storage solution for transferring large amounts of data to GCP. This appliance is useful when transferring large datasets that might be impractical to move over the internet due to their size and latency constraints.

7. What is Cloud IoT Edge in GCP?

Show answer Cloud IoT Edge extends Google Cloud's capabilities to the edge for IoT devices, enabling edge computing by providing a framework to run IoT applications and machine learning models directly on IoT devices. It enables local data processing and analysis, reducing latency and optimizing bandwidth usage.

8. What is BigQuery Omni in GCP?

Show answer BigQuery Omni allows users to analyze data across multiple clouds within a single pane of glass. It allows users to analyze and gain insights from data stored in multiple clouds, enabling seamless data processing and analytics.

9. What is Bigtable in GCP?

Show answer Cloud Bigtable is a NoSQL database service for handling large analytical and operational workloads at scale.
* Scalability and Performance: Bigtable is designed for high scalability and performance, capable of handling petabytes of data with low latency.

10. What is profiler in GCP ? How is it useful ?

Show answer Profiler in GCP is a tool for identifying performance bottlenecks in applications. It analyzes code execution and provides insights into performance issues, helping developers optimize and fine-tune their applications.

Under the hood: Cloud Profiler uses statistical sampling with minimal overhead (~0.5% CPU). It captures CPU, heap, and thread profiles.

Analogy: Profiler is to production code what a blood test is to health — continuous low-overhead monitoring that catches problems early.

11. What is Cloud Trace in GCP?

Show answer Cloud Trace is a distributed tracing system for generating latency reports. It provides detailed information about how long it takes for a request to travel through various components of a distributed application, allowing for performance improvements and troubleshooting.

12. What is Dialogflow in GCP?

Show answer Dialogflow is a natural language understanding platform for building conversational applications. It enables developers to design and deploy conversational interfaces, supporting multiple platforms and languages, facilitating natural and rich interactions with users.

13. What is Cloud Storage Transfer Service?

Show answer It's a service for transferring large amounts of data from other cloud providers or on-premises to GCP storage. It allows seamless and secure transfers while handling the complexities of large-scale data migration.

14. What are network tags and how are they different from labels?

Show answer As the name suggests, network tags can be applied only to network resources.
While labels don't affect the resources on which they are applied, network tags do affect resources (e.g. firewall access and networking routes)

15. What does the following command do? gcloud deployment-manager deployments create

Show answer Deployment Manager creates a new deployment.

Under the hood: Deployment Manager uses YAML/Jinja2/Python templates to define GCP resources declaratively — GCP's native IaC tool.

Gotcha: Deployment Manager is being superseded by Terraform and Pulumi in most organizations. Consider those for new projects.

Analogy: Deployment Manager is to GCP what CloudFormation is to AWS.

🟡 Medium (14)

1. What all types of databases are suppported by GCP ?

Show answer AlloyDB, BigTable, Firestore, MemoryStore, Spanner, SQL

Remember: GCP database types by workload: AlloyDB (PostgreSQL-compatible), Bigtable (wide-column NoSQL), Firestore (document), Memorystore (Redis/Memcached cache), Spanner (global relational), Cloud SQL (managed MySQL/PostgreSQL/SQL Server).

Gotcha: Spanner is the only globally-consistent relational database — but it is significantly more expensive than Cloud SQL.

2. Explain TensorFlow on GCP.

Show answer TensorFlow is an open-source machine learning platform, and GCP provides infrastructure and services to leverage TensorFlow efficiently.
Key features:
* High-Performance Computing: GCP offers powerful compute resources to train and deploy TensorFlow models.

3. What are NFS shares ?

Show answer NFS (Network File System) shares allow multiple instances to access and share a common file system over a network. It's a distributed file system protocol that enables a client to access files over a network as if they were on its local disks. In GCP, NFS shares can be set up using Google Cloud Filestore, providing high-performance, fully managed NFS file servers to store and access data for applications that need shared file systems.

4. Describe Google Cloud VPN.

Show answer Cloud VPN allows secure connections between an on-premises network and GCP using IPsec VPN protocols.
* Secure Connectivity: Cloud VPN establishes an encrypted IPsec tunnel between the on-premises network and GCP. This ensures secure communication over the public internet.

5. What is a RISK Manager ?

Show answer In the context of GCP, Risk Manager is a tool that allows organizations to identify and manage various types of risks associated with their GCP environment. It enables continuous risk assessment, monitoring, and response by providing insights into vulnerabilities, compliance issues, and threats. Risk Manager helps in creating a risk-aware culture by centralizing risk management activities and streamlining risk mitigation processes.

6. Explain each of the above databases types.

Show answer 1. AlloyDB
AlloyDB is Google's scalable, distributed, in-memory database. It combines the benefits of traditional relational databases with the scale and performance of NoSQL databases, ideal for high-throughput transactional applications.
2. BigTable

7. What is Secure Boot and how does it protect the boot process?

Show answer Secure Boot is a component of Shielded VMs and is a UEFI feature that ensures the system boots only with signed and verified code, guaranteeing that the operating system and bootloader haven't been tampered with. This prevents the loading of unauthorized firmware and helps protect against boot-level malware or rootkits.

8. Describe Edge AI in GCP.

Show answer Edge AI enables running ML models on edge devices to process data locally. GCP provides tools and services for deploying machine learning models to edge devices, allowing real-time processing and decision-making at the edge.

9. What is Google Cloud Disaster Recovery (DR)?

Show answer It's a set of strategies and services to recover critical systems and data in case of a disaster. GCP offers various features and capabilities to facilitate disaster recovery planning, such as data replication, failover mechanisms, and geographic redundancy. By leveraging GCP's distributed infrastructure and data replication services, organizations can design and implement robust disaster recovery plans to ensure business continuity in case of unexpected disruptions.

10. What is Async Replication ?

Show answer Async Replication, in the context of databases or storage systems, refers to an asynchronous method of data replication. It involves copying and synchronizing data from a source to a destination in a non-blocking manner. The replication process doesn’t require immediate confirmation of data synchronization and can continue independently, potentially leading to a small delay in data consistency between the source and destination.

11. Explain Cloud Monitoring in GCP.

Show answer Cloud Monitoring provides visibility into the performance, uptime, and overall health of cloud resources. It provides insights through dashboards, alerts, and other tools, helping users monitor, troubleshoot, and optimize their cloud-based systems.

12. What is Stackdriver in GCP?

Show answer Stackdriver is a monitoring, logging, and diagnostics service, providing insights into applications on GCP. It includes:
* Monitoring: Real-time performance metrics and uptime monitoring for applications and infrastructure.
* Logging: Centralized log management and analysis across applications and systems.
* Error Reporting: Insights into application errors and exceptions for debugging and improvement.

13. How do you list buckets?

Show answer Two ways to do that:

$ gsutil ls

$ gcloud alpha storage ls

Remember: gsutil is the legacy CLI. gcloud storage is the modern replacement (faster, unified). Both work.

Gotcha: gsutil ls shows all buckets in the current project. Use gsutil ls gs://bucket-name/ to list objects inside a bucket.

14. Explain the Cloud Migration Service on GCP.

Show answer The service helps move on-premises workloads to GCP efficiently and securely. The service offers tools, methodologies, and best practices to streamline the migration process.

Remember: GCP migration tools include Migrate for Compute Engine (VM migration), Migrate for GKE (container migration), and Database Migration Service.

Gotcha: Migration is not just lift and shift — plan for networking, IAM, and service account differences.

🔴 Hard (5)

1. Explain Google Cloud Trace.

Show answer Cloud Trace is a performance monitoring tool for understanding and optimizing latency in applications. It provides:
* Performance Insights: Traces the latency of requests across different services to identify performance bottlenecks.
* Request Analysis: Helps understand the performance of individual requests and their paths through distributed systems.
* Debugging and Optimization: Helps in debugging and optimizing the performance of applications.

2. Explain AI Platform in GCP.

Show answer AI Platform provides a comprehensive set of ML services for building, deploying, and managing models. It supports various tasks, such as data preprocessing, model training, and model deployment at scale. AI Platform provides a collaborative environment for data scientists and machine learning engineers to develop and operationalize ML models.

3. Explain Google Data Studio.

Show answer Data Studio is a free business intelligence and data visualization tool that turns data into informative reports and dashboards.
It allows users to create customizable, informative reports and dashboards using various data sources. Key features include:
* Data Connectivity: Connects to a wide range of data sources.
* Interactive Dashboards: Enables the creation of interactive and visually appealing reports.
* Collaboration: Supports sharing and collaboration on reports within teams.

4. Describe Edge TPU in GCP.

Show answer Edge TPU is Google's purpose-built ASIC designed to run machine learning (ML) models for edge devices. It's optimized for running TensorFlow Lite models for efficient machine learning tasks on edge devices. Edge TPUs enable low-latency, high-throughput, and power-efficient machine learning inference on devices, such as IoT devices or local servers, without the need for continuous cloud connectivity.

5. Explain Google Cloud AutoML.

Show answer AutoML is a suite of machine learning products for developers with limited ML expertise to train high-quality models. It provides tools for building high-quality custom machine learning models with minimal coding. AutoML includes:
* AutoML Vision: Enables the creation of custom image recognition models.
* AutoML Natural Language: Allows the training of custom text analysis models.
* AutoML Tables: Supports building predictive models for structured data without requiring deep ML expertise.