Core Google Cloud Platform Services¶

Cloud Digital Leader Focus: This section covers GCP services from a BUSINESS perspective - understanding WHAT services do, WHY to choose them, and WHEN to use them for business value. Focus on business outcomes, cost optimization, and strategic decision-making.

Compute Services Overview¶

Compute services form the foundation of your cloud infrastructure. The key business question is: How much operational responsibility do you want?

Business Decision Framework: - Full control needed? Choose Compute Engine (IaaS) - Focus on code, not infrastructure? Choose App Engine (PaaS) - Containerized applications? Choose GKE or Cloud Run - Event-driven, short tasks? Choose Cloud Functions - Serverless containers? Choose Cloud Run

Compute Services¶

Compute Engine (IaaS)¶

What it is: Virtual machines running in Google's data centers - maximum control and flexibility

Business Value: - Cost optimization: Custom machine types prevent over-provisioning (only pay for exact resources needed) - Legacy migration: Easiest path for "lift-and-shift" migrations from on-premises - No vendor lock-in: Standard VMs work like any server environment - Business continuity: Live migration means zero downtime during maintenance - Budget control: Sustained use discounts (up to 30%) automatic, no upfront commitment

Key Business Features: - Custom machine types: Tailor CPU and memory to your needs - optimize costs by avoiding oversized instances - Preemptible instances: Up to 80% cost savings for fault-tolerant workloads (batch processing, rendering) - Live migration: Zero-downtime maintenance - Google moves VMs between hosts transparently - Sustained use discounts: Automatic discounts (up to 30%) for long-running instances - no commitment required - Committed use discounts: 57% savings for 1-year, 70% for 3-year commitments - Global load balancing: Distribute traffic across regions for performance and reliability

When to Choose Compute Engine (Business Perspective): - Migrating existing applications with minimal changes (lift-and-shift strategy) - Applications requiring specific OS configurations or licensed software - Need full control over the operating system and environment - Running workloads where you want to optimize every aspect of cost and performance - Legacy applications that cannot be easily containerized or refactored

Business Use Cases: - Web servers and application hosting for traditional architectures - Development and testing environments that need various configurations - High-performance computing (HPC) for research, financial modeling, rendering - SAP, Oracle, and enterprise applications requiring specific configurations - Disaster recovery infrastructure (cost-effective with preemptible instances)

Cost Considerations: - Most flexible pricing with per-second billing (1-minute minimum) - Sustained use discounts automatic - good for always-on workloads - Preemptible VMs for 80% savings on non-critical workloads - Custom machine types prevent waste from oversized instances - Budget visibility: easy to track and forecast costs

Pricing Model: Pay per second with a 1-minute minimum; automatic discounts for sustained use

App Engine (PaaS)¶

What it is: Platform for building and hosting web applications - focus on code, Google manages infrastructure

Business Value: - Faster time-to-market: Developers focus on business logic, not infrastructure management - Automatic cost optimization: Scale to zero when idle, no paying for unused capacity - Reduced operational overhead: No need for DevOps team to manage servers, patching, scaling - Traffic management: Built-in A/B testing and gradual rollouts reduce deployment risk - Developer productivity: Integrated services (caching, task queues, authentication) accelerate development

Key Business Features: - Automatic scaling: Scale to zero (pay nothing when idle) or handle traffic spikes automatically - Multiple language support: Java, Python, PHP, Go, Node.js, .NET, Ruby - use your team's expertise - Integrated services: Built-in services for common needs (reduce third-party service costs) - Version management: Deploy multiple versions and split traffic for A/B testing and safe rollouts - No server management: Eliminate operational overhead - no patching, no capacity planning

When to Choose App Engine (Business Perspective): - Startups and small teams without dedicated operations staff - Applications with unpredictable or variable traffic (scales automatically) - Rapid application development and prototyping (fastest path to production) - Web applications and APIs that fit standard patterns - When reducing operational costs and complexity is the priority

Business Use Cases: - Web applications with variable traffic (e-commerce sites, campaign landing pages) - RESTful APIs for mobile and web applications - Microservices architecture for agile development - Rapid prototyping and MVP development for startups - Internal business applications with intermittent usage

Cost Considerations: - Pay only for actual usage - scales to zero when idle - No infrastructure management costs (staff time savings) - Standard environment: Lower cost, faster scaling, more restrictions - Flexible environment: Higher cost, more flexibility, slower scaling - Good for variable workloads that benefit from auto-scaling

Environments: - Standard: Preconfigured runtime environments, fastest scaling, lowest cost, some restrictions - Flexible: Docker containers with more flexibility, slower scaling, higher cost

Google Kubernetes Engine (GKE)¶

What it is: Managed Kubernetes service for containerized applications - enterprise-grade container orchestration

Business Value: - Application portability: Avoid vendor lock-in with industry-standard Kubernetes - Multi-cloud strategy: Run same applications on GCP, on-premises, or other clouds (with Anthos) - Operational efficiency: Google manages Kubernetes complexity, your team focuses on applications - Resource optimization: Bin-packing and autoscaling maximize infrastructure utilization - Enterprise security: Built-in compliance, security scanning, and identity management

Key Business Features: - Managed Kubernetes: Google manages the control plane (no management overhead, 99.95% SLA) - Auto-scaling: Horizontal pod autoscaling and cluster autoscaling optimize costs and performance - Security: Workload Identity, Binary Authorization, vulnerability scanning reduce security risk - Integrated monitoring: Built-in logging and monitoring provide operational visibility - Multi-cloud: Anthos enables consistent deployment across clouds and on-premises

When to Choose GKE (Business Perspective): - Containerized applications requiring complex orchestration - Multi-cloud or hybrid cloud strategy (avoid vendor lock-in) - Microservices architecture with many independent services - Teams with Kubernetes expertise or investment - Applications requiring high availability and sophisticated deployment patterns

Business Use Cases: - Enterprise microservices platforms requiring scalability and resilience - Hybrid deployments connecting cloud and on-premises data centers - CI/CD pipelines for automated testing and deployment - SaaS platforms requiring tenant isolation and resource management - Applications requiring gradual rollouts, blue-green deployments

Cost Considerations: - Pay for worker nodes (VMs) running your containers - Control plane is free for standard clusters, charged for Autopilot - Autopilot mode: Google manages nodes, you pay only for pod resources (simplified billing) - Standard mode: You manage nodes, more control, more complexity - Spot nodes available for 80% savings on fault-tolerant workloads

Node Types: - Standard nodes: Traditional VMs for containers - full control - Spot nodes: Preemptible instances for 80% cost savings on batch/fault-tolerant workloads

Cloud Functions (Serverless)¶

What it is: Event-driven serverless compute platform - pay only for execution time

Business Value: - Zero infrastructure cost: No servers to manage or pay for when idle - Instant scalability: Handle traffic spikes automatically without capacity planning - Rapid development: Simple deployment model accelerates time-to-market - Event-driven architecture: React to business events in real-time - Cost efficiency: Sub-second billing, pay only for actual execution

Key Business Features: - Event-driven: Triggered by events from GCP services (automate business processes) - Auto-scaling: Scale from zero to handle traffic (no wasted capacity) - Pay-per-use: Only pay for execution time (sub-100ms billing increments) - Multiple triggers: HTTP, Cloud Storage, Pub/Sub, Firestore (integrate with existing systems) - Language support: Node.js, Python, Go, Java, .NET, Ruby, PHP (use existing skills)

When to Choose Cloud Functions (Business Perspective): - Simple, single-purpose functions responding to events - Glue code connecting different services and APIs - Workloads with unpredictable, sporadic usage patterns - Real-time data processing pipelines - Budget-conscious projects requiring minimal infrastructure cost

Business Use Cases: - API backends and webhooks for integrations (Slack bots, payment webhooks) - Data processing and ETL (process files uploaded to Cloud Storage) - IoT data processing (respond to sensor data in real-time) - Real-time file processing (image thumbnails, document conversion) - Scheduled tasks and automation (data exports, cleanup jobs)

Cost Considerations: - Extremely low cost for low-volume workloads (generous free tier) - Pay only for execution time (to nearest 100ms) - No cost when functions are not running - Best for short-lived tasks (max 9 minutes execution) - Invocations are also metered (very low cost per million)

Trigger Types: - HTTP triggers: Direct HTTP requests (webhooks, APIs) - Background triggers: Cloud events from storage, database, messaging services

Cloud Run¶

What it is: Fully managed serverless platform for containerized applications - serverless + containers

Business Value: - Application portability: Use standard containers, avoid vendor lock-in - Developer flexibility: Run any language, any library, any binary - Cost efficiency: Scale to zero, pay only for actual requests and compute time - Fast deployment: Deploy containers in seconds without infrastructure management - No Kubernetes complexity: Get container benefits without Kubernetes overhead

Key Business Features: - Container-based: Deploy any containerized application (Docker, etc.) - ultimate flexibility - Automatic scaling: Scale to zero (no cost when idle) or handle traffic spikes instantly - Request-based billing: Pay only for requests and compute time (to the nearest 100ms) - Portable: Standard container runtime - migrate easily to other platforms - Integrated security: Built-in TLS, DDoS protection, IAM integration

When to Choose Cloud Run (Business Perspective): - Containerized applications needing serverless benefits (cost, scaling) - Applications using languages or libraries not supported by App Engine - Existing containerized applications you want to run serverless - Teams already using containers but wanting to avoid Kubernetes complexity - Web applications and APIs requiring fast, automatic scaling

Business Use Cases: - Containerized web applications and REST APIs - Microservices that need to scale independently - Batch processing jobs that run on-demand - Event-driven applications processing business events - Migration of containerized apps from other platforms (preserving investments)

Cost Considerations: - Pay only for requests and compute time (100ms billing increments) - Scales to zero - no cost when idle - More expensive per request than always-on VMs, but cheaper for variable workloads - No cluster management costs (unlike GKE) - Free tier includes 2 million requests per month

Cloud Run vs. GKE Decision: - Choose Cloud Run: Simpler applications, variable traffic, want zero operational overhead - Choose GKE: Complex orchestration needs, multi-cloud portability, sophisticated deployment patterns

Storage Services Overview¶

Storage is often the largest cloud cost after compute. The key business question is: How frequently will you access this data?

Business Decision Framework: - Unstructured data (files, images, videos)? Choose Cloud Storage - VM disk storage? Choose Persistent Disks - Shared file storage across VMs? Choose Filestore - Access frequency determines cost: Standard > Nearline > Coldline > Archive

Storage Services¶

Cloud Storage (Object Storage)¶

What it is: Scalable object storage for unstructured data - unlimited capacity, global availability

Business Value: - Cost optimization: Pay only for what you use, no capacity planning, automatic lifecycle management - 99.999999999% durability: Data is virtually never lost (11 nines of durability) - Global reach: Serve content globally with low latency through Google's network - No infrastructure management: No storage arrays, no capacity planning, no administration - Seamless scaling: Store exabytes without performance degradation

Storage Classes (Critical for Cost Optimization): - Standard: Frequently accessed data (>1x/month) - highest cost, no retrieval fees, best for hot data - Nearline: Infrequently accessed data (1x/month) - 50% cheaper, small retrieval fee, 30-day minimum - Coldline: Rarely accessed data (1x/quarter) - 70% cheaper, higher retrieval fee, 90-day minimum - Archive: Long-term archival (1x/year) - 90% cheaper, highest retrieval fee, 365-day minimum

Key Business Features: - Global accessibility: Access from anywhere on the internet via API, console, or command line - Automatic redundancy: Data replicated across multiple locations automatically (no extra cost) - Lifecycle management: Automatic data class transitions based on age or access patterns (optimize costs automatically) - Security: Encryption at rest and in transit by default, granular access controls - Integration: Works seamlessly with all GCP services (BigQuery, Dataflow, Vertex AI) - Versioning: Keep multiple versions of objects for data protection and compliance

When to Choose Cloud Storage (Business Perspective): - Storing unstructured data (files, images, videos, backups, logs) - Need unlimited storage without capacity planning - Serving content globally to users or applications - Building data lakes for analytics (BigQuery can query directly) - Long-term data retention for compliance (cheaper than tape)

Business Use Cases: - Website content and media files (images, videos, downloads) - Data backup and archival (replace tape backups, 11 nines durability) - Data lake for analytics (store raw data, query with BigQuery) - Content distribution for global applications - Disaster recovery storage (off-site backup location) - Machine learning training datasets (integrates with Vertex AI) - Log storage and compliance archives

Cost Considerations: - Choose storage class based on access frequency (biggest cost factor) - Standard: $0.020/GB/month, no retrieval costs - Nearline: $0.010/GB/month, $0.01/GB retrieval - Coldline: $0.004/GB/month, $0.02/GB retrieval - Archive: $0.0012/GB/month, $0.05/GB retrieval - Lifecycle policies automate transitions to cheaper classes - Egress charges apply when data leaves GCP

Cost Optimization Strategy: - Use Standard for frequently accessed data (active data) - Automatically transition to Nearline after 30 days (monthly access) - Move to Coldline after 90 days (quarterly access) - Archive for compliance and long-term retention (annual access) - Object Lifecycle Management automates this (set it and forget it)

Persistent Disks¶

What it is: High-performance block storage for virtual machines - like hard drives for your VMs

Business Value: - Performance options: Match disk performance to application needs, control costs - Data protection: Automatic encryption and snapshot backups protect business data - High availability: Regional disks provide automatic replication across zones - Flexibility: Resize disks without downtime as business needs grow - Reliability: 99.999% annual uptime for regional persistent disks

Types (Choose Based on Performance Needs): - Standard persistent disks: HDD-based storage - lowest cost, good for sequential access (file storage, backups) - SSD persistent disks: High-performance SSD storage - balanced price/performance (most databases, applications) - Extreme persistent disks: Highest performance for demanding workloads (SAP HANA, high-transaction databases)

Key Business Features: - Encryption: Encrypted by default at no extra cost (compliance requirement) - Snapshots: Point-in-time backups for disaster recovery and data protection - Resizing: Resize disks without downtime (adapt to changing business needs) - Regional disks: Replicated across zones for high availability (99.999% SLA vs 99.9% for zonal)

When to Choose Persistent Disks (Business Perspective): - VM-based applications requiring block storage - Databases running on Compute Engine - Applications requiring high IOPS and low latency - Business-critical applications needing high availability

Business Use Cases: - Database storage for SQL Server, Oracle, MySQL on VMs - File systems for application servers - Boot disks for virtual machines - Development and test environments

Cost Considerations: - Pay for provisioned capacity (not usage) - Standard: $0.040/GB/month (lowest cost) - SSD: $0.170/GB/month (4x more, much faster) - Extreme: $0.290/GB/month (highest performance) - Snapshots charged separately ($0.026/GB/month) - Regional disks cost 2x zonal but provide replication

Filestore¶

What it is: Fully managed Network File System (NFS) for applications - shared file storage across VMs

Business Value: - Application compatibility: Supports legacy applications requiring shared file systems - No management overhead: Fully managed NFS service, no file server administration - High performance: Enterprise-grade performance for demanding applications - Simplified migration: Easy lift-and-shift for applications needing shared storage - Automatic backups: Built-in data protection for business continuity

Key Business Features: - High performance: Up to 16 GB/s throughput, low latency for demanding workloads - POSIX compliance: Standard file system interface, works with existing applications - Automatic backups: Scheduled and on-demand backups for data protection - Integration: Works seamlessly with GKE and Compute Engine - Scaling: Scale capacity independently of performance

When to Choose Filestore (Business Perspective): - Migrating applications that require shared file storage - Multiple VMs or containers need to access the same files - Applications requiring NFS protocol support - High-performance file operations (media rendering, data analytics)

Business Use Cases: - Application migration requiring shared storage (lift-and-shift) - Content management systems with shared assets - Media processing and rendering workloads - Data analytics workloads with shared datasets - Developer environments requiring shared code/data

Cost Considerations: - Pay for provisioned capacity (minimum 1TB) - Basic tier: $0.20/GB/month (1TB minimum, up to 60 MB/s per TB) - High Scale tier: $0.30/GB/month (10TB minimum, up to 120 MB/s per TB) - Enterprise tier: $0.30/GB/month (1TB minimum, highest performance) - More expensive than Cloud Storage but provides file system semantics

Database Services Overview¶

Choosing the right database is critical for business success. The key question is: What type of data and what scale?

Business Decision Framework: - Relational data, regional scale? Choose Cloud SQL - Relational data, global scale? Choose Cloud Spanner - Document data for mobile/web apps? Choose Firestore - Massive analytical workloads? Choose BigQuery - IoT or time-series data? Choose Bigtable

Database Selection Matrix: | Requirement | Service | Why | |-------------|---------|-----| | Traditional SQL app | Cloud SQL | Easiest migration, lowest cost | | Global SQL application | Cloud Spanner | Global consistency, unlimited scale | | Mobile/web app data | Firestore | Real-time sync, offline support | | Analytics/reporting | BigQuery | Serverless, petabyte-scale | | IoT/time-series | Bigtable | High throughput, low latency |

Database Services¶

Cloud SQL¶

What it is: Fully managed relational database service - MySQL, PostgreSQL, SQL Server without the operational overhead

Business Value: - Zero database administration: No patching, backups, replication management (reduce IT overhead) - Cost predictability: Pay for what you provision, no licensing complexity for MySQL/PostgreSQL - Fast migration: Compatible with on-premises databases, minimal code changes - High availability: Automatic failover, 99.95% SLA reduces business disruption - Focus on business logic: Developers build features, not manage databases

Supported Engines: - MySQL: Open-source relational database - no licensing costs, broad ecosystem - PostgreSQL: Advanced open-source database - enterprise features without cost - SQL Server: Microsoft's relational database - familiar to Windows shops, bring-your-own-license or included

Key Business Features: - Automatic backups: Point-in-time recovery up to 7 days (protects against data loss) - High availability: Multi-zone deployments with automatic failover (99.95% SLA) - Security: Encryption at rest/transit, IAM integration, private IP (compliance ready) - Scaling: Read replicas for read-heavy workloads, vertical scaling for growth - Maintenance: Automatic updates and patches during maintenance windows (no downtime) - Connection pooling: Efficient connection management reduces resource waste

When to Choose Cloud SQL (Business Perspective): - Migrating existing MySQL, PostgreSQL, or SQL Server databases - Traditional web and business applications - Applications requiring ACID transactions and relational data - Regional applications (single region or nearby regions) - Teams with SQL expertise, wanting managed operations

Business Use Cases: - E-commerce platforms and shopping carts (transactional integrity) - Content management systems (WordPress, Drupal) - ERP and CRM applications (business data with relationships) - Customer databases and user management - Inventory management systems

Cost Considerations: - Pay for instance size (CPU/RAM) + storage + network egress - MySQL/PostgreSQL: No licensing fees (major cost savings vs commercial databases) - SQL Server: License included or BYOL (bring your own license) - Shared-core instances for dev/test (start at ~$10/month) - High availability costs ~2x (worth it for production) - Read replicas add cost but improve performance

Cloud Spanner¶

What it is: Globally distributed, strongly consistent relational database - SQL at global scale

Business Value: - Global business reach: Serve customers worldwide with low latency from single database - No trade-offs: Get both global scale AND strong consistency (revolutionary capability) - Unlimited growth: Scale horizontally without limits (add capacity as business grows) - Mission-critical SLA: 99.999% availability (5 nines = 5 minutes downtime per year) - Future-proof architecture: Built for global, always-on applications

Key Business Features: - Global scale: Horizontal scaling across regions - serve billions of transactions - Strong consistency: ACID transactions globally (no eventual consistency workarounds) - SQL interface: Standard SQL queries (use existing SQL skills) - High availability: 99.999% SLA (five nines) - essentially always available - Automatic sharding: Transparent data distribution (Google handles complexity) - No downtime scaling: Add/remove capacity without maintenance windows

When to Choose Cloud Spanner (Business Perspective): - Global applications serving customers worldwide - Financial services requiring strong consistency (banking, payments) - Applications outgrowing regional databases - Mission-critical systems requiring 99.999% availability - When you need both SQL and global scale (unique combination)

Business Use Cases: - Global financial services (payments, trading, banking) - Gaming leaderboards and player data (worldwide players) - Supply chain management across regions - Global retail platforms - SaaS applications with worldwide customers

Cost Considerations: - Premium service with premium pricing (most expensive GCP database) - Pay for node capacity + storage (minimum 1 node = ~$750/month per region) - Regional config: 1 region, lower cost, 99.99% SLA - Multi-region config: 3+ regions, higher cost, 99.999% SLA - Only choose when global scale and consistency are required - Consider Cloud SQL first - migrate to Spanner when you outgrow it

Firestore¶

What it is: NoSQL document database for mobile and web applications - real-time, offline-capable

Business Value: - Faster app development: Real-time sync and offline support built-in (reduce development time) - Better user experience: Apps work offline, sync automatically when online - Global reach: Multi-region replication serves users worldwide with low latency - Scales automatically: From zero to millions of users without capacity planning - Mobile-first: Purpose-built for modern mobile and web applications

Key Business Features: - Real-time synchronization: Live updates to all connected clients (collaborative features built-in) - Offline support: Local caching and automatic sync when online (works anywhere) - Multi-region replication: Data replicated globally for low-latency access - Security rules: Declarative security at document level (protect user data) - ACID transactions: Strong consistency guarantees when needed - Flexible data model: Document structure adapts as business needs change

When to Choose Firestore (Business Perspective): - Mobile applications requiring offline capability - Real-time collaborative applications - Applications with unpredictable, variable traffic - Rapid prototyping and MVP development - When you want serverless, zero-ops database

Business Use Cases: - Mobile applications (iOS, Android apps) - Real-time chat and messaging applications - Collaborative tools (shared documents, project management) - Social media platforms (user profiles, feeds, likes) - Gaming applications (player state, matchmaking)

Cost Considerations: - Pay for document reads, writes, deletes (not storage primarily) - Generous free tier (50K reads, 20K writes, 20K deletes per day) - Good for mobile apps with many inactive users - Can be expensive for write-heavy workloads - Storage: $0.18/GB/month (relatively expensive for large data)

Cloud Bigtable¶

What it is: NoSQL wide-column database for large analytical and operational workloads - petabyte-scale, low latency

Business Value: - Handle massive scale: Billions of rows, millions of operations per second - Consistent performance: Single-digit millisecond latency at any scale - Cost-effective at scale: Cheaper than traditional databases for large data volumes - IoT-ready: Purpose-built for time-series and high-throughput data - Real-time insights: Process massive data streams for immediate business intelligence

Key Business Features: - High throughput: Millions of operations per second (handle any traffic) - Low latency: Single-digit millisecond latency (real-time applications) - Petabyte-scale: Store and process petabytes of data efficiently - Integration: Works with Hadoop, Spark, Dataflow, and HBase APIs - Automatic replication: Multi-cluster replication for global applications

When to Choose Bigtable (Business Perspective): - IoT applications collecting massive sensor data - Time-series data (financial tick data, monitoring metrics) - High-throughput operational workloads - Real-time analytics on large datasets - When you have consistently large workloads (minimum scale requirements)

Business Use Cases: - Time-series data (stock prices, sensor readings, application metrics) - IoT applications (connected devices, telemetry) - Financial services (trading data, fraud detection) - AdTech and MarTech (user behavior tracking, personalization) - Real-time analytics and monitoring

Cost Considerations: - Minimum cluster size: 1 node (~$650/month) - not cost-effective for small workloads - Cost-effective at scale (petabytes of data) - Pay for node capacity + storage (SSD or HDD) - Storage is cheap: $0.17/GB/month for SSD - Only economical for consistently large workloads - Consider Firestore or Cloud SQL for smaller datasets

BigQuery (Data Warehouse)¶

What it is: Serverless, petabyte-scale data warehouse - analytics without infrastructure

Business Value: - Business intelligence without infrastructure: No servers, no database administration - Analyze all your data: Petabyte-scale - analyze your entire business data in one place - Real-time insights: Stream data and query immediately for timely business decisions - Democratize data: Standard SQL enables business analysts to query data themselves - Built-in ML: Business analysts can build ML models without data science expertise

Key Business Features: - Serverless: Zero infrastructure management, automatic scaling (focus on insights, not operations) - SQL interface: Standard SQL queries (use existing skills, broad tool compatibility) - Real-time analytics: Stream data for immediate insights (milliseconds from event to query) - BigQuery ML: Build machine learning models using SQL (democratize ML) - Data sharing: Secure data sharing across organizations (collaborate with partners) - Separation of compute and storage: Query performance independent of data size

When to Choose BigQuery (Business Perspective): - Business intelligence and reporting needs - Data warehousing without infrastructure management - Ad-hoc analysis of large datasets - Real-time dashboards and analytics - When you want analysts querying data directly (self-service analytics)

Business Use Cases: - Business intelligence and executive dashboards - Customer analytics and segmentation - Marketing attribution and campaign analysis - Log analysis and security monitoring - Financial reporting and forecasting - Supply chain analytics

Cost Considerations: - Two pricing models: On-demand or flat-rate (capacity reservations) - On-demand: $5 per TB queried (pay per query) - Storage: $0.02/GB/month for active data (first 10GB free) - Flat-rate: $2,000/month per 100 slots (predictable cost for heavy users) - Free tier: 1TB queries, 10GB storage per month - Cost optimization: Partition tables, cluster columns, use appropriate data types

Networking Services Overview¶

Networking affects both user experience (performance) and security. Key business considerations: global reach, performance, and cost.

Business Decision Framework: - Private network isolation? VPC is foundational - Global application? Cloud Load Balancing for traffic distribution - Improve website performance? Cloud CDN for content caching - Connect to on-premises? Cloud VPN (encrypted) or Cloud Interconnect (dedicated)

Networking Services¶

Virtual Private Cloud (VPC)¶

What it is: Software-defined networking for GCP resources - your private network in the cloud

Business Value: - Security and isolation: Private network protects business resources - Global infrastructure: Single network spans all GCP regions (simplifies architecture) - Cost savings: Private Google Access reduces egress costs - Compliance: Network isolation helps meet regulatory requirements - Hybrid connectivity: Securely extend on-premises network to cloud

Key Business Features: - Global scope: Single VPC spans all regions (unlike AWS regional VPCs) - Firewall rules: Control traffic to/from instances (security) - Private Google Access: Access Google services without external IPs (reduce egress costs) - VPC peering: Connect VPCs privately (multi-project, multi-organization networking) - Shared VPC: Centralized network management across projects (enterprise governance)

When to Choose VPC (Business Perspective): - Every GCP deployment needs VPC (foundational service) - Multi-tier applications requiring network segmentation - Compliance requirements for network isolation - Hybrid cloud connecting on-premises and cloud - Multi-project environments needing centralized networking

Business Use Cases: - Secure multi-tier applications (web, app, database tiers) - Hybrid cloud with on-premises data center connectivity - Multi-environment deployments (dev, test, prod isolation) - Compliance-driven architectures requiring network controls - Multi-project organizations needing centralized networking

Cost Considerations: - VPC itself is free (no charge for the network) - Pay for resources in VPC (VMs, load balancers, etc.) - Egress charges when data leaves Google network - Private Google Access reduces egress costs

Cloud Load Balancing¶

What it is: Distribute incoming traffic across multiple backend instances - global scale, built-in redundancy

Business Value: - Global reach with single IP: Serve users worldwide from one endpoint - High availability: Automatic failover to healthy instances reduces downtime - Performance: Route users to nearest/fastest backend (better user experience) - Scalability: Handle traffic spikes automatically (Black Friday, viral events) - Cost optimization: Distribute load efficiently, auto-scale backends

Types (Choose Based on Application Needs): - Global HTTP(S): Web applications - global reach, SSL termination, URL-based routing - Global SSL Proxy: Non-HTTP SSL traffic - global reach for SSL applications - Global TCP Proxy: Non-HTTP TCP traffic - global reach for custom protocols - Regional Network: Regional TCP/UDP - high throughput, preserve source IP - Regional Internal: Internal load balancing - private traffic within VPC

Key Business Features: - Global reach: Single anycast IP serves globally (simplify DNS, best performance) - Auto-scaling: Scale backends automatically based on traffic (handle demand spikes) - Health checks: Route traffic only to healthy instances (improve reliability) - SSL termination: Handle SSL/TLS encryption (offload work from backends) - Content-based routing: Route traffic based on URL, headers (support microservices)

When to Choose Load Balancing (Business Perspective): - Applications requiring high availability (minimize downtime) - Global applications serving users worldwide - Applications with variable traffic patterns - Multi-region deployments for disaster recovery - Microservices needing intelligent traffic routing

Business Use Cases: - Global web applications and e-commerce sites - API gateways for mobile and web applications - Multi-region deployments for performance and reliability - SaaS platforms serving global customers - Applications requiring zero-downtime deployments (blue-green, canary)

Cost Considerations: - Pay for ingress traffic processed and instance hours - Global load balancers: ~$25/month + $0.008-$0.012 per GB - Regional load balancers: ~$18/month + per-GB charges - Premium tier networking (default): Better performance, higher cost - Standard tier: Lower cost, no global load balancing

Cloud CDN¶

What it is: Global content delivery network - cache content close to users for fast delivery

Business Value: - Improved user experience: Faster page loads and content delivery (higher conversion rates) - Reduced costs: Cache hits reduce backend load and egress charges - Global reach: 200+ edge locations worldwide serve content locally - Higher capacity: Offload traffic from origin servers (handle more users) - Better SEO: Page speed affects search rankings

Key Business Features: - Global edge locations: 200+ locations cache content close to users - Integration: Works seamlessly with Cloud Load Balancing - Cache control: Programmatic cache invalidation and management - Security: DDoS protection and SSL support included - Analytics: Detailed logs and metrics on cache performance

When to Choose Cloud CDN (Business Perspective): - Websites with global audience (improve international performance) - Media-heavy sites (images, videos consume bandwidth) - High-traffic websites (reduce origin server load) - E-commerce sites (faster pages = higher conversions) - APIs serving static or cacheable content

Business Use Cases: - Website acceleration (faster page loads worldwide) - Video streaming and media delivery - Software and file distribution (downloads, updates) - E-commerce sites (product images, static content) - Mobile app backends (reduce latency, data costs)

Cost Considerations: - Pay for cache egress and HTTP/HTTPS requests - Cache egress cheaper than origin egress ($0.04-$0.08/GB vs $0.08-$0.23/GB) - Cache hit ratio determines savings (80%+ hit ratio = significant savings) - Typical savings: 50-90% reduction in egress costs - Also reduces origin infrastructure costs (less load)

Cloud Interconnect and VPN¶

What it is: Hybrid connectivity options - connect on-premises infrastructure to GCP

Business Value: - Hybrid cloud: Leverage existing on-premises investments while moving to cloud - Data sovereignty: Keep sensitive data on-premises while using cloud for other workloads - Migration path: Gradual migration with connectivity to legacy systems - Lower costs: Reduce egress charges (Interconnect) for large data transfers - Better performance: Lower latency for hybrid applications

Cloud VPN (IPsec over Internet): - Site-to-site VPN: Connect on-premises network to GCP VPC - IPsec encryption: Secure tunnel over public internet - High availability: 99.9% SLA with redundant gateways - Cost-effective: ~$0.05/hour per tunnel + egress charges - Quick setup: Days to deploy, not months

Cloud Interconnect (Dedicated Connection): - Dedicated Interconnect: Direct physical connection to Google (10/100 Gbps) - Partner Interconnect: Connection through service provider (50 Mbps - 50 Gbps) - Higher bandwidth: Up to 200 Gbps (multiple connections) - Lower latency: Direct connection to Google network (no internet hops) - Reduced egress costs: Egress discounts for large data transfers

When to Choose Each (Business Perspective): - Cloud VPN: Small data transfers, quick setup, budget-conscious, test/dev - Dedicated Interconnect: Large data transfers (10+ Gbps), lowest latency, predictable performance - Partner Interconnect: Medium bandwidth needs, don't have own facility near Google

Business Use Cases: - Hybrid cloud architecture (workloads split between on-premises and cloud) - Disaster recovery (replicate data to/from cloud) - Gradual cloud migration (maintain connectivity during transition) - Burst to cloud (handle peak loads in cloud, regular workloads on-premises) - Access on-premises data from cloud applications

Cost Considerations: - Cloud VPN: $0.05/hour per tunnel (~$36/month) + standard egress - Dedicated Interconnect: Port fee ($1,700-$27,000/month) + reduced egress - Partner Interconnect: Service provider fee + Google fee + reduced egress - Egress savings: Interconnect can reduce egress from $0.08/GB to $0.02/GB - Break-even: Calculate based on data transfer volume (Interconnect better for >10TB/month)

AI and Machine Learning Services Overview¶

AI/ML enables business innovation without requiring deep technical expertise. The key question: Build custom models or use pre-trained?

Business Decision Framework: - Standard AI tasks (vision, language, speech)? Use Pre-trained APIs (fastest, cheapest) - Custom ML models needed? Use Vertex AI AutoML (no ML expertise required) - Advanced custom models? Use Vertex AI Custom Training (ML expertise required) - SQL-based ML for analysts? Use BigQuery ML (democratize ML)

AI and Machine Learning Services¶

Vertex AI¶

What it is: Unified ML platform for building, deploying, and scaling ML models - Google's comprehensive ML solution

Business Value: - Competitive advantage: Build custom AI capabilities unique to your business - Democratize ML: AutoML enables business analysts to build models without data scientists - Faster time-to-value: Pre-built components and workflows accelerate ML projects - MLOps built-in: Production-grade model deployment and management - Reduce ML complexity: Unified platform eliminates tool sprawl and integration challenges

Key Business Features: - AutoML: Build custom models without coding (business analysts can use) - Custom training: Train models with your own code (for data scientists) - Model deployment: Serve models with managed endpoints (no infrastructure management) - MLOps: End-to-end workflow management (from training to production) - Explainable AI: Understand model predictions (regulatory compliance, trust)

When to Choose Vertex AI (Business Perspective): - Pre-trained APIs don't fit your specific use case - Need custom models trained on your proprietary data - Building competitive advantage through unique ML capabilities - Require explainability for regulatory or business reasons - Have data science team or want to develop ML capabilities

Business Use Cases: - Custom computer vision (product defect detection, medical imaging) - Personalized recommendations unique to your business - Fraud detection tailored to your transaction patterns - Custom natural language models for industry-specific terminology - Predictive maintenance based on your equipment data

Cost Considerations: - Pay for training compute, prediction compute, and storage - AutoML: Higher cost but no data science team needed - Custom training: Lower cost but requires ML expertise - Prediction costs: Batch predictions cheaper than real-time - Consider total cost: service cost + staff cost

Pre-trained AI APIs (Quick Wins for Business)¶

What they are: Ready-to-use AI services requiring no ML expertise - call an API, get results

Business Value: - Immediate ROI: Deploy AI capabilities in days, not months - No ML expertise required: Developers use simple APIs - Google-quality AI: Leverage Google's AI research and infrastructure - Pay-per-use: No upfront costs, scale from zero - Focus on business problems: Build solutions, not train models

Vision AI¶

What it is: Image and video analysis through simple API calls

Key Business Features: - Image analysis: Detect objects, faces, text, inappropriate content in images - Video analysis: Analyze video content, detect scene changes, extract metadata - Optical Character Recognition (OCR): Extract text from images and documents - Product Search: Visual search for retail applications

Business Use Cases: - Content moderation (filter inappropriate user-generated content) - Retail product search (customers upload photos to find products) - Accessibility (generate image descriptions for visually impaired) - Quality control (detect defects in manufacturing) - Insurance claims processing (analyze damage photos)

Cost Considerations: - Pay per image analyzed (first 1,000 images/month free) - $1.50 per 1,000 images for most features - Custom models via AutoML Vision more expensive - Much cheaper than building own computer vision system

Natural Language AI¶

What it is: Text analysis and translation through API calls

Key Business Features: - Sentiment analysis: Determine positive, negative, neutral sentiment in text - Entity extraction: Identify people, organizations, locations, dates in text - Content classification: Categorize content by topic - Translation API: Translate text between 100+ languages

Business Use Cases: - Customer feedback analysis (analyze support tickets, reviews, surveys) - Content localization (translate website, apps, documents) - Email routing (automatically categorize and route customer emails) - Social media monitoring (track brand sentiment) - Compliance monitoring (flag sensitive content)

Cost Considerations: - Sentiment analysis: $1 per 1,000 text records - Entity extraction: $1 per 1,000 text records - Translation: $20 per million characters - Very cost-effective compared to manual analysis

Speech-to-Text and Text-to-Speech¶

What they are: Convert between audio and text through API calls

Key Business Features: - Speech-to-Text: Convert audio to text with 125+ language support - Text-to-Speech: Convert text to natural-sounding speech - Real-time streaming: Process audio streams in real-time - Speaker diarization: Identify different speakers in conversation

Business Use Cases: - Call center analytics (transcribe and analyze customer calls) - Voice assistants and IVR systems (interactive voice response) - Accessibility (closed captions, screen readers) - Meeting transcription and note-taking - Voice-controlled applications

Cost Considerations: - Speech-to-Text: $0.006 per 15 seconds of audio - Text-to-Speech: $4-$16 per million characters - Significant savings compared to manual transcription - Enables previously impossible use cases (analyze 100% of calls)

Document AI¶

What it is: Extract structured data from documents through AI

Key Business Features: - Document parsing: Extract data from invoices, receipts, contracts - Form processing: Process structured and unstructured forms - Custom extractors: Build custom parsers for your document types - Pre-trained parsers: Ready-to-use for common document types

Business Use Cases: - Invoice and receipt processing (accounts payable automation) - Contract analysis and data extraction - ID document verification (KYC processes) - Medical records digitization - Loan application processing

Cost Considerations: - Pay per page processed - Pre-trained parsers: $0.01-$0.10 per page depending on type - Significant labor savings (reduce manual data entry) - Faster processing = better customer experience

BigQuery ML¶

What it is: Build and deploy ML models using SQL queries in BigQuery

Business Value: - Democratize ML: Business analysts can build models using SQL (no Python/coding) - Data stays in place: No data movement (faster, more secure) - Familiar tools: Use existing BI tools and SQL knowledge - Serverless: No infrastructure to manage - Integrated workflow: From data to insights to predictions in one platform

Business Use Cases: - Customer churn prediction (identify at-risk customers) - Sales forecasting (predict future revenue) - Product recommendations (cross-sell, upsell opportunities) - Demand forecasting (inventory optimization) - Customer lifetime value prediction

Cost Considerations: - Pay for BigQuery query processing ($5 per TB) - Model training and prediction queries charged like normal queries - Much cheaper than separate ML platform (no data movement, no separate infrastructure) - Enables business analysts to build models (no data science team cost)

Data Analytics Services (Supporting Business Intelligence)¶

Pub/Sub¶

What it is: Messaging service for event-driven systems and real-time data pipelines

Business Value: - Real-time insights: Process business events as they happen - Reliable messaging: Never lose important business events (guaranteed delivery) - Scalability: Handle millions of messages per second - Decouple systems: Connect services without tight integration - Foundation for streaming: Enable real-time analytics and processing

Business Use Cases: - Real-time data ingestion for analytics - Event-driven microservices architecture - IoT data collection and processing - Streaming data to BigQuery for real-time dashboards - Order processing and notification systems

Dataflow¶

What it is: Fully managed stream and batch data processing service

Business Value: - Real-time and batch processing: One service for both use cases - Serverless: No infrastructure management - Auto-scaling: Handle variable data volumes automatically - Unified programming model: Apache Beam open standard (avoid lock-in)

Business Use Cases: - Real-time ETL (extract, transform, load) pipelines - Stream processing for real-time analytics - Batch data processing for reporting - Data enrichment and cleansing - Real-time fraud detection

Operations and Monitoring (Observability)¶

Business Value of Observability: - Reduce downtime: Detect and fix issues before customers notice - Faster troubleshooting: Find root cause quickly (reduce MTTR) - Performance optimization: Identify bottlenecks and inefficiencies - Cost optimization: Identify waste and unused resources - SLA compliance: Prove you're meeting service level agreements

Cloud Monitoring¶

What it is: Infrastructure and application monitoring service - visibility into your systems

Key Business Features: - Metrics collection: System and custom business metrics - Alerting: Proactive notifications prevent downtime - Dashboards: Visual displays for stakeholders and operations teams - Uptime monitoring: Website and API availability tracking (SLA reporting) - Integration: Works with third-party tools (ServiceNow, PagerDuty)

Business Value: - Early problem detection reduces customer impact - SLA tracking and reporting for contracts - Capacity planning through trend analysis - Cost optimization by identifying unused resources

Cloud Logging¶

What it is: Centralized logging service for applications and infrastructure - audit trail and troubleshooting

Key Business Features: - Log ingestion: Collect logs from all sources (applications, infrastructure, audit) - Search and filter: Query logs for troubleshooting and compliance - Export: Send logs to BigQuery for analysis, Cloud Storage for archival - Retention: Configurable policies for compliance requirements

Business Value: - Compliance and audit requirements (maintain audit trail) - Security investigation and forensics - Troubleshooting and root cause analysis - Business intelligence (analyze user behavior from logs)

Cloud Trace¶

What it is: Distributed tracing system for application performance analysis

Business Value: - Improve user experience by identifying slow operations - Optimize costs by finding inefficient operations - Troubleshoot performance issues in complex microservices - Capacity planning based on actual usage patterns

Error Reporting¶

What it is: Real-time error monitoring and alerting - know about errors before customers complain

Business Value: - Proactive error detection improves customer satisfaction - Prioritize fixes based on error frequency and impact - Faster time to resolution with detailed error information - Reduce support costs by fixing issues proactively

Business Scenarios: Choosing the Right Services¶

Scenario 1: E-commerce Startup (Cost-Conscious, Fast Time-to-Market)¶

Requirements: Build online store quickly, minimal operational overhead, unpredictable traffic

Recommended Services: - Compute: App Engine (auto-scaling, no infrastructure management) - Database: Firestore (serverless, real-time sync for cart, product catalog) - Storage: Cloud Storage (product images, Standard class) - CDN: Cloud CDN (fast image delivery worldwide) - Analytics: BigQuery (analyze sales data, customer behavior) - AI: Vision AI for product search, Recommendations AI for cross-selling

Why: Serverless services minimize operational overhead and cost. Pay only for what you use. Auto-scaling handles traffic spikes during sales events.

Scenario 2: Global Enterprise (Multi-Region, High Availability)¶

Requirements: Global customer base, 99.99%+ uptime, strong consistency, compliance

Recommended Services: - Compute: GKE with multi-region clusters (portability, high availability) - Database: Cloud Spanner (global consistency, 99.999% SLA) - Storage: Cloud Storage multi-region buckets - Network: Global Load Balancing, Cloud CDN, Cloud Armor (security) - Analytics: BigQuery (global analytics) - Operations: Full observability stack (Monitoring, Logging, Trace)

Why: Global reach with consistent low latency. High availability architecture meets enterprise SLAs. Compliance through audit logging and monitoring.

Scenario 3: Data Analytics Company (Process Massive Datasets)¶

Requirements: Analyze petabytes of data, real-time and batch processing, business intelligence

Recommended Services: - Data Warehouse: BigQuery (serverless, petabyte-scale) - Data Lake: Cloud Storage (Nearline/Coldline for historical data) - Real-time Processing: Pub/Sub + Dataflow (streaming data pipelines) - Batch Processing: Dataflow (batch ETL jobs) - ML: BigQuery ML (SQL-based models for analysts) - BI Tools: Looker, Data Studio (visualization)

Why: Serverless architecture eliminates infrastructure management. Separation of compute and storage optimizes costs. Analysts can work independently with SQL.

Scenario 4: Healthcare Provider (Compliance, Security, Hybrid Cloud)¶

Requirements: HIPAA compliance, data sovereignty, integrate with on-premises systems

Recommended Services: - Compute: Compute Engine in private VPC (full control for compliance) - Database: Cloud SQL with private IP (regional, compliance-friendly) - Storage: Cloud Storage with customer-managed encryption keys - Networking: Cloud VPN or Interconnect (secure hybrid connectivity) - AI: Healthcare API (HIPAA-compliant document processing) - Operations: Cloud Logging for audit trails, VPC Service Controls

Why: Full control and visibility for compliance. Hybrid connectivity maintains integration with on-premises systems. Healthcare-specific APIs designed for regulated environments.

Scenario 5: Mobile Gaming Company (Global Players, Real-Time)¶

Requirements: Global mobile game, real-time leaderboards, player data sync, scale to millions

Recommended Services: - Backend: Cloud Run (containerized game services, auto-scaling) - Database: Firestore (real-time sync, offline support for mobile) - Leaderboards: Cloud Spanner (global consistency for leaderboard integrity) - Player Data: Cloud Storage (game assets, saved games) - Analytics: BigQuery (player behavior analysis, game metrics) - AI: Recommendations AI (personalized offers, game suggestions)

Why: Real-time sync keeps all players in sync. Global infrastructure provides low latency worldwide. Auto-scaling handles player spikes during events.

Scenario 6: Financial Services (Regulated, High Security)¶

Requirements: Financial transactions, regulatory compliance, audit trails, strong consistency

Recommended Services: - Compute: GKE with binary authorization (control what runs) - Database: Cloud Spanner (ACID transactions globally, audit logging) - Storage: Cloud Storage with retention policies (compliance) - Security: VPC Service Controls, Cloud Armor, Security Command Center - Analytics: BigQuery (fraud detection, regulatory reporting) - AI: Fraud Detection AI, Explainable AI (regulatory compliance)

Why: Strong consistency essential for financial accuracy. Comprehensive audit logging for compliance. Advanced security controls protect sensitive data.

Scenario 7: IoT Manufacturing (High Volume, Time-Series Data)¶

Requirements: Millions of sensors, time-series data, predictive maintenance, real-time monitoring

Recommended Services: - Data Ingestion: Pub/Sub (collect sensor data at scale) - Data Processing: Dataflow (real-time processing for alerts) - Database: Bigtable (time-series data, high throughput) - Analytics: BigQuery (historical analysis, reporting) - ML: Vertex AI (predictive maintenance models) - Visualization: Data Studio (real-time dashboards for operations)

Why: Pub/Sub handles millions of messages per second. Bigtable optimized for time-series data. Real-time processing enables immediate alerts for anomalies.

Scenario 8: Media Streaming Platform (High Bandwidth, Global Audience)¶

Requirements: Video streaming worldwide, high bandwidth, low latency, cost efficiency

Recommended Services: - Storage: Cloud Storage (video files, lifecycle management for old content) - CDN: Cloud CDN (cache video close to users globally) - Compute: Cloud Run (transcode videos on-demand) - Networking: Premium tier for consistent performance - Analytics: BigQuery (viewer analytics, content performance) - AI: Video Intelligence API (content moderation, metadata extraction)

Why: Cloud CDN reduces origin load and egress costs. Cloud Storage lifecycle management archives old content. Global infrastructure serves viewers worldwide with low latency.

Service Selection Decision Framework¶

Quick Selection Guide¶

For Compute: 1. Lift-and-shift migration? Compute Engine 2. Zero infrastructure management? App Engine 3. Containers + simplicity? Cloud Run 4. Containers + complex orchestration? GKE 5. Event-driven, short tasks? Cloud Functions

For Storage: 1. Unstructured data (files, images)? Cloud Storage (choose class by access frequency) 2. VM disk storage? Persistent Disks 3. Shared file storage? Filestore

For Database: 1. Relational, regional? Cloud SQL 2. Relational, global? Cloud Spanner 3. Document, mobile/web? Firestore 4. Analytics/data warehouse? BigQuery 5. Time-series, IoT? Bigtable

For AI/ML: 1. Standard AI tasks (vision, language)? Pre-trained APIs 2. Custom models, no ML expertise? Vertex AI AutoML 3. Advanced custom models? Vertex AI Custom Training 4. SQL-based ML for analysts? BigQuery ML

Cloud Digital Leader Exam Tips¶

Service Questions Patterns¶

Pattern 1: "Choose the right service" - Focus on business requirements, not technical implementation - Key factors: cost, operational overhead, time-to-market, scale - Remember: serverless = lower operational overhead = faster time-to-market

Pattern 2: "Cost optimization" - Serverless services often more cost-effective for variable workloads - Storage classes: choose based on access frequency - Preemptible/Spot instances: 80% savings for fault-tolerant workloads - Sustained use and committed use discounts for predictable workloads

Pattern 3: "Global scale requirements" - Cloud Spanner for global SQL - Global Load Balancing for global web apps - Cloud CDN for global content delivery - Multi-region Cloud Storage for global data access

Pattern 4: "Compliance and security" - Audit logging available on all services - Private IP options for databases - Customer-managed encryption keys available - VPC Service Controls for additional security perimeter

Pattern 5: "Migration scenarios" - Lift-and-shift: Compute Engine - Replatform: Cloud SQL, managed services - Refactor: Serverless (App Engine, Cloud Run, Cloud Functions) - Database migration: Database Migration Service

Key Exam Concepts¶

Compute Services: - Understand the trade-off: control vs. operational overhead - IaaS (Compute Engine) = most control, most operational work - PaaS (App Engine) = less control, less operational work - Serverless (Cloud Run, Cloud Functions) = least control, zero operational work - Know when to choose each based on business needs

Storage Services: - Storage class selection is critical for cost optimization - Understand minimum storage duration for Nearline, Coldline, Archive - Cloud Storage for unstructured data (objects) - Persistent Disks for VM storage (block storage) - Filestore for shared file systems (NFS)

Database Services: - Match database type to use case (relational, document, time-series, analytics) - Cloud SQL: regional relational (start here for most use cases) - Cloud Spanner: global relational (premium service, higher cost) - Firestore: document database for mobile/web (real-time sync) - BigQuery: analytics and data warehouse (serverless, petabyte-scale) - Bigtable: time-series and IoT (high throughput, requires scale)

Networking Services: - VPC is foundational (every deployment uses VPC) - Global Load Balancing enables global applications with single IP - Cloud CDN reduces costs and improves performance - Cloud VPN for quick hybrid connectivity - Cloud Interconnect for high-bandwidth hybrid connectivity

AI/ML Services: - Pre-trained APIs: fastest path to AI capabilities (no ML expertise needed) - Vertex AI AutoML: custom models without ML expertise - BigQuery ML: democratize ML for SQL users - Focus on business value, not technical details

Data Analytics: - BigQuery is the central service (data warehouse + analytics + ML) - Pub/Sub for real-time data ingestion - Dataflow for data processing (ETL) - Know when to use real-time vs. batch processing

Common Exam Traps¶

1. Don't over-engineer: Choose simplest service that meets requirements
2. Serverless usually wins for cost: Unless you have consistent 24/7 load
3. Global doesn't always mean Cloud Spanner: Consider multi-region Cloud SQL
4. Not everything needs Kubernetes: Cloud Run often simpler choice
5. Pre-trained APIs before custom ML: Don't build what already exists
6. Consider total cost: Include operational overhead, not just service cost
7. Time-to-market matters: Managed services accelerate delivery

Key Success Factors for Exam¶

• Think like a business leader: Focus on outcomes, not technical details
• Cost optimization: Always consider cost implications
• Operational overhead: Serverless reduces operational complexity
• Time-to-market: Managed services enable faster delivery
• Scalability: Cloud services scale automatically
• Global reach: Google's global infrastructure is a key advantage
• Business value: Always connect technical capabilities to business outcomes

Remember: Cloud Digital Leader is Business-Focused¶

The exam tests business understanding of cloud services: - WHAT services do for business - WHY choose one service over another - WHEN to use each service based on business requirements - Business outcomes (cost, time-to-market, scalability, etc.)

NOT testing: - HOW to implement technically - Command-line syntax or console navigation - Deep technical architecture details - Coding or scripting