Design Multi-Cloud Architectures
Provides a multi-cloud decision framework across AWS, Azure, GCP: service comparison, patterns, cost, and migration.
Why it matters
Architect cloud-agnostic applications and make informed decisions for service selection across AWS, Azure, and GCP. This skill provides a decision framework and patterns for designing resilient and cost-effective multi-cloud strategies.
Outcomes
What it gets done
Design multi-cloud strategies and select optimal services.
Implement cloud-agnostic architectures using best practices.
Compare compute, storage, and database services across major cloud providers.
Develop phased migration strategies and cost optimization plans.
Install
Add it to your toolbox
Run in your project directory:
curl -fsSL https://spark.entire.vc/get/ag-multi-cloud-architecture | bash Overview
Multi-Cloud Architecture
Provides a multi-cloud decision framework covering AWS/Azure/GCP service comparison, architecture patterns, cost optimization, and phased migration strategy. Use when designing a multi-cloud strategy, migrating between providers, or building cloud-agnostic architecture.
What it does
Provides a decision framework and architecture patterns for building applications across AWS, Azure, and GCP - covering service comparison, multi-cloud patterns, cloud-agnostic abstraction, cost optimization, and migration strategy.
When to use - and when NOT to
Use this skill when designing a multi-cloud strategy, migrating between cloud providers, selecting cloud services for a specific workload, implementing a cloud-agnostic architecture, or optimizing costs across providers. Not a fit for single-cloud-only projects with no cross-provider requirements or for application-level (non-infrastructure) architecture decisions.
Inputs and outputs
Provides direct service comparison tables across AWS/Azure/GCP for compute (EC2/VMs/Compute Engine, ECS/Container Instances/Cloud Run, EKS/AKS/GKE, Lambda/Functions/Cloud Functions), storage (S3/Blob/Cloud Storage, EBS/Managed Disks/Persistent Disk, EFS/Azure Files/Filestore, Glacier/Archive Storage), and database services (RDS/SQL Database/Cloud SQL, DynamoDB/Cosmos DB/Firestore, Aurora/Cloud Spanner, ElastiCache/Cache for Redis/Memorystore).
Defines four multi-cloud patterns: single provider with disaster recovery in another cloud (with database replication and automated failover), best-of-breed (using each provider's strongest service, e.g. AI/ML on GCP, enterprise apps on Azure, general compute on AWS), geographic distribution (serving users from the nearest cloud region for data sovereignty and global load balancing), and cloud-agnostic abstraction (Kubernetes for compute, PostgreSQL for database, S3-compatible storage like MinIO, open source tooling).
For cloud-agnostic architecture, recommends specific cloud-native alternatives: Kubernetes (EKS/AKS/GKE) for compute, PostgreSQL/MySQL for database, Apache Kafka for message queuing, Redis for caching, S3-compatible APIs for object storage, Prometheus/Grafana for monitoring, and Istio/Linkerd for service mesh - organized through an abstraction layer stack (application to Terraform-based infrastructure abstraction to cloud provider APIs).
Cost guidance compares compute pricing models (AWS on-demand/reserved/spot/savings plans, Azure pay-as-you-go/reserved/spot, GCP on-demand/committed use/preemptible) and lists eight cost optimization strategies including reserved capacity (30-70% savings), spot/preemptible instances, right-sizing, and cost allocation tags.
Migration strategy follows four phases: assessment (infrastructure inventory, dependency mapping, cost estimation), pilot (select and test a pilot workload), migration (incremental migration with dual-run period), and optimization (right-sizing, cloud-native services adoption). Best practices include using infrastructure as code (Terraform/OpenTofu), CI/CD pipelines, designing for cross-cloud failure, and testing disaster recovery procedures.
Integrations
Compares and integrates AWS, Azure, and GCP native services, with cloud-agnostic alternatives built on Kubernetes, PostgreSQL, Apache Kafka, Redis, Terraform/OpenTofu, Prometheus/Grafana, and Istio/Linkerd.
Who it's for
Cloud architects designing, migrating, or optimizing multi-cloud systems who need direct service-to-service comparisons, concrete architecture patterns, and a phased migration approach rather than a single-provider deep dive.
Application Layer -> Infrastructure Abstraction (Terraform) -> Cloud Provider APIs -> AWS / Azure / GCP
FAQ
Common questions
Discussion
Questions & comments · 0
Sign In Sign in to leave a comment.