Complete Feature Mapping: AWS Analytics to Azure
Comparative positioning note
This document is written from the perspective of Microsoft Azure, Cloud Scale Analytics, and CSA Loom. Any description of third-party or competing products, services, pricing, or capabilities is derived from publicly available documentation and sources believed accurate at the time of writing, and is provided for general comparison only. We do not claim expertise in, or authority over, any non-Microsoft product or service; the respective vendor's official documentation is the authoritative source for their offerings, which may change over time. Nothing here is intended to disparage any vendor — where a competing product has genuine advantages, we aim to note them honestly. Verify all third-party details against the vendor's current official documentation before making decisions.
The definitive feature-by-feature reference for mapping every AWS analytics capability to its Microsoft Azure equivalent.
Audience: Platform architects, migration leads, and technical evaluators Last updated: 2026-04-30
Summary
This reference maps 103 AWS analytics and infrastructure features across 12 capability domains to their Azure equivalents. Each mapping includes migration complexity, the CSA-in-a-Box evidence path (where the pattern exists in the repository), and notes on gaps or limitations.
| Metric | Count |
| Total features mapped | 103 |
| Full parity or better (XS-M effort) | 90 |
| Partial parity (L effort) | 10 |
| Known gaps (XL or no equivalent) | 3 |
Migration complexity key
| Rating | Description | Typical effort |
| XS | Drop-in replacement or native Azure capability | < 1 day |
| S | Minor configuration or adaptation required | 1-3 days |
| M | Moderate development; requires design decisions | 1-3 weeks |
| L | Significant development; architectural changes | 1-3 months |
| XL | Major initiative; phased delivery | 3+ months |
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 1 | S3 Standard | General-purpose object storage | ADLS Gen2 (hot tier) / OneLake | S | csa_platform/unity_catalog_pattern/onelake_config.yaml | Hierarchical namespace provides directory-level operations |
| 2 | S3 Intelligent-Tiering | Automatic tier optimization | ADLS Gen2 lifecycle management policies | S | N/A --- use Azure native | Rule-based tiering: hot, cool, archive |
| 3 | S3 Infrequent Access | Low-cost infrequent storage | ADLS Gen2 cool tier | XS | N/A --- use Azure native | Direct cost mapping |
| 4 | S3 Glacier / Deep Archive | Archival storage | ADLS Gen2 archive tier | XS | N/A --- use Azure native | Rehydration latency differs; plan accordingly |
| 5 | S3 Object Lock (WORM) | Immutable write-once storage | ADLS Gen2 immutable storage (time-based / legal hold) | XS | N/A --- use Azure native | 1:1 compliance-driven retention |
| 6 | S3 Versioning | Object version history | ADLS Gen2 blob versioning + Delta Lake time travel | S | ADR-0003 | Delta time travel provides richer versioning than object-level |
| 7 | S3 Lifecycle Policies | Automated tier transitions and expiration | ADLS Gen2 lifecycle management policies | XS | N/A --- use Azure native | 1:1 rule-set translation |
| 8 | S3 Access Points | Simplified per-application access | Private endpoints + RBAC + ABAC on containers | S | docs/SELF_HOSTED_IR.md | ACL-level access maps to RBAC + ABAC |
| 9 | S3 Cross-Region Replication | Geo-redundant replication | ADLS Gen2 object replication + GRS/GZRS | S | docs/DR.md, docs/MULTI_REGION.md | Azure provides multiple redundancy options |
| 10 | S3 Event Notifications | Event triggers on object changes | Event Grid (BlobCreated / BlobDeleted) | S | csa_platform/data_activator/ | Event Grid is the native Azure event routing service |
| 11 | S3 Select | In-place query of S3 objects | ADLS Gen2 query acceleration (preview) / Fabric SQL endpoint | M | N/A --- use Azure native | Fabric SQL endpoint is the recommended path |
| 12 | EBS (Elastic Block Store) | Block storage for EC2 | Azure Managed Disks | XS | N/A --- infrastructure, not analytics | Direct equivalent; not part of analytics migration |
| 13 | EFS (Elastic File System) | Managed NFS file storage | Azure Files (NFS or SMB) / Azure NetApp Files | S | N/A --- use Azure native | ANF for high-performance NFS workloads |
2. Compute: data warehousing (Redshift)
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 14 | Redshift RA3 clusters | Managed columnar data warehouse | Databricks SQL Warehouses + Delta Lake on ADLS Gen2 | M | csa_platform/unity_catalog_pattern/, ADR-0003 | RA3 decoupling maps to Databricks compute-storage separation |
| 15 | Redshift Serverless | On-demand serverless warehouse | Databricks SQL Serverless | S | ADR-0002, ADR-0010 | RPU model maps to DBU model |
| 16 | Spectrum (external tables) | Query S3 data without loading | OneLake shortcuts + Databricks Lakehouse Federation | S | csa_platform/unity_catalog_pattern/onelake_config.yaml | Zero-copy read pattern preserved |
| 17 | Materialized views | Pre-computed query results | dbt incremental models + Databricks materialized views | M | domains/shared/dbt/dbt_project.yml, ADR-0001 | Most MVs re-express as dbt incremental |
| 18 | Stored procedures | PL/pgSQL-style server-side logic | dbt macros + Databricks SQL UDFs + notebook jobs | L | domains/finance/dbt/macros/, domains/shared/dbt/macros/ | Complex imperative SPs require notebook rewrite |
| 19 | WLM (Workload Management) | Queue-based query prioritization | Databricks SQL Warehouse sizing + serverless auto-scale | M | csa_platform/multi_synapse/rbac_templates/ | Each WLM queue becomes a SQL Warehouse |
| 20 | Distribution + sort keys | Physical data distribution strategy | Delta partitioning + Z-ordering | S | ADR-0003 | Distribution key becomes partition column; sort keys become ZORDER columns |
| 21 | Concurrency scaling | Auto-scale for burst query load | Databricks SQL Warehouse serverless auto-scale | XS | ADR-0010 | Serverless handles burst natively |
| 22 | Data sharing | Cross-account data sharing | Delta Sharing + OneLake shortcuts | M | csa_platform/data_marketplace/ | Open protocol; Purview data-product registry |
| 23 | Federated queries | Query RDS/Aurora from Redshift | Databricks Lakehouse Federation + ADF linked services | M | csa_platform/unity_catalog_pattern/ | Native connectors for Postgres/MySQL/SQL Server |
| 24 | Redshift ML | In-warehouse ML training via SageMaker | Databricks ML + Feature Store + MLflow | M | csa_platform/ai_integration/model_serving/ | Tighter ML integration than Redshift ML |
3. Compute: Spark and Hadoop (EMR)
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 25 | EMR on EC2 | Managed Spark/Hadoop/Hive/Presto cluster | Azure Databricks | M | csa_platform/unity_catalog_pattern/, ADR-0002 | Almost every Spark workload maps to Databricks |
| 26 | EMR Serverless | Serverless Spark/Hive execution | Databricks Serverless Jobs | S | ADR-0002, ADR-0010 | Per-job compute maps to serverless jobs |
| 27 | EMR Studio | Managed notebook IDE | Databricks Workspace Notebooks + Git integration | S | domains/shared/notebooks/ | 1:1 notebook-and-repo UX |
| 28 | EMR on EKS | Kubernetes-native Spark | Databricks (managed containers) / AKS + Spark Operator | L | N/A | Databricks manages own containers; AKS for custom K8s |
| 29 | Bootstrap actions | Cluster initialization scripts | Databricks init scripts + cluster policies | XS | csa_platform/unity_catalog_pattern/deploy/ | 1:1 semantic mapping |
| 30 | Managed scaling | Auto-scale cluster workers | Databricks cluster autoscaling + serverless | XS | ADR-0002 | Serverless removes tuning burden |
| 31 | Spot instances (EMR) | Low-cost interruptible compute | Databricks Spot on Azure (Azure Spot VMs) | XS | csa_platform/unity_catalog_pattern/deploy/ | Direct 1:1 mapping |
| 32 | Hive Metastore (EMR) | Metadata catalog for Hive tables | Unity Catalog (primary) + external Hive metastore | M | csa_platform/unity_catalog_pattern/unity_catalog/ | Unity Catalog is target; bridge via external metastore |
| 33 | EMRFS | S3-backed file system for EMR | ADLS Gen2 (abfss://) + OneLake | S | N/A --- use Azure native | Direct path substitution in Spark configs |
4. Compute: ad-hoc queries (Athena)
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 34 | Athena SQL queries | Serverless SQL over S3 | Databricks SQL + OneLake shortcuts to S3 | S | csa_platform/unity_catalog_pattern/onelake_config.yaml | S3 stays read-only during migration |
| 35 | Athena workgroups | Cost/access isolation per group | Databricks SQL Warehouses (one per workgroup) | XS | docs/COST_MANAGEMENT.md | Auto-stop + Azure budgets for cost control |
| 36 | Athena federated queries | Query non-S3 sources (DynamoDB, RDS) | Databricks Lakehouse Federation | S | ADR-0002 | Native connectors for Postgres/MySQL/SQL Server/Snowflake |
| 37 | Athena ACID (Iceberg) | ACID transactions on Athena | Delta Lake ACID (primary) + Iceberg read compatibility | S | ADR-0003 | Databricks reads Iceberg natively during migration |
| 38 | Athena Spark | Interactive Spark sessions in Athena | Databricks Interactive Notebooks | S | domains/shared/notebooks/ | Richer notebook experience |
| 39 | CTAS / INSERT OVERWRITE | Create table as select patterns | dbt models + MERGE INTO on Delta | S | domains/shared/dbt/ | Idempotent merges replace CTAS idioms |
| 40 | Athena Provisioned Capacity | Dedicated compute reservation | Databricks SQL Pro warehouses | S | ADR-0002 | Dedicated compute with auto-scale |
5. ETL and orchestration (Glue and Step Functions)
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 41 | Glue Data Catalog | Centralized metadata catalog | Unity Catalog (runtime) + Purview (business catalog) | M | csa_platform/csa_platform/governance/purview/, ADR-0006 | Unity Catalog holds runtime metadata; Purview holds lineage and glossary |
| 42 | Glue ETL Jobs (PySpark) | Managed Spark ETL | Databricks Jobs + dbt models + ADF activities | M | domains/shared/notebooks/, domains/shared/pipelines/adf/ | PySpark moves to Databricks; SQL logic to dbt |
| 43 | Glue Python Shell | Lightweight Python jobs | Azure Functions / small Databricks Python tasks | S | csa_platform/functions/ | Serverless functions for lightweight jobs |
| 44 | Glue Crawlers | Schema discovery and catalog population | Purview scan jobs + Databricks Auto Loader schema inference | M | csa_platform/csa_platform/governance/purview/purview_automation.py | Purview for governance; Auto Loader for runtime schema |
| 45 | Glue Studio (visual ETL) | Drag-and-drop ETL designer | ADF Mapping Data Flows / Fabric Data Factory visual | M | domains/shared/pipelines/adf/ | Visual design in ADF; transformation logic in dbt |
| 46 | Glue Streaming | Spark Structured Streaming ETL | Databricks Structured Streaming + Event Hubs | M | ADR-0005, examples/iot-streaming/ | Kinesis source replaced with Event Hubs |
| 47 | Glue DataBrew | Visual data prep tool | Power Query (Fabric) + dbt + Databricks SQL | S | domains/shared/dbt/dbt_project.yml | Most transforms re-express as Power Query or dbt |
| 48 | Glue Data Quality | Assertion-based data checks | dbt tests + Great Expectations + data-product contracts | S | domains/finance/data-products/invoices/contract.yaml | dbt tests are more expressive |
| 49 | Step Functions | Serverless workflow orchestration | ADF pipeline activities + Logic Apps | M | domains/shared/pipelines/adf/ | ADF for data orchestration; Logic Apps for integration workflows |
| 50 | EventBridge | Event bus for decoupled services | Event Grid + Service Bus | S | csa_platform/data_activator/ | Event Grid for events; Service Bus for messaging |
6. Business intelligence (QuickSight)
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 51 | QuickSight dashboards | Interactive dashboards and analyses | Power BI reports and dashboards | M | N/A --- use Power BI native | Manual rebuild; no automated migration tool |
| 52 | SPICE | In-memory analytics engine | Power BI Import mode / Direct Lake | S | N/A --- use Power BI native | Direct Lake eliminates import refresh entirely |
| 53 | QuickSight Q | Natural language querying | Power BI Copilot | S | N/A --- use Power BI native | Copilot uses GPT-4 for richer NL interaction |
| 54 | Calculated fields | Custom computed columns | DAX measures and calculated columns | M | N/A --- use Power BI native | DAX is more expressive but has a learning curve |
| 55 | Parameters | Dashboard parameterization | Power BI slicers + parameters + bookmarks | S | N/A --- use Power BI native | Richer parameterization options |
| 56 | Row-level security | Per-user data filtering | Power BI RLS + Entra ID groups | S | N/A --- use Power BI native | Dynamic RLS via DAX + Entra ID |
| 57 | QuickSight embedding | Embed dashboards in apps | Power BI Embedded / Power BI embed in Teams | S | N/A --- use Power BI native | Broader embedding targets (Teams, SharePoint, custom apps) |
7. Streaming (Kinesis and MSK)
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 58 | Kinesis Data Streams | Real-time data streaming | Event Hubs | M | ADR-0005 | Shard model maps to partition model |
| 59 | Kinesis Data Firehose | Managed delivery to storage/analytics | Event Hubs Capture / ADF streaming | S | ADR-0005 | Event Hubs Capture writes directly to ADLS Gen2 |
| 60 | Kinesis Data Analytics | SQL/Flink-based stream processing | Stream Analytics / Fabric Real-Time Intelligence | M | N/A --- use Azure native | Stream Analytics for SQL; Fabric RTI for complex event processing |
| 61 | MSK (Managed Kafka) | Managed Apache Kafka | Event Hubs with Kafka protocol (AMQP + Kafka wire) | M | ADR-0005 | Kafka clients connect with endpoint/config change only |
| 62 | MSK Connect | Managed Kafka Connect | Event Hubs + ADF connectors / Kafka Connect on AKS | M | N/A --- use Azure native | ADF connectors cover most source/sink patterns |
8. AI and ML (SageMaker and Bedrock)
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 63 | SageMaker Studio | Managed ML IDE | Azure ML Studio / Databricks ML / AI Foundry | M | csa_platform/ai_integration/ | Multiple options depending on workflow |
| 64 | SageMaker Training | Managed training compute | Azure ML Compute + Databricks ML clusters | M | csa_platform/ai_integration/model_serving/ | Direct mapping; GPU SKUs available |
| 65 | SageMaker Endpoints | Real-time ML inference | Azure ML Managed Endpoints / AKS | M | csa_platform/ai_integration/model_serving/ | Managed endpoints simplify deployment |
| 66 | SageMaker Pipelines | ML workflow orchestration | Azure ML Pipelines / Prompt Flow | M | N/A --- use Azure native | Prompt Flow for LLM-centric workflows |
| 67 | SageMaker Feature Store | Managed feature store | Databricks Feature Store / Azure ML Feature Store | M | N/A --- use Azure native | Databricks Feature Store integrates with Unity Catalog |
| 68 | Bedrock | Managed LLM access | Azure OpenAI Service | S | csa_platform/ai_integration/ | GPT-4o, GPT-4.1, o3, o4-mini available in Azure Gov |
| 69 | Bedrock Agents | Autonomous AI agents | Azure AI Agents / Copilot Studio | M | N/A --- use Azure native | Copilot Studio for no-code; AI Agents SDK for code |
| 70 | Bedrock Knowledge Bases | RAG with managed retrieval | Azure AI Search + Azure OpenAI | M | N/A --- use Azure native | AI Search provides vector + hybrid search |
9. Governance and security
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 71 | IAM roles and policies | Identity and access management | Entra ID + Azure RBAC + ABAC | M | csa_platform/multi_synapse/rbac_templates/ | Role-per-service maps to RBAC assignments |
| 72 | Lake Formation | Fine-grained data access control | Purview + Unity Catalog access control | L | csa_platform/csa_platform/governance/purview/, ADR-0006 | Column-level and row-level security in Unity Catalog |
| 73 | KMS | Key management and encryption | Azure Key Vault | S | N/A --- use Azure native | CMK for all data-at-rest encryption |
| 74 | Secrets Manager | Secret storage and rotation | Azure Key Vault secrets | XS | N/A --- use Azure native | Direct mapping; auto-rotation supported |
| 75 | CloudTrail | API audit logging | Azure Monitor Activity Log + Diagnostic Settings | S | N/A --- use Azure native | Richer integration with Log Analytics |
| 76 | GuardDuty | Threat detection | Microsoft Defender for Cloud | S | N/A --- use Azure native | Broader threat detection across Azure services |
| 77 | CloudWatch | Monitoring and alerting | Azure Monitor + Log Analytics | S | N/A --- use Azure native | Unified monitoring across all Azure services |
| 78 | X-Ray | Distributed tracing | Application Insights | S | N/A --- use Azure native | Part of Azure Monitor; OpenTelemetry compatible |
10. DevOps and infrastructure
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 79 | CloudFormation | Infrastructure as Code | Bicep (primary) / Terraform | M | ADR-0004 docs/adr/0004-bicep-over-terraform.md | Bicep chosen for Azure policy evidence |
| 80 | CDK | Programmatic IaC | Bicep with modules / Terraform CDK | M | ADR-0004 | Bicep modules provide composability |
| 81 | CodePipeline | CI/CD pipeline | GitHub Actions / Azure DevOps Pipelines | S | .github/workflows/ | Standard CI/CD; richer marketplace |
| 82 | CodeBuild | Managed build service | GitHub Actions runners / Azure DevOps hosted agents | S | .github/workflows/ | Direct mapping |
| 83 | AWS Organizations | Multi-account management | Azure Management Groups + Subscriptions | M | N/A --- use Azure native | 4-subscription pattern in CSA-in-a-Box |
| 84 | Service Control Policies | Organizational guardrails | Azure Policy + Blueprints | M | N/A --- use Azure native | Azure Policy provides deny/audit/deploy-if-not-exists |
| 85 | AWS Config | Resource configuration tracking | Azure Policy + Azure Resource Graph | S | N/A --- use Azure native | Resource Graph enables advanced queries |
11. Networking and data transfer
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 86 | VPC | Virtual private cloud | Azure Virtual Network (VNet) | M | N/A --- use Azure native | Similar concepts; different defaults |
| 87 | VPC Endpoints (Gateway) | Private access to S3/DynamoDB | Service Endpoints | XS | N/A --- use Azure native | Route to service via backbone |
| 88 | VPC Endpoints (Interface) | Private access to other services | Private Endpoints | S | N/A --- use Azure native | Private IP for PaaS service |
| 89 | AWS PrivateLink | Private service connectivity | Azure Private Link | S | N/A --- use Azure native | Same concept |
| 90 | Security Groups | Stateful instance-level firewall | Network Security Groups (NSGs) | S | N/A --- use Azure native | Stateful; applied at NIC or subnet |
| 91 | NACLs | Stateless subnet-level firewall | NSGs (at subnet level) | S | N/A --- use Azure native | NSGs are stateful; applied at subnet scope |
| 92 | Direct Connect | Dedicated private connectivity | ExpressRoute | M | N/A --- use Azure native | Dedicated private connection |
| 93 | Transit Gateway | Hub-and-spoke networking | Azure Virtual WAN / VNet Peering | M | N/A --- use Azure native | Hub-and-spoke or mesh topology |
| 94 | NAT Gateway | Outbound internet for private subnets | Azure NAT Gateway | XS | N/A --- use Azure native | Direct equivalent |
| 95 | S3 Transfer Acceleration | Accelerated upload to S3 | Azure CDN / Front Door (for upload patterns) | S | N/A --- use Azure native | Different approach; CDN for distribution |
12. Application integration
| # | AWS feature | Description | Azure equivalent | Complexity | CSA-in-a-Box evidence | Notes |
| 96 | Lambda | Serverless compute | Azure Functions | M | csa_platform/functions/ | Direct equivalent; different trigger model |
| 97 | API Gateway | Managed REST/WebSocket API | Azure API Management | M | N/A --- use Azure native | Richer policy engine |
| 98 | SQS | Managed message queue | Azure Queue Storage / Service Bus | S | N/A --- use Azure native | Service Bus for enterprise messaging |
| 99 | SNS | Managed pub/sub notifications | Event Grid / Service Bus Topics | S | N/A --- use Azure native | Event Grid for event-driven |
| 100 | DynamoDB | Managed NoSQL database | Cosmos DB (NoSQL API) | M | N/A --- use Azure native | Multi-model; global distribution |
| 101 | ElastiCache | Managed Redis/Memcached | Azure Cache for Redis | S | N/A --- use Azure native | Direct equivalent |
| 102 | RDS / Aurora | Managed relational database | Azure SQL / Azure Database for PostgreSQL | S | N/A --- use Azure native | Direct equivalents per engine |
| 103 | Cognito | User authentication and authorization | Entra External ID / Azure AD B2C | M | N/A --- use Azure native | Different architecture |
Migration complexity summary
By effort level
| Effort | Count | Percentage | Description |
| XS | 18 | 17% | Drop-in replacement; < 1 day |
| S | 33 | 32% | Minor adaptation; 1-3 days |
| M | 39 | 38% | Moderate development; 1-3 weeks |
| L | 10 | 10% | Significant development; 1-3 months |
| XL | 3 | 3% | Major initiative; 3+ months |
By domain
| Domain | Features | Avg. complexity | Highest risk |
| Storage | 13 | S | Minimal; strong parity |
| Data warehousing | 11 | M | Stored procedure migration (L) |
| Spark/Hadoop | 9 | S-M | EMR on EKS (L) |
| Ad-hoc queries | 7 | S | Partition projection (M) |
| ETL/orchestration | 10 | M | Glue Streaming (M) |
| BI | 7 | S-M | Dashboard rebuild is manual |
| Streaming | 5 | M | MSK Connect connectors vary |
| AI/ML | 8 | M | SageMaker Pipeline conversion (M) |
| Security/governance | 8 | S-M | Lake Formation tag-based access (L) |
| DevOps | 7 | S-M | CloudFormation to Bicep (M) |
| Networking | 10 | S-M | Transit Gateway to Virtual WAN (M) |
| Application integration | 8 | S-M | DynamoDB to Cosmos DB (M) |
Migration priority recommendation
For a typical federal analytics migration, the recommended order based on dependency and risk:
- Storage (S3 to ADLS/OneLake): Foundation for everything else; OneLake shortcuts enable immediate bridge
- Identity (IAM to Entra ID/RBAC): Required before any workload migration
- Catalog (Glue to Unity Catalog/Purview): Required for compute migration
- Compute (Redshift/EMR/Athena to Databricks): Core workload migration
- ETL (Glue to ADF/dbt): Depends on catalog and compute
- Streaming (Kinesis/MSK to Event Hubs): Independent; can parallelize
- BI (QuickSight to Power BI): Depends on compute and catalog
- AI/ML (SageMaker/Bedrock to Azure AI): Often independent track
- Monitoring (CloudWatch to Azure Monitor): Throughout migration
- Networking (VPC to VNet): Deploy early; configure throughout
Gap summary
| # | AWS feature | Gap description | Workaround | Severity |
| 1 | EMR on EKS | No direct Kubernetes-native Spark equivalent in Databricks | Use AKS + Spark Operator for K8s-specific requirements; Databricks manages containers internally | Low --- affects only K8s-native Spark users |
| 2 | Glue DataBrew visual transforms | Power Query + dbt covers most cases; some point-and-click transforms require SQL rewrite | Document each DataBrew job; rewrite as dbt model or Power Query step | Low --- documented pattern |
| 3 | Athena partition projection | No direct equivalent for dynamic partition inference | Use Delta Lake auto-partitioning + Databricks partition pruning | Low --- Delta handles this differently but effectively |
| 4 | Lake Formation tag-based access | Unity Catalog uses catalog/schema/table grants; tag-based access is a roadmap item | Use Unity Catalog row filters and column masks for fine-grained access | Medium --- different model but functional |
| 5 | Redshift SUPER type | No native semi-structured column type in Delta | Store as STRING with JSON functions; use : notation for field access in Databricks SQL | Low --- JSON functions cover all use cases |
| 6 | Redshift Concurrency Scaling free tier | No equivalent free burst capacity | Databricks Serverless auto-scales without a free-tier concept; cost is per-DBU | Low --- serverless pricing is competitive |
AWS services explicitly out of scope
The following AWS services are not part of the analytics migration and are not mapped in this document:
| Service | Reason | Azure equivalent (if relevant) |
| EC2 (general compute) | Infrastructure, not analytics | Azure Virtual Machines |
| ECS / Fargate | Container orchestration | Azure Container Apps / AKS |
| Route 53 | DNS management | Azure DNS |
| CloudFront | CDN | Azure Front Door / CDN |
| Elastic Load Balancing | Load balancing | Azure Load Balancer / Application Gateway |
| AWS Backup | Backup management | Azure Backup |
| Systems Manager | Operations management | Azure Automation |
These services may be relevant to a broader cloud migration but are not addressed in this analytics-focused feature mapping.
How to use this document
- For migration planning: Filter to your specific AWS services. Not every row applies to every migration.
- For effort estimation: Use the complexity column to build a rough work-breakdown structure. XS and S items can often be handled in parallel; L items need dedicated sprint capacity.
- For gap assessment: Review the gap summary to identify areas requiring architectural decisions before migration.
- For executive communication: Use the migration complexity summary to communicate risk and effort to stakeholders.
Last updated: 2026-04-30 Maintainers: CSA-in-a-Box core team Related: Migration Center | Why Azure over AWS | Migration Playbook
