Edit

Share via

Azure Cloud Services migration decision matrix

This article provides a decision matrix to help evaluate migration options for an Azure solution currently built on Cloud Services. The matrix compares various Azure platform technologies to help guide your rearchitecting decisions.

Decision matrix

Criteria / Option Service Fabric Azure Functions Azure Web Apps Azure App Service (Container Apps, Logic Apps, etc.) Azure Kubernetes Service (AKS) Virtual Machine Scale Sets Azure Virtual Machines
Migration Complexity Moderate-to-high. Requires rearchitecting roles to stateful/stateless services. Often a learning curve for state management and orchestration. Low-to-moderate. Best suited for event-driven, lightweight tasks. Significant rework if the app isn't designed for serverless. Low-to-moderate. Often similar to current web role structure. Minimal rework required. Low-to-moderate. Migration complexity varies by service but well-supported with migration tools. High. Containerizing your app requires a deep understanding of microservices and orchestration. Moderate. Closer to Cloud Services model, offering easier lift-and-shift with autoscaling capabilities. High. Lift-and-shift may be simpler, but it doesn't leverage cloud-native benefits.
Scalability & Performance High. Designed for complex microservices and can scale stateful/stateless workloads effectively. High. Automatic scaling based on events, but can be subject to cold starts. Moderate-to-high. Built-in scaling features, but may need manual configuration for advanced scenarios. High. Automated scaling with various triggers (HTTP requests, event-based, schedule). Very high. Container orchestration allows fine-grained scaling and resilience. High. Automatic scaling based on metrics, with predictive autoscaling capabilities. Limited. Scalability depends on manual scaling and VM size/configuration.
Control & Customization High. Full control over service orchestration and state management. Limited. Managed service with constrained runtime environment and execution limits. Moderate. Managed platform with some configuration options; less control over underlying infrastructure. Moderate. More flexibility than Web Apps with container support, but still PaaS constraints. High. Maximum control over container orchestration and runtime environment. High. Control over VM configuration, networking, and load balancing, with autoscaling policies. Very high. Complete control over the OS, middleware, and runtime.
Operational Overhead Moderate. Requires management of cluster health, upgrades, and stateful services. Low. Managed service abstracts infrastructure, reducing operational overhead. Low. Microsoft manages the platform, though some app-specific issues remain. Low. Managed service with minimal infrastructure management required. High. Requires in-depth operational expertise (monitoring, updates, networking, etc.). Moderate. Handles instance scaling automatically but requires VM image maintenance and updates. High. Full responsibility for maintenance, patching, and security.
Ecosystem Integration Strong. Integrates well with Azure's service-oriented ecosystem, but may require further configuration. Strong. Natively integrated with Azure's Event Grid, logic apps, and other serverless services. Strong. First-class integration with CI/CD, monitoring, and other Azure services. Very strong. Designed to work together with other Azure services with minimal configuration. Strong. Works well with modern DevOps tools, though integration complexity increases with microservices. Strong. Integrates with Azure Load Balancer, Application Gateway, and Azure Monitor. Variable. While integration is possible, it often requires more custom development.
Use Case Suitability Ideal for applications needing fine-grained control over microservices, including stateful services. Best for event-driven workloads, batch processing, and where latency is acceptable. Well-suited for web apps and APIs with a standard request/response model. Good for mixed workloads combining web, mobile backends, integration logic, and API scenarios. Excellent for microservices and containerized workloads that require high availability and scalability. Well-suited for stateless applications with variable load patterns and IaaS requirements. Good for legacy applications that require minimal changes, but less cloud-optimized.

Additional considerations

  • Legacy vs. Re-architected:
    If you're looking to modernize your architecture beyond a lift-and-shift, then options like Service Fabric, AKS, or even a serverless approach with Functions might provide better long-term benefits despite higher initial rearchitecting costs.

  • Skill Set and Team Expertise:
    Evaluate your team's familiarity with microservices (Service Fabric, AKS) versus serverless or PaaS (Web Apps, Functions). Training and adoption curves could influence your decision.

  • Cost Implications:
    Managed services (Functions, Web Apps) may reduce operational costs compared to managing your own container clusters (AKS) or VMs, but the pricing models (consumption vs. reserved instances) can vary significantly.

  • Future-Proofing:
    Consider how each option aligns with long-term product strategy. For example, AKS and containerized architectures are gaining popularity for their portability and ease of integration with CI/CD pipelines.

  • Performance and Latency Requirements:
    Some workloads might benefit from the granular control offered by Service Fabric or AKS, while others might be well served by the ease-of-use and autoscaling features of Functions or Web Apps.

Next steps

Assess your workloads:
Identify which components of your solution are best suited for each architecture—consider both compute-bound and stateful components.

Prototype migration:
Consider building a proof-of-concept for the top two or three options. This can help uncover unforeseen challenges and better estimate migration complexity.

Cost-benefit analysis:
Include not only migration effort and operational overhead but also long-term benefits such as ease of updates, resilience, and scalability.

Stakeholder review:
Share the decision matrix with your team and other stakeholders for feedback and to align on priorities.

Review migration guides:

  • Last updated on