## Why Lift-and-Shift Isn't Enough
Moving your applications to the cloud on the same architecture they ran on-premises — "lift and shift" — gets you off your data centre hardware but doesn't give you most of the cloud's benefits. Your applications still run as they always did; you're just paying cloud prices for server-like resources instead of data-centre prices. In many cases, this is actually more expensive than your previous on-premises setup.
The full benefits of cloud — elastic scalability that matches infrastructure to actual demand, automatic resilience that handles hardware failures transparently, development velocity from managed services, and cost efficiency from paying only for what you use — require cloud-native architecture. Applications need to be designed (or redesigned) for the cloud operating model rather than just running on cloud infrastructure.
Replatforming is the middle path between lift-and-shift (minimal change) and complete rebuild (maximum change). It makes targeted changes to the application — containerising it, replacing embedded databases with managed services, restructuring for horizontal scaling — without fundamentally changing the application logic or user interface. Done well, replatforming delivers most of the cloud benefit at a fraction of the rebuild cost.
## The Replatforming Process
Assessment comes first. Before changing anything, understand what the application does, how it's used, what it depends on, and what its non-functional requirements are (performance, availability, RTO/RPO). This assessment surfaces the specific technical blockers to cloud-native operation: hardcoded server references, local file system dependencies, session affinity requirements, external dependencies that won't work in cloud networking.
Containerisation is usually the first replatforming step. Packaging the application in a Docker container doesn't change the application, but it does give you portability, consistent deployment, and a foundation for Kubernetes orchestration. Most applications containerise without code changes; the work is in the Dockerfile and build pipeline.
Managed service substitution replaces self-managed components with cloud managed equivalents: self-managed MySQL becomes Amazon RDS; self-managed Redis becomes ElastiCache; self-managed Kafka becomes Amazon MSK or Azure Event Hubs. The application continues to use the same protocols and APIs, but the operational overhead and availability characteristics improve dramatically.
Configuration externalisation — moving environment-specific configuration out of the application code and into cloud-native configuration management — is essential for cloud deployment. Applications that assume they're running on a specific server with specific local file paths need to be updated to read configuration from environment variables and secrets management services.
## Kubernetes Migration Strategy
Deploying containerised applications to Kubernetes rather than individual VMs provides automatic scaling, self-healing, rolling deployments, and efficient resource utilisation. For most organisations, using a managed Kubernetes service (EKS, AKS, GKE) is the right choice — the operational overhead of managing Kubernetes control planes is significant, and managed services handle it transparently.
The blue-green deployment pattern is the standard approach for zero-downtime migration to Kubernetes. You deploy the application to Kubernetes alongside the existing deployment, validate that it works correctly under load, and then switch traffic from the old deployment to the new one. If anything goes wrong, you switch back instantly. This eliminates the deployment risk that makes many replatforming projects stressful.
Service mesh adoption (Istio, Linkerd) is often a later step in the replatforming journey, adding observability, traffic management, and security capabilities at the infrastructure layer without changes to application code. Start without a service mesh and add it when the operational complexity is justified by the capabilities you need.
*Lara IT Solutions provides cloud-native migration services and replatforming programme delivery. Contact 0330 043 1930.*