[Hyperscale lock-in]
A practical response to hyperscaler lock-in
Leaving a large cloud provider is rarely about virtual machines. The hard parts are managed services, identity, data, habits, and team knowledge.
Find the decisions that became dependencies
We map provider services, network choices, data stores, monitoring, deploy paths, contracts, and the skills needed to run the system. This shows what you need and what you can replace.
Three viable responses
Moving a workload is not useful if it becomes less reliable or more costly. The usual choices are:
- Optimise: reduce waste and improve visibility where you are
- Diversify: place workloads across private, local, and cloud systems
- Exit: move workloads when control and stable cost justify the work