Dynamic and cloud-based systems are typically deployed on virtualized infrastructure, so their performance risk profile is a superset of the risks affecting virtualized systems (e.g., resource contention and overhead from the hypervisor layer) plus additional risks specific to dynamic provisioning, elasticity, auto-scaling, and multi-tenant resource sharing across unpredictable, externally managed cloud resources. This is why the performance-risk analysis for dynamic/cloud-based architectures must account for virtualization concerns on top of its own category-specific concerns, unlike multi-tier, purely distributed, or client-server architectures which are not inherently virtualization-based.
Community Discussion