Feature Description
When enabled, WaveRider allows OpCenter to optimize cloud resources by moving jobs as the resource demands change.
Resource utilization (CPU, memory and storage) is not constant while a job runs. Users face a difficult choice: over-provision and overpay, or under-provision and risk out-of-memory conditions. With WaveRider, users can start with a small virtual machine (VM) and dynamically migrate to bigger VMs (and back down to smaller VMs) as the application demands change during a complete job run.
Job migration is initiated in three ways:
- Manual: migration initiated using CLI commands or the OpCenter web interface.
- Policy-driven: migration triggered by rules-based policy configured per job (policy based on CPU and memory utilization thresholds).
- Programmatic: migration executed at specific program breakpoints defined by job script.