Consumption Metrics and Visualization in DECTRIS Cloud
Efficient use of computing and storage resources is essential in modern experimental workflows. In DECTRIS Cloud, resource usage is tracked and visualized through clear, intuitive dashboards. This ensures scientists and lab managers can monitor consumption in real time, optimize workloads, and plan ahead for future needs.
- CPU hours (CPUh)
- GPU hours (GPUh)
- Elastic Storage usage (Hot & Archived)
CPUh – CPU Hours
- What it is: CPUh indicates the total hours of computing performed by CPU cores.
How it works in DECTRIS Cloud:- Each experiment’s CPU usage is measured and deducted from the available quota.
- For example, using 10 cores for 2 hours consumes 20 CPUh.
- The dashboard displays remaining CPUh alongside already consumed CPUh, as shown in the image.
- Why it matters: Helps track workloads and manage costs tied to CPU-intensive experiments.
GPUh – GPU Hours
- What it is: GPUh measures GPU computing usage in hours, a critical metric for image analysis, AI/ML tasks or accelerated data analysis.
- How it works in DECTRIS Cloud:
- GPU usage is tracked in the same way as CPUh, displayed as remaining vs. used GPUh.
- This ensures allocation of limited, high-value GPU resources.
- Why it matters: Enables researchers to balance workloads between CPU and GPU.
Elastic Storage
DECTRIS Cloud provides elastic, scalable storage with two tiers:
- Hot Storage: Fast-access storage for active experiments.
- Archived Storage: Long-term storage for completed or infrequently accessed data.
How it works in DECTRIS Cloud:
- The dashboard shows total available capacity (e.g., 100 GiB Hot Storage or 11 TiB Archived Storage).
- Usage is displayed against quota, allowing researchers to plan data movement between hot and archived tiers.
Why it matters:
- Prevents storage bottlenecks during experiments.
- Supports cost optimisation by archiving older data instead of keeping everything in expensive hot storage.