Output

During a simulation, OpenDC regularly takes a sample of the state of the data center and writes it to Parquet files. The export model decides which files are written, which columns they have, and how often a sample is taken.

Where the results are written

The results of an experiment are written to outputFolder/name/, where outputFolder and name are set in the experiment. Each run of a scenario gets its own folder:

output/
└── example/                  the name of the experiment
    ├── trackr.json           the settings of every scenario
    ├── simulation-analysis/
    └── raw-output/
        ├── 0/                scenario 0
        │   ├── seed=0/       the run with seed 0
        │   │   ├── host.parquet
        │   │   ├── task.parquet
        │   │   ├── service.parquet
        │   │   ├── powerSource.parquet
        │   │   └── battery.parquet
        │   └── seed=1/
        └── 1/                scenario 1

simulation-analysis/ is filled by M3SA, when it is used.

Output files

File	One row per
host.parquet	host, per sample
task.parquet	task, per sample
service.parquet	sample
powerSource.parquet	power source, per sample
battery.parquet	battery, per sample

Time

Every file has two times: timestamp is in milliseconds since the start of the simulation, and timestamp_absolute is in milliseconds since the Unix epoch.

Energy use, carbon emissions, uptime, downtime and CPU and GPU times are measured over the time since the previous sample. To get the total, add up the values of all samples.

Example

Reading the host file of a run with Python:

import pandas as pd

host = pd.read_parquet("output/example/raw-output/0/seed=0/host.parquet")
host["time"] = pd.to_datetime(host["timestamp_absolute"], unit="ms")
total_energy = host.groupby("host_name")["energy_usage"].sum()