Vessim

Vessim is a co-simulation testbed for carbon-aware systems that allows you to simulate how your computing systems interact with local renewable energy sources, battery storage, and the public grid. It connects domain-specific simulators for power generation and energy storage with real software and hardware.

What can I do with Vessim?

Vessim helps you to understand and optimize how your (distributed) computing system interacts with (distributed) renewable energy sources and battery storage.

• Carbon-aware applications: Develop applications that automatically reduce their energy consumption when the grid is powered by fossil fuels, and increase activity when renewable energy is abundant.
• Energy system composition: Experiment with adding solar panels, wind turbines, or batteries to see how they would affect your energy costs and carbon emissions.
• Plan for outages and extreme events: Simulate power outages or renewable energy fluctuations to understand risks and test backup strategies.
• Quality assurance: Apply Vessim in continuous integrating testing to validate software roll-outs in a controlled environment.

Vessim can run simulations faster than real-time, includes historical datasets for realistic scenarios, and can simulate multiple microgrids in parallel. You can test scenarios using historical data or connect real applications and hardware to simulated energy systems.

Check out the official documentation!

Simple scenario

The scenario below simulates a microgrid consisting of a simulated computing system (which consistently draws 700W), a single producer (a solar power plant who's production is modelled based on a dataset provided by Solcast), and a battery. The Monitor periodically stores the energy system state in a CSV file.

```python import vessim as vs

environment = vs.Environment(simstart="2022-06-15", stepsize=300) # 5 minute step size

microgrid = environment.addmicrogrid( name="datacenter", actors=[ vs.Actor(name="server", signal=vs.StaticSignal(value=-700)), # negative = consumes power vs.Actor(name="solarpanel", signal=vs.Trace.load("solcast2022_global", column="Berlin", params={"scale": 5000})), # 5kW maximum ], storage=vs.SimpleBattery(capacity=100), )

Write results to CSV

monitor = vs.Monitor([microgrid], outfile="./results.csv") environment.add_controller(monitor)

environment.run(until=24 * 3600) # 24h simulated time ```

Software-in-the-loop scenario

Vessim can simulate the energy systems of real applications and hardware. The following example

1. pulls energy consumption data from a Prometheus monitoring system
2. exposes the simulated microgrid via a REST API, including the current marginal carbon intensity of the public grid (provided by Watttime).

```python import vessim as vs

environment = vs.Environment(simstart="2022-06-15", stepsize=5) # 5 second step size

microgrid = environment.addmicrogrid( name=f"gpuclusterinberlin", actors=[ vs.Actor(name="gpucluster", signal=vs.PrometheusSignal( prometheusurl="http://localhost:30826/prometheus", query=f"sum(DCGMFIDEVPOWERUSAGE)", # sum of all GPUs' power consumption username="username", password="xxxxxxxxxx" )) ], gridsignals={"mciindex": vs.WatttimeSignal( username="username", password="xxxxxxxxxx", location=(52.5200, 13.4050), # Berlin coordinates )}, )

restapi = vs.Api([microgrid], exportprometheus=True) environment.addcontroller(restapi)

environment.run(until=24 * 3600, rt_factor=1) # 24h in real-time mode ```

Installation

You can install the latest release of Vessim via pip:

pip install vessim

If you require software-in-the-loop (SiL) capabilities, you should additionally install the sil extension:

pip install vessim[sil]

Work in progress

Our team at the Distributed and Operating Systems group at TU Berlin is actively working to improve Vessim. We are currently working on the following aspects and features:

• Vessim X Exalsius: We are working on integrating Vessim into Exalsius' GPU provisioning and scheduling.
• Vessim X Flower: We are working on integrating Vessim into the federated learning framework Flower.
• Vessim X Vidur: We are working on integrating Vessim into the LLM simulator Vidur.
• System Advisor Model (SAM): We are working on integrating NREL's SAM as a subsystem in Vessim, allowing for better simulation of solar arrays, wind farms, and other types of renewable energy generators.
• Battery degradation: We are working on integrating NREL's BLAST-Lite for modeling battery lifetime and degradation
• Calibration: We are working on a methodology for calibrating Vessim simulations on real hardware testbeds.

Datasets

Vessim comes with ready-to-user datasets for solar irradiance and average carbon intensity provided by

and

We're working on documentation on how to include custom datasets for your simulations.

Publications

If you use Vessim in your research, please cite our paper:

• Philipp Wiesner, Ilja Behnke, Paul Kilian, Marvin Steinke, and Odej Kao. "Vessim: A Testbed for Carbon-Aware Applications and Systems." 3rd Workshop on Sustainable Computer Systems (HotCarbon). 2024.

For details in Vessim's software-in-the-loop simulation methodology, refer to our journal paper:

• Philipp Wiesner, Marvin Steinke, Henrik Nickel, Yazan Kitana, and Odej Kao. "Software-in-the-Loop Simulation for Developing and Testing Carbon-Aware Applications." Software: Practice and Experience, 53 (12). 2023.

For BibTeX citations and more related papers, please refer to the documentation.