Table 4 and 5

Scenario 1

Platform		Image
	Vanilla		opt. RT
	x_m →	µ ↑	x_m →	µ ↑
Bare-metal	16.7%	0.33	60.0%	3.30
Virtual Machine	25.0%	2.58	58.3%	4.00
Container	16.7%	1.50	66.7%	3.83

Scenario 2

Platform		Image
	Vanilla		opt. RT
	x_m →	µ ↑	x_m →	µ ↑
Bare-metal	100.0%	1.33	58.3%	0.83
Virtual Machine	66.7%	2.58	50.0%	1.25
Container	71.4%	2.29	83.3%	1.42

Reproducibility

CSV Timestamp Data

All sampled CSV Data for scenario1 are available under https://doi.org/10.14459/2023mp1718824.

All sampled CSV Data for scenario2 are available under https://doi.org/10.14459/2024mp1736868.

EVT Modeling and Evaluation

Create EVT Models


# Install dependencies
apt install -y zstd
pip install evt
pip install sympy

# Get EVT script
The script is located at https://github.com/wiednerf/container-in-low-latency/blob/main/scenario2/scripts/evt/evt-model.py

python3 evt-model.py --data <filename.csv.zst>

# Results are stored in <filename.csv.evt.json>
# Plots of the MLE fitting in <filename.csv.mle.pdf>