Published September 2, 2024 | Version v2
Dataset Open

Evaluating the Overhead of the Performance Profiler Cloudprofiler With MooBench

  • 1. ROR icon Kiel University
  • 2. Lancaster University Leipzig

Description

The dataset is the experimental results of the paper "Evaluating the Overhead of the Performance Profiler Cloudprofiler With MooBench". The results include Cloudprofiler and all target performance frameworks of MooBench in the current main branch. The experimental environment is an on-premise, bare metal server with two Intel Xeon E5-2650 CPUs and 64 GiB RAM on each NUMA domain. It runs the Debian 12.6 Linux on a single 480 GB SSD. The JVM-based performance benchmarks ran with OpenJDK 17.0.2, and Cloudprofiler's C++ library was built with GCC 12.2.0. The Kieker Python framework ran with Python 3.11.2.

MooBenchs target frameworks and benchmark scripts are maintained separately under "frameworks/<target benchmkark>" (the <target_benchmark> label is the placeholder for each performance frameworks), and the MooBench execution results are stored under "frameworks/<target benchmkark>/results-<target_benchmark>". This dataset is a single compressed archive, which contains all MooBench execution results from each target_benchmark:

results-Cloudprofiler-java/raw-${LOOP_NUM}-${REC_DEPTH}-${CONF_NUM}.csv
results-inspectIT-java/raw-*.csv
results-Kieker-java/raw-*.csv
results-Kieker-python/raw-*.csv
results-OpenTelemetry-java/raw-*.csv

The experimental data is separated as files by the current configuration number ${CONF_NUM}, the current loop number ${LOOP_NUM}, and the current recursion depth ${REC_DEPTH}.

The overall results are depicted as boxplots in the results.pdf file. The Y-axis is the execution time in microseconds, and the configurations on the X-axis are ordered by ${CONF_NUM} for each target framework.

Cloudprofiler is hosted at https://github.com/shinhyungyang/cloud_profiler, and MooBench is hosted at https://github.com/kieker-monitoring/moobench.

Files

results.pdf

Files (1.2 GB)

Name Size Download all
md5:907a31c3fc60d1793525fe22fa6815e9
15.0 kB Preview Download
md5:3a4132a5746d0dd0957bddbd18331e14
1.2 GB Download