ShareBench is a tool set for real-world performance analysis studies of distributed resource-sharing mechanisms and policies, developed as part of the Computer Science Bachelor Thesis on Performance Characterization of Distributed Resource-Sharing Mechanisms by Lennart K. M. Schulz.
The work consists of two components: ShareBench-Base and ShareBench-Gen. The former is an infrastructure framework for automated real-world performance analysis studies. The latter is a workload generator for OLAP workloads based on the TPC-DS data set and queries.
Documentation for each of the components can be found in the docs folder.
The only supported composition is Spark on Kubernetes. Support for other Application Frameworks or Resource Managers is not currently planned.
The work was initially based on the spark-data-generator by Sacheendra Talluri. Code annotations indicate what material is copied/adapted from that source.
To reduce the size of this repository, the data obtained through the experiments is not directly included but rather provided in sharebench-data.
When using ShareBench for research, please use the following BibTeX entry for citations:
@thesis{schulz2024Sharebench,
author = {Schulz, Lennart K. M.},
title = {{ShareBench}: Performance Characterization of Distributed Resource-Sharing Mechanisms},
institution = {{VU} Amsterdam},
type = {Bachelor Thesis},
date = {2024},
}