Published on Fri Oct 29 2021
A Demonstration of Benchmarking Time Series Management Systems in the Cloud
Time Series Management Systems (TSMS) are Database Management Systems that
have been configured with the primary objective of processing and storing time
series data. With the IoT expanding at exponential rates and there becoming
increasingly more time series data to process and analyze, several TSMS have
been proposed and are used in practice. Each system has its own architecture
and storage mechanisms and factors such as the dimensionality of the dataset or
the nature of the operators a user wishes to execute can cause differences in
system performance. This makes it highly challenging for practitioners to
determine the most optimal TSMS for their use case. To remedy this several TSMS
benchmarks have been proposed, yet these benchmarks focus primary on simple and
supported operators, largely disregarding the advanced analytical operators
(ie. Normalization, Clustering, etc) that constitute a large part of the use
cases in practice. In this demo, we introduce a new benchmark that enables
users to evaluate the performance of four prominent TSMS (TimescaleDB, MonetDB,
ExtremeDB, Kairos-H2) in their handling of over 13 advanced analytical
operators. In a simple and interactive manner, users can specify the TSMS(s) to
compare, the advanced analytical operator(s) to execute, and the dataset(s) to
utilize for the comparison. Users can choose from over eight real-world
datasets with varying dimensions or upload their own dataset. The tool then
provides a report and recommendation of the most optimal TSMS for the
parameters chosen.