This repository contains the time series of two projects (P4 and P5). These were used to evaluate different time series databases. The evaluation is described in the paper (https://www.usenix.org/system/files/conference/fast17/fast17-lautenschlager.pdf). In contrast to the time series used in the paper, these time series are made anonymous (names of the time series).
The file names contain information separated by "_": <host>_<process>_<logical-group>_<name>_<date>
A pair of a time series consists of a timestamp and a numeric value.
Example: 12.08.2015 17:25:40.964;213,385,216 (note that they are ;-separted)
All files are zipped to reduce space (*.csv.gz).
P4 contains 498 time series with 3.8 mio pairs of time stamp and numeric value (data of a few hours measurement). The traces (strace) and the lsof-files are not included.
P5 contains 24,055 time series with 3.8 billion pairs of time stamp and numeric value (data of up to three months).