A Time-Series Database Analysis Based on a Multi-attribute Maturity Model

摘要

Time Series Database (TSDB), is a particular type of data repository. TSDBs are capable of diverse functionalities regarding operations on time series data and are developed using different technologies. A large number of TSDB solutions, available both as open source and commercial software, has emerged in the last years. Selecting the proper TSDB is a challenging endeavour for a potential enterprise client. In this paper, a set of open-source TSDBs which includes InfluxDB, Graphite, RRDTool, Prometheus, OpenTSDB and TimescaleDB - has been selected for analysis, evaluation and ranking. A comparison of TSDBs implies the establishment of a set of attributes. A set of quantitative and qualitative attributes which have different scales and units of measure has been selected. The problem of selecting or ranking TSDBs evaluated in this paper with a set of attributes is a Multi-Attribute Decision Making problem (MADM). For solving the TSDBs evaluation, analysis and selection a multi-attribute TSDBs maturity model has been proposed. The model for the selected TSDBs set and for the 18 attributes, 10 quantitative and 8 qualitative, has been validated. The model for all the attributes, both quantitative and qualitative, has been solved and, finally, a comparison between the obtained rankings has been made. The comparison of the ranks obtained shows that the leading TSDB is InfluxDB.