Time Series Management Systems: A 2022 Survey
Søren Kejser Jensen, Torben Bach Pedersen, Christian Thomsen · VBN Forskningsportal (Aalborg Universitet) · 2022
Enormous amounts of time series are being collected in many different domains. These include, but are not limited to, aviation, computing, energy, finance, logistics, and medicine. However, general-purpose Database Management Systems (DBMSs) are not optimized for times series management and thus significantly limit the amount of time series that can be efficiently stored and analyzed. As a remedy, specialized Time Series Management Systems (TSMSs) have been developed. This chapter, provides a thorough survey and classification of TSMSs that are developed through academic or industrial research and documented through peer-reviewed papers. To document their design and novel contributions, a summary of each system is provided. The systems are primarily classified based on their architecture. In addition, the systems are classified based on: when and why each system was developed, how it can be deployed, how mature its implementation is, how scalable it is, how it processes time series, what interfaces it provides, the type of approximation it supports, how low latency it can achieve, how it stores time series, and the types of queries it supports. The chapter concludes with a collection of open research problems based on the limitations of the surveyed systems.