Time Series Management Systems: A Survey

📝 Abstract

The collection of time series data increases as more monitoring and automation are being deployed. These deployments range in scale from an Internet of things (IoT) device located in a household to enormous distributed Cyber-Physical Systems (CPSs) producing large volumes of data at high velocity. To store and analyze these vast amounts of data, specialized Time Series Management Systems (TSMSs) have been developed to overcome the limitations of general purpose Database Management Systems (DBMSs) for times series management. In this paper, we present a thorough analysis and classification of TSMSs developed through academic or industrial research and documented through publications. Our classification is organized into categories based on the architectures observed during our analysis. In addition, we provide an overview of each system with a focus on the motivational use case that drove the development of the system, the functionality for storage and querying of time series a system implements, the components the system is composed of, and the capabilities of each system with regard to Stream Processing and Approximate Query Processing (AQP). Last, we provide a summary of research directions proposed by other researchers in the field and present our vision for a next generation TSMS.

💡 Analysis

The collection of time series data increases as more monitoring and automation are being deployed. These deployments range in scale from an Internet of things (IoT) device located in a household to enormous distributed Cyber-Physical Systems (CPSs) producing large volumes of data at high velocity. To store and analyze these vast amounts of data, specialized Time Series Management Systems (TSMSs) have been developed to overcome the limitations of general purpose Database Management Systems (DBMSs) for times series management. In this paper, we present a thorough analysis and classification of TSMSs developed through academic or industrial research and documented through publications. Our classification is organized into categories based on the architectures observed during our analysis. In addition, we provide an overview of each system with a focus on the motivational use case that drove the development of the system, the functionality for storage and querying of time series a system implements, the components the system is composed of, and the capabilities of each system with regard to Stream Processing and Approximate Query Processing (AQP). Last, we provide a summary of research directions proposed by other researchers in the field and present our vision for a next generation TSMS.

📄 Content

1 Time Series Management Systems: A Survey Søren Kejser Jensen, Torben Bach Pedersen, Senior Member, IEEE, Christian Thomsen Abstract—The collection of time series data increases as more monitoring and automation are being deployed. These deployments range in scale from an Internet of things (IoT) device located in a household to enormous distributed Cyber-Physical Systems (CPSs) producing large volumes of data at high velocity. To store and analyze these vast amounts of data, specialized Time Series Management Systems (TSMSs) have been developed to overcome the limitations of general purpose Database Management Systems (DBMSs) for times series management. In this paper, we present a thorough analysis and classification of TSMSs developed through academic or industrial research and documented through publications. Our classification is organized into categories based on the architectures observed during our analysis. In addition, we provide an overview of each system with a focus on the motivational use case that drove the development of the system, the functionality for storage and querying of time series a system implements, the components the system is composed of, and the capabilities of each system with regard to Stream Processing and Approximate Query Processing (AQP). Last, we provide a summary of research directions proposed by other researchers in the field and present our vision for a next generation TSMS. Index Terms—Approximation, Cyber-physical systems, Data abstraction, Data compaction and compression, Data storage representations, Data structures, Database architectures, Distributed databases, Distributed systems, Internet of things, Scientific databases, Sensor data, Sensor networks, Stream processing, Time series analysis ! 1 INTRODUCTION T HE increase in deployment of sensors for monitoring large industrial systems and the ability to analyze the collected data efficiently provide the means for automation and remote management to be utilized at an unprecedented scale [1]. For example, the sensors on a Boeing 787 produce upwards of half a terabyte of data per flight [2]. While the use of sensor networks can range from an individual smart light bulb to hundreds of wind turbines distributed throughout a large area, the readings from any sensor network can be represented as a sequence of values over time, more precisely as a time series. Time series are finite or unbounded sequences of data points in increasing order by time. Data series generalize the concept of time series by removing the requirement that the ordering is based on time. As time series can be used to represent readings from sensors in general, the development of methods and systems for efficient transfer, storage, and analysis of time series is a necessity to enable the continued increase in the scale of sensor network and their deployment in additional domains [1], [3], [4], [5]. For a general introduction to storage and analysis of time series see [6], [7], a more in-depth introduction to sensor data management, data mining, and stream processing is provided by [8], [9], [10]. While general Database Management Systems (DBMSs), and in particular Relational Database Management Systems (RDBMSs), have been successfully deployed in many situa- tions, they are unsuitable to handle the velocity and volume of the time series produced by the large scale sensor networks deployed today [3], [4], [5], [11]. In addition, analysis of the collected time series often requires exporting the data to another application such as R or SPSS, as these provide • S. K. Jensen, T. B. Pedersen, and C. Thomsen, are with the Department of Computer Science at Aalborg University, Denmark. E-mail: {skj, tbp, chr}@cs.aau.dk. additional capabilities and a simpler interface for time series analysis compared to an RDBMS, adding complexity to the analysis pipeline [12]. In correspondence with the increasing need for systems that efficiently store and analyze time series, Time Series Management Systems (TSMSs)1 have been proposed for multiple domains including monitoring of industrial machinery, analysis of time series collected from scientific experiments, embedded storage for Internet of things (IoT) devices, and more. For this paper we define a TSMS as any system developed or extended for storing and querying data in the form of time series. Research into TSMSs is not a recent phenomenon and the problems using RDBMSs for time series have been demonstrated in the past. In the 1990s Seshadri et al. developed the system SEQ and the SQL-like query language SEQUIN [13]. SEQ was built specifically to manage sequential data using a data model [14] and a query optimizer that utilize that the data is stored as a sequence and not a set of tuples [15]. SEQ was implemented as an extension to the object-relational DBMS PREDATOR with the resulting system supporting storage and querying of relational and sequential data together. While additional support for sequences was added to the S

This content is AI-processed based on ArXiv data.