Choosing a TSDB for Smart City Public Facility Platforms

As urbanization accelerates, traditional municipal management models can no longer meet the demands of running a modern city efficiently. Smart city initiatives have become a major strategy for urban development worldwide, and public facilities are central to the digital transformation of city operations. Amid the flood of data generated by smart cities, time-series databases, with their strength in processing time-stamped data efficiently, have become an important technology for building public facility data platforms.

1. Core characteristics of smart city public facility data

1.1 Multi-source, heterogeneous data ingestion

Smart city public facilities draw data from an extraordinarily wide range of sources. In smart water management, for example, data may come from SCADA systems at water treatment plants, flow meters and pressure sensors across the pipe network, PLC controllers at pump stations, and online instruments at water quality monitoring stations. These devices use different communication protocols such as Modbus, OPC UA, MQTT, and LoRa, with varying data formats and sampling frequencies. A mid-sized city commonly has over 100,000 connected IoT terminals. Efficiently integrating this multi-source, heterogeneous data is one of the first considerations when selecting a time-series database.

1.2 Massive terminals and high-concurrency writes

Public facility monitoring is characterized by broad coverage and high density. Pressure monitoring points are deployed every few hundred meters along urban water supply networks, gas pipeline leak detection sensors are distributed throughout underground networks, and environmental monitoring stations are positioned across every part of the city. These massive numbers of terminals continuously generate high-frequency data, placing enormous pressure on the write concurrency of the data platform. A capable time-series database should handle high write throughput, potentially millions of data points per second in writes to support the stable operation of a city-scale public facility monitoring network.

1.3 Real-time and historical requirements coexist

Public facility management requires both real-time monitoring and alerting as well as long-term historical data for trend analysis and decision support. Pump station operating status demands second-level or even millisecond-level real-time response, while pipe network leakage analysis may need to trace back several years of historical data. A time-series database must deliver low-latency real-time queries while also providing efficient historical data compression, storage, and fast retrieval to ensure both real-time and historical capabilities.

2. Typical application scenarios

2.1 Smart water management: full-chain digital operations

Smart water management is one of the most mature public facility domains for time-series database adoption. From water source intake, treatment plant purification, and pipeline distribution to end-user consumption, every stage generates large volumes of time-series data.

At the treatment plant stage, parameters such as sedimentation tank turbidity, filter head loss, chemical dosing amounts, and effluent residual chlorine require continuous monitoring. In the pipeline network stage, flow and pressure data from DMA zones are core inputs for leakage control. At the pump station stage, pump operating current, vibration, and temperature data relate directly to equipment safety.

By building a unified data platform on a time-series database, organizations can achieve end-to-end data connectivity from source to tap to provide a high-quality data foundation for hydraulic network models and leakage early-warning algorithms.

2.2 Smart gas: safety monitoring and dispatch optimization

Gas safety is a matter of urban public security. Pressure, flow, and temperature monitoring at gas gate stations, operating status monitoring at pressure regulator boxes, and alarm data from pipeline leak detection sensors all need to be aggregated in real time at the data center. A time-series database can support gas utilities in building SCADA monitoring, pipeline network simulation, and gas consumption forecasting applications to shift from reactive emergency repair to proactive prevention.

2.3 Smart heating: precise regulation and energy savings

District heating systems in northern cities are a classic time-series data-intensive scenario. Boiler operating parameters at heat source plants, temperature and pressure readings from primary and secondary networks at heat exchange stations, and room temperature monitoring data from end users together form the data foundation for heating regulation. A smart heating platform built on a time-series database can achieve system-wide data integration across source, network, station, and user, supporting energy-saving strategies such as climate compensation and zoned scheduling by time and area to deliver significant consumption reductions while maintaining heating quality.

2.4 Environmental monitoring: multi-dimensional urban sensing

Urban environmental monitoring networks include air quality monitoring stations, noise monitoring points, and automatic water quality monitoring stations. These stations report monitoring data at regular intervals according to standard protocols, and the data has strictly time-series characteristics. A time-series database can effectively support environmental quality trend analysis, pollution source tracing, early warning, and forecasting applications to provide data-driven support for urban environmental management.

3. Core selection criteria for time-series databases

3.1 Multi-protocol ingestion capability

Public facility sites deploy a wide variety of equipment with diverse communication protocols. When evaluating options, examine whether the time-series database provides comprehensive data collection tools or interfaces, whether it supports direct connections for industrial protocols such as MQTT, Modbus, and OPC UA, and whether it can integrate seamlessly with existing SCADA systems and IoT platforms. A mature data ingestion ecosystem can significantly reduce system integration costs.

3.2 High-concurrency write and query performance

A city-scale public facility platform must support concurrent data writes from hundreds of thousands or even millions of measurement points. During evaluation, verify the database’s write stability and query response speed under high-concurrency scenarios through practical testing, focusing on key performance indicators such as data compression ratio, write throughput, and time-range query efficiency.

3.3 Data subscription and sharing mechanisms

Smart cities emphasize data sharing and cross-department collaboration. A time-series database should support flexible data subscription mechanisms that allow different business systems to obtain real-time or historical data on demand. For example, a pipeline GIS system may need to subscribe to pressure monitoring data for pipe burst location, while a billing system requires water consumption data for usage-based charging. Robust data sharing capability is key to breaking down information silos.

3.4 Multi-tier deployment and edge computing support

Smart city public facilities are inherently characterized by distributed collection and centralized management. A strong time-series database should support a cloud-edge-device collaborative multi-tier deployment architecture, allowing data preprocessing and local storage at edge nodes to ensure business continuity during network interruptions, and automatically synchronizing data to the central platform when the network recovers.

4. Data model design best practices

4.1 Hierarchical data organization

Smart city public facility data has a clear hierarchical structure: city, district, site, device, and measurement point. The data model should map to this physical hierarchy, using a multi-level organization of databases, Supertables, and Subtables. For example, databases can be created by business domain (such as water, gas, heat), with Supertables defined within each database by device type, and each specific device instance corresponding to a Subtable.

4.2 Supertable modeling strategy

The Supertable is a useful data modeling tool in time-series databases, particularly well-suited to public facility scenarios where many instances of the same device type exist. In smart water management, for example, a flow meter Supertable can be defined with common fields such as flow rate, pressure, and temperature, while each physically installed flow meter becomes a Subtable that automatically inherits the schema. This modeling approach ensures schema consistency while enabling efficient management of massive device data.

4.3 Tag system design

Well-designed tags are critical for data querying and analysis. It is recommended to configure each measurement point with complete tag information, including area code, site name, device type, installation location, and responsible department. These tags not only support multi-dimensional data filtering and aggregation queries but are also the key link for data association with GIS systems and asset management systems.

5. Integrated applications with GIS and BIM

5.1 Joint analysis of spatial and time-series data

By maintaining spatial coordinate tags for each measurement point in the time-series database, time-series data can be linked with GIS maps. In a pipe burst analysis scenario, the burst location can be identified via GIS while the time-series database is queried for data changes at surrounding pressure monitoring points before and after the incident to enable rapid impact assessment. Time-series databases like TDengine support efficient data filtering through tags, making such spatial-temporal joint analysis perform well in practice.

5.2 Data-driven BIM models

In scenarios such as water treatment plants and pump stations, BIM models provide detailed three-dimensional representations of facilities. By binding BIM components to their corresponding monitoring points, model-as-data visualization management becomes possible. Clicking a pump component in the BIM model displays its real-time operating parameters and historical trends, significantly improving operational efficiency.

6. City-scale deployment architecture

6.1 Three-tier architecture

For a city-scale smart public facility platform, a three-tier deployment architecture of edge node, district center, and city platform is recommended.

Edge node layer: deployed at water treatment plants, pump stations, and heat exchange stations, this layer handles local device protocol connections, data collection, and edge computing. Edge nodes run lightweight time-series database instances for local data caching and preprocessing to ensure no data loss during network interruptions.

District center layer: deployed by administrative division or business area, this layer aggregates data from edge nodes within its jurisdiction and handles district-level data storage, analysis, and business applications. District centers deploy standard time-series database clusters to support district monitoring, statistical analysis, and report generation.

City platform layer: this layer is the city-level data hub. It aggregates key data from all district centers and enables cross-district comprehensive analysis and decision support. The city platform deploys a High Availability (HA) time-series database cluster and interfaces with upper-layer applications such as the city brain and digital twin platforms.

6.2 Data synchronization and disaster recovery strategy

Data synchronization across the three tiers should follow the principle of on-demand upload and tiered storage. Edge nodes retain recent full-frequency high-resolution data, district centers retain medium-frequency data over longer periods, and the city platform primarily stores aggregated key indicator data. In parallel, data backup and disaster recovery mechanisms should be established at every tier to ensure the security and reliability of smart city public facility data.

7. Conclusion

As the core engine of a smart city public facility data platform, the choice of time-series database directly determines the quality of the technology foundation for the entire smart city initiative. Faced with challenges including multi-source heterogeneous data ingestion, massive high-concurrency workloads, and the coexistence of real-time and historical requirements, municipal enterprises and technical teams should evaluate time-series databases from the perspective of actual business scenarios, comprehensively assessing multi-protocol ingestion capability, performance, data sharing mechanisms, and deployment flexibility.

For data model design, teams should use hierarchical organization and Supertable modeling to build a standardized, extensible data system. The integrated application of time-series databases with spatial technologies such as GIS and BIM should also be actively explored to make better use of spatio-temporal data.

Through well-considered time-series database selection and architecture design, smart city public facility data platforms will provide solid data support for the modernization of urban governance.

If you are planning a smart city public facility data platform, start from your specific business scenarios, conduct POC testing and validation of time-series databases, and choose the technology solution that best fits your needs to lay a data foundation for the sustainable development of your smart city.