A systems architecture for an innovative solution
The project has an infrastructure for obtaining, storing, processing and integrating data. Intelligent telemetry systems have been implemented for capturing and processing information, with the following elements:
1 Data acquisition processes.
The electrical measurement data is obtained from the i-de website and is downloaded in CSV files manually for each user, while robotization methods are being tested to automate this process.
The water consumption measurement data is downloaded through a flow using the Apache NiFi tool, which connects periodically to the HIDRAQUA API and is then stored in the data space.
2. Infrastructure and hosting.
The solution is deployed on its servers; a CPD with “TIER III characteristics of the ANSI/TIA-942 standard for Data Centers” with an availability of 99.982%, preventing maintenance from affecting computing.
Its main features are:
Security. Backup with virtual machines in a VMware environment is complemented by a dedicated fibre storage network (SAN).
Smart City Platform. Developed according to UNE 178104, it allows horizontality, multi-entity, multi-service, and transversality. Interoperability. It is open, scalable and modular, and integrates data from different sources, sensors, applications, and social networks to offer decision-making support services.
In the knowledge phase, there are processes in charge of performing real-time analysis, batch processes for heavy calculations and machine learning processes that will serve as support for performing analysis of the platform's operation as well as generating events and notifications.
It has an object storage system compatible with the S3 interface. In this space, a "bucket" or cube has been created to deposit the data, in 3 types of folders. The "raw" folder stores the original data obtained from both sources. The “bronze” and “silver” folders store intermediate data obtained from the previous ones and which have been treated to carry out data cleaning and adaptation processes. Finally, the “gold” folder stores the final data obtained by the analytical and machine learning processes.
3. Alarm management.
With the data stored in the enabled data space, machine learning algorithms are implemented for this purpose. The generated alerts are stored in the data space and are consulted from the website that will be developed for this alarm management. The data generated on alarms can be consulted by third-party applications through a REST API.
