Introduction

Context

We were asked to monitor Deustotech’s power consumption. We needed to gather the energy usage of each workplace, as well as conference rooms and printers power waste.

These were all the requirements:

• Monitoring of work place sockets, meeting rooms and printers (meeting rooms and offices we treat them the same).

• Data storage in Influx.

• Specification of the monitored devices, using well-defined identifiers and their locations in Deustotech.

• System and code documentation in Read The Docs.

• RESTful API implementation and specification for data queries.

• RESTful API implementation and specification for device control (turn off, on).

• Specification and documentation of Hooks 1 to facilitate integration with different devices and middlewares.

• Implementation of error alerts.

• Backup mechanism implementation.

1

Hook: techniques used to alter or increase the behavior of an operating system, applications or other software components by intercepting function calls or messages or events passed between software components

Proposed solution

As a result we created an IoT system where each workplace, group room and printer is plugged to an SmartPlug.

The gathered data is stored in a time series database: influxDB.

Also a RESTful API is exposed through which the data can be retrieved, and SmartPlugs are controlled: turned ON/OFF.

System’s overview

The system architecture is shown in the following diagram:

• API: Abstracts data queries and control requests into a REST API web server. The API specifications are presented in API Usage. Behind the scenes it is simple its implementation: it receives a REST request, it formats the message, and it publishes to MQTT, where a Component executes the query. Finally, it waits for the Component response, which is published in another MQTT topic.

• Component: It is the application that subscribes to a specific MQTT topic, where receives data queries. It responsible of executing the queries; usually this means to look up in a Database. Once the query is executed, it publishes the result to a MQTT topic where the API is listening.

• Hook: It is responsible of reading from an MQTT topic the data generated by the devices and taking care of it. For example, in case of the SmartPlugs, it stores them into InfluxDB database.

• Agent: It is the middleware that deals with the devices directly. Agents can be classified into two groups: controllers and collectors. The controllers send commands to the devices, such as “turn ON/OFF”. And the collectors are responsible of retrieving the data from the devices. Both of them use MQTT as the communication techonolgy with other components or applications.

Request flow: Influx query

1. Client makes HTTP request to api and api publishes influx query to MQTT

2. influx_query reads the command from MQTT

3. influx_query sends query to InfluxDB

4. InfluxDB executes query and returns the result to influx_query

5. influx_query publishes the result to MQTT

6. api reads query result from MQTT and returns an HTTP response

Request flow: SmartPlug command

1. Client makes HTTP request to api and api publishes control command to MQTT

2. sp_command reads the command from MQTT

3. sp_command sends command to the SmartPlug device(s)

4. SmartPlug executes the command and sp_command reads the resonse

5. sp_command publishes the response to MQTT

6. api reads control command response from MQTT and returns an HTTP response