Striim Founder and CTO, Steve Wilkes, shows how to collect and analyze large amounts of IoT data while ensuring scalability, reliability and optimal use of resources using Striim agents for edge processing and Striim servers for analytics.
This demo uses an Arduino unit to generate IoT data, a Dell IoT Gateway and Striim Agent to collect data, and a Striim server for processing and analytics.
Today we’re going to look at IoT edge processing and analytics.
You’ve probably heard that iot generates huge amounts of data. The question is how do you collect and analyze it all while making sure your system scales is reliable and doesn’t overload your existing resources including networks. We would recommend that you use Striim agents at the edge for data collection and initial processing and service to perform more complex analytics. In this demo we’re using an Arduino, you know to generate IoT data at Dell IoT gateway, running a stream agent to collect and initially processed the data and the stream server for further processing and running analytics. The arduino sends MQTT data to an MQTT broker running in the Dell gateway. It is collected and initially processed by the stream agent before being delivered to the stream server. A separate program is used to control the Arduino by it MQTT messages stream splits are processing over the agent and the server during initial pressing at the edge and analysis in the server. The results of this processing can then be shown you some visualizations in the dashboard. The device we are using is not, we know Uno with a light sensor and an Allie do it’s code send sensor data over MQTT and can be used to switch the led on and off and start sensing light levels. The Dell IoT gateway access and MQTT broker by utilizing the mosquito MQTT broker software. It also runs our stream agent which smartly connects to the selected a stream cluster. The Striim server is running on a macbook pro and will provide control of the whole end to end data flow.
A separate program is used to send MQTT messages to the Ardwino causing it to change modes, turn the led on and off and start sensing light levels with stream running as both a server and as an agent in the Iot gateway. When we deploy our data flow we can choose for part of it to be deployed at the edge. The agent then does data collection and edge processing within the gateway. The edge processing collects MQTT data and splits any light sensor data from command information, the server processing calculate, which can then be previewed in the data flow before we visualize it. And then this data is stored as well as any command information that we may have received. The dashboard visualizes all of this data in real time, including the statistics and Deming. The amount of light received by the light sensor or increasing it by utilizing the flashlight on a camera has an immediate realtime effect on the data visualizations showing the full end to end data flow. In future videos, we will integrate with more iot devices, including examples from healthcare, industrial, Iot, and manufacturing. We will also show how edge processing can be used to spot patterns in events and taking immediate action by talking back to them.