Services:Real-time Data Analytics
Real-time Data Analytics
Anomaly Detection for IoT
With the increasing usage of IoT in every industry, network security threats, malicious control, wrong setups are also becoming common in these domains. That is the reason why anomalies in IoT (connected car infrastructure, smart cities and smart homes, smart retail and customer 360, intelligent manufacturing) or fraud transaction detection, especially in industries like banking or insurance, require swift reactions.
​
Technaura has created a platform that can analyse data from multiple IoT services in real-time. It triggers alerts in case of detection of anomalies.
​
Some of the advantages are:
​
-
Lightweight pipeline
​
-
Build for poor connectivity
​
-
High latency
​
-
Support various clients’ connections (tens of thousands per MQTT server)
Technical Insight
-
This concept is about to create a pipeline using Kafka and Kafka connect MQTT in which we can integrate multiple IoT devices and the data pipeline can sink the respective data with other MQTT and DB sources. The data will be ingested into Imply (Druid) and analyzed with Machine Learning and in case of detection of any anomaly, it will trigger an alert.
​
-
All the flow is managed by Kafka connect MQTT connectors with zero lagging expectations. Imply is the heart of this concept with machine learning algorithms.
​
-
In case of an anomaly detection, an alerting stream will be triggered by Imply
​
-
Data from different IoT devices can also be stored in another microservice destination.
​
-
Furthermore, this is a complete pipeline without any code or application. This is lightweight, built for high latency scenarios, and supports many client connections (tens of thousands per MQTT server).
