Design Architecture Blueprint for IOT and AI based Smart-Cities
Smart City is all about interactive and futuristic city which leverages IOT and connectivity to enable scenarios that were not possible in the past, both new and existing cities can benefit from emerging use cases of smart cities. This blog post focuses rather on actually how at platform level smart city can be implemented and what technologies are needed and where exactly to place them to achieve the overall smart city.
At the aggregation layer, before data goes into integration layer, we often times need to work on data which is aggregated, like hourly sum or sum of whole day of values etc. AWS Lambdas or Azure Functions both can be leveraged to perform aggregation.
Integration layer basic goal is just to connect multiple systems. We cannot use Node-Red here, but we required a performant system here which can scale both horizontally and vertically. So here we use Apache Camel Karavan (open source GUI based integration platform: see footnotes for more info on it). Kubernetes is base platform to host Karavan solution for cloud scale elasticity. We have connectors for each technology to allow development of all types of flows between applications.
Besides Integration layer is AI layer which basically takes data from Integration layer. You might need to perform some traffic prediction per hour for example so data would be flowing from aggregation layer all the way to integration and then to ML component of AI layer to generate and tune the model for traffic per hour as example.
Lastly the Apps for smart cities which implement the use cases leverages information from integration layer as well as AI layer to perform their tasks.
Apache Camel Karavan Technical Review: SRK-TechBlog | Review of Apache Camel Karavan as GUI Based Integration Platform (emetalsoft.com)