Model-Driven IoT App Stores: Deploying Customizable Software Products to Heterogeneous Devices
summary
A model-driven app store maps OCL constraints on components to device object diagrams of a provided hardware ontology, generates facts and rules as Prolog rule engine code, and then proposes deployments using Docker/Kubernetes/Azure/AWS or the EU Genesis prototype. A smart‑home case study enforces a 4‑megapixel camera requirement, matches available Raspberry Pi cameras, produces alternative fixes when no match exists, and deploys handwritten drivers (C++/Python) to capable edge devices while excluding microcontrollers that cannot receive software.
AI-written summary · about
chapters (10)
- 0:00Model-driven IoT app stores
- 11:22Hardware vs device capabilities
- 13:53Application features and rules
- 16:58Validating sensor requirements
- 18:46Deploying to heterogeneous devices
- 19:37Smart home case study
- 20:10Handling rule mismatches
- 22:01Hardware platform requirements
- 24:08Integrating device properties
- 25:13Q&A
related talks
41mOne Model to Rule Them All
28mAutomotive Cloud Development Paradigms - Diving into Connecting IoT...
50mDeveloping (on) for the Edge
49mEnable IoT with Edge Computing and Machine Learning
25mImplementing IoT Architectures with Open Source Software
43mIndustrial IoT in practice
54mImplementing IoT Architectures with Open Source
31mBuild Your Intelligent Edge