Overview of IoT Ecosystems Components
Jan Vossaert [email protected] SMIOT Introduction IoT Devices
• A device includes hardware and software that directly interacts with the world o Devices connect to a network to communicate with each other, or to centralized applications o Devices might be directly or indirectly connected to the Internet
• IoT device development o Develop own hardware o Use commercial development platforms o Integrate dedicated commercial IoT device IoT Devices
• Device selection/design criteria Internet of things IoT Devices
Memory ~16KB Memory ~500MB
Memory ~2GB
Memory > 4GB Memory ~2GB Memory ~100KB Memory ~8KB Memory > 4GB IoT Devices
• Low-power communication stack • Standard IT communication stack • Cache only small amounts of data • Intelligence • Limited processing power
• Low resource software stack • IT SW stack • Network stack • WindRiver VxWorks • RIOT OS • Android Things • ARM Mbed OS • Green Hills Integrity • Contiki • Windows IoT Core • Nucleus RTOS Gateways
• Functionality o Protocol conversion data data o Move data to the cloud
o Access control to IoT functionality o Manage IoT devices in local network
o Intelligence on the edge: processes data on behalf of a cluster of devices • Pre-processing of data • Fast response times
o (End-user application platform) Gateways
Custom dedicated gateways Commercial gateways -Embedded platforms -Contain gateway software -COTS mobile platform with custom -Tied to specific application domains apps -Typically linked to back-end -Third-party software stacks platform IoT Platforms
• IoT Software-as-a-Service (SaaS) focused platforms
• Infrastructure/Platform-as-a-Service (IaaS/PaaS) back ends IoT Platforms: IoT Software-as-a-Service (SaaS) focused platforms
• Simplify the development of IoT ecosystems & applications o Device management o Data aggregation & storage o Can provide rule-based event engines o User access management
• Hide IoT complexity behind cloud APIs o All IoT devices in one platform o Multi-platform heterogeneity IoT Platforms: IoT Software-as-a-Service (SaaS) focused platforms
• Simplify the development of IoT ecosystems & applications o Device management o Data aggregation & storage o Can provide rule-based event engines o User access management
o Examples • Kaa (http://www.kaaproject.org) • Weave (https://developers.google.com/weave/) • Thingworx (http://www.thingworx.com) • AllThingsTalk (http://www.allthingstalk.com/) • Cariots (https://www.carriots.com/) • Waylay http://www.waylay.io) • Sierra Wireless’ AirVantage (https://airvantage.net) IoT Platforms: IoT Software-as-a-Service (SaaS) focused platforms
• Simplify the development of IoT ecosystems & applications o Device management o Data aggregation & storage o Can provide rule-based event engines o User access management
o Application-domain specific back-ends • Google Nest • FitBit IoT Platforms: Infrastructure/Platform-as-a-Service (IaaS/PaaS) back ends
• Focus on integration of business logic • Provide data mining libraries • Some platforms provide modules for IoT integrations
• Examples o IBM Bluemix (https://www.ibm.com/cloud-computing/bluemix/) o Microsoft Azure (https://azure.microsoft.com) IoT Platforms
• IoT platform selection criteria o Functionality • What analytics are supported • Device management • Should be as complex as, but not more complex than, the tasks it must perform
o Security and privacy • Multi-user/tenancy support • Platform security measures • Delivery models (cloud vs on-premises)
o Potential for scaling up o Pay for support (open source solutions) vs pay for platform IoT Platforms
• IoT platform selection criteria o Integration • Integration with existing applications • Device & communication protocols support
o IoT platform pitfalls • Data ownership & processing • Vendor lock-in IoT Ecosystems in Practice: MyDriving
• Enables you to record trips in your car o MyDriving mobile application and off the shelf OBD devices o Mobile device is used as gateway & application interface o Azure Services in the backend to analyze car telemetry data for both • Real-time insights like fuel consumption • Long term patterns and trends like driver rating IoT Ecosystems in Practice: Smart Cities
• Smart traffic sign platform in Antwerp o Uses smart city platform of Rombit o Digital traffic signs are equipped with Sigfox, NFC, GPS and E-ink screens
o Periodically upload location o Validity of traffic signs can be checked via online platform o Dynamic authenticated content IoT Ecosystems in Practice: Smart Cities Conclusions
• Cloud-managed solution o IoT devices directly communicate with cloud platform o Protocol conversions step via gateway
• Gateway-managed solution o Dedicated or mobile gateway interacts with IoT devices o Data processing and IoT device management on local gateway Conclusions
-Increased autonomy -Remote management -Local access to -Long-term data storage functionality -Flexible access control -Time critical functionality -Processing at the edge Conclusions