Whether one is dealing with a manufacturing plant or a smart city, the architecture of a distributed Internet of Things (IoT) system depends on several design considerations, says Ingo Freise, software architect at Deutsche Telekom.
One is the ability to combine legacy sub-systems, such as a final-assembly cell, with greenfield deployments such as an image processing system for quality control. Another design consideration relates to the challenge of combining multiple connectivity and data management technologies, with the added complication that these are sourced from multiple vendors.
Deutsche Telekom dealt with such issues when working on the MySMARTLife project in a consortium with the cities of Nantes, Hamburg, and Helsinki. Naturally, our involvement in the project focused on the city of Hamburg.
The city’s platform, based on a standard developed by the Open Geospatial Consortium (OGC), is a common starting point for many cities. It supports the geo-location, surface planning and activities that municipal agencies deal with on a day-to-day basis.
An open standards and interoperable framework
The city of Hamburg operates under a transparency law that requires all data from public sources to be published. This ranges from static data, such as school opening times, to live data streams. An example of the latter is the status of electric vehicle charging stations.
Our solution addressed two issues. One was the issue of sourcing and publishing city data, going beyond legacy, geo-spatial sources. The second issue was to make it easy for data users to access data through an API (application program interface). While the existing Hamburg system included an API, our analysis identified the need for one with enhanced capabilities.
New features would enable a system administrator to assign access policies to end-point devices and data sources. An enhanced API would empower data providers to manage access and security at a granular level with different types of data consumer. These requirements illustrate how a basic API can be enhanced to address operational and digital transformation needs.
Foundations for digital transformation
Since our team focuses on applied innovation, we did not want to invent a new technology or IoT platform. We evaluated FiWARE and oneM2M specifications, eventually preferring oneM2M which had the added benefit of aligning with our Java expertise and Scrum organising methodology.
In addition to being an open and internationally recognised standard, oneM2M is a middleware technology. It sits in a horizontal layer between IoT applications on the upper level and a lower level of connected devices and other data sources. oneM2M defines a standard set of tools for building interoperable IoT systems.
Examples include tools for connectivity management, subscription management, security, and device management. The last of these is a good example where oneM2M avoids reinvention and reuses existing technology in the form of the LWM2M standard.
The Deutsche Telekom team worked with the Fraunhofer Institute of Optronics, System Technologies and Image Exploitation (IOSB) who built a server to collect city data from parking lots and traffic lights. My team developed a oneM2M data management platform to gather data from other sources and to publish it to third party users. We built a bridge between the two systems using an interworking proxy entity (IPE), another standardised capability in the oneM2M toolkit.
This architecture gathers city data from many different sources and to combine with the primarily geospatial data from Hamburg’s in-house data platform.
This is an elegant way to bring together several brownfield systems and gradually add greenfield sensor deployment and data sources. It is also valuable for data users who now have a single way to access a wide and growing range of city data.
As factory owners, city administrations and their service partners embark on Industrie 4.0, smart city, and digital transformation initiatives, system architects need to plan for interoperable and extensible systems using open standards, such as oneM2M.
An important design principle for service providers, system integrators and IT departments is to enable seamless interworking between new and legacy systems to work across departmental and operational boundaries.
The author is Ingo Freise, software architect, Deutsche Telekom.