The current role of the Internet of Thing (IoT) is no more limited only to electronic identification of objects but it is perceived as a way to act, measure, or provide services based on real-world entities. Advancements in networking technologies and sensor/actuator capabilities provide a large number of physical world objects with communication and computation capabilities to interact with their surrounding environment.
However, the emerging IoT platforms are currently hard to deploy and operate, requiring except in the most trivial cases the intervention of domain experts to interpret the sensor data and, eventually, to come up with actuation commands.
This approach is clearly too expensive and time-consuming, and simply does not scale with the increasing number of IoT devices. In addition, the development of IoT application is highly scenario and technology dependent, due to the heterogeneity of devices. That is why powerful middleware solutions are required to integrate heterogeneous devices and dynamic services for building complex systems for the IoT.
We believe that the service-oriented approach provides adequate abstractions for application developers, and that it is a good approach to integrate heterogeneous sensors and different sensor network technologies with Cloud platforms through the Internet, by paving the way for new IoT applications.
The collection of diverse information produced by the proliferation of inter-connected sensing entities is one of the biggest challenge that should be adequately addressed. The huge amounts of raw data produced by IoT devices need to be structured, stored, analysed, correlated and mined in a reliable and scalable way.
The size of the produced data, and the high rate at which data are being produced, suggest that we need new solutions that combine tools for data management and services capable of promptly structuring, aggregating and mining data even just at the time they are produced.
In this paper we propose a OSGI-based middleware to be deployed on top of physical sensors and sensor networks, capable of abstracting sensor devices from their proprietary interfaces, and offering them to third party applications in an as-a-Service fashion for prompt and universal use. The middleware also offers tool to elaborate real-time measurements produced by sensors. A prototype of the middleware has been implemented. .
This middleware, called Sensor Node Plug-in System (SNPS), incorporates sensors which are no longer low-level devices producing raw measurement data, but can be seen as “services” to be used and composed over the Internet in a simple and standardized way in order to build even complex and sophisticated applications.
The primary aim of this SNPS is to bring any physical sensor/actuator on an abstraction level that allows for easier and standardized management tasks (switch on/off, sampling), in a way that is independent of the proprietary sensor’s specification. The SNPS middleware is then organized into several components, each of which is implemented as a software module (or “bundle”) within the OSGi framework.
By the time a sensor is “plugged” into the middleware, it will constitute a resource/service capable of interacting with other resources (be them other sensors plugged into the middleware or third party services) in order to compose high-value services to be accessed in SOA-fashion. The middleware also offers a set of complimentary services and tools to support the management of the entire life cycle of sensors and to sustain the overall QoS provided by them.
To read this external content in full, download the complete paper from the authors online archives.