For a long time, the Internet of Things was considered the vision of interconnecting every device, leading to fundamentally new and pervasive application scenarios. In practice, however, the projected growth and realization of IoT scenarios is often impeded by practical problems. The BASE MoVE research project, a cooperation between universities and industrial partners, took a holistic look at requirements and inhibitors for investing in IoT solutions, using Ambient Assisted Living as an application domain example. The perspectives of all stakeholders involved were taken into account during the design of a solution architecture, from the user to the manufacturer to the service provider and housing association. This paper presents the resulting modular base platform for IoT applications. Power supply through battery and energy harvesting enables low installation costs, since no expensive installation of cable ducts in existing buildings is required, which reduces the initial investment. The use of open source software and the support of several common smart home protocols also prevents a lock-in effect and dependency on a single manufacturer. This makes it possible to protect investments from market-driven changes from one manufacturer’s ecosystem to another. Over-the-air updates allow for secure operation as well as remote maintenance, no longer requiring expensive in-person maintenance. Finally, the manufacturing of the solution as a hardware module, as realized in BASE MoVE, also allows for easier creation and certification of new sensor devices in a company’s product portfolio. To evaluate the developed solution, an apartment was equipped with different sensor devices and a smart home scenario was implemented. The feasibility study could demonstrate that it is indeed possible to create a base platform that meets the requirements of the stakeholders involved. In addition to the scientific results, the project gives an assessment about components’ maturity and cost, which is valuable for the commercial project partner and its market entry strategy.

The following topics are dealt with: embedded systems; multiprocessing systems; real-time systems; scheduling; Internet of Things; cyber-physical systems; delays; processor scheduling; field programmable gate arrays; security of data.

Aktuellen Systemplattformen für Internet der Dinge-Geräte mangelt es oft an Isolationskonzepten. Verglichen mit traditionell isolierten Eingebetteten Systemen vergrößert die Anbindung an das Internet die Angriffsfläche, so dass durch Fehler die Funktionalität des gesamten Gerätes beeinträchtigt werden kann.

Abstract The current development trend of Internet of Things (IoT) aims for a tighter integration of mobile and stationary devices via various networks. This includes communication of vehicles to roadside infrastructure (V2I), as well as intelligent sensors / actors in Logistics and smart home environments. Compared to isolated traditional embedded systems, the exposure to open networks increases the attack surface, and errors in the networking components could compromise the safety and security of the embedded application or the whole network. But often current system architectures for mass-market IoT devices lack the required isolation concepts. Using a partitioning microkernel and enforcing the use of a microcontroller’s memory protection unit (MPU) facilities, we compare different isolation concepts for a publish/subscribe middleware implementing OMG’s Data Distribution Service (DDS) standard and we evaluate our results on an STM32F4 microcontroller. The results of this case study show moderate costs for increased memory usage and additional context switches.

The introduction of modern model-driven software development methodologies into the industrial practise still proves to be a challenge. Especially small or medium-sized enterprises (SMEs) need an incremental and continuous modernisation process, which incorporates existing projects, is customised and cost-effective. Particularly, suitable solutions for model-based or -driven testing with test automation to increase the efficiency are in demand. This paper presents an approach for integrating existing proprietary system models of an SME partner for describing industrial automation processes into a model-driven test process, utilising a domain-specific language for the test specification. The test objectives focuses on the correct implementation of the communication and synchronisation of distributed state machines. The presented approach is integrated into a test framework, which is based on the Eclipse Modelling Framework (EMF) and the Eclipse Test and Performance Tools Platform Project (TPTP) framework. To separate the possibly changeable system and DSL-specific models from the implementation of the test framework, a stable and more generic test meta model was defined.

Im Bereich der industriellen Automatisierung hat die Software-Qualität und damit das systematische Testen eine hohe Bedeutung. In dieser Arbeit wird die Erweiterung eines Testframeworks um die Validierung von Echtzeiteigenschaften für verteilte Zustandsmaschinen vorgestellt. Die Arbeit wurde zusammen mit einem industriellen mittelständischen Kooperationspartner im Bereich industrieller Ablaufsteuerungen durchgeführt. Der Aufbau der domänenspezifischen Testmodellierung und die Testauswertung der Echtzeiteigenschaften als Teil eines modellgetriebenen Testprozesses werden beschrieben, und als Konsequenz aus der Synchronisationsungenauigkeit der Uhren des verteilten zu testenden Systems wird eine Fehlerabschätzung gegeben, die bei der Testauswertung berücksichtigt wird.

In the last years many different protocols, middleware approaches and software stacks were developed for Wireless Sensor Networks (WSN) and the Internet of Things (IoT). Limited interoperability and support for heterogeneous target platforms are often problematic. A good match for this problem domain would be the OMG Data Distribution Service (DDS) standard, which specifies a data-centric publish/subscribe middleware with real-time capabilities and a rich set of Quality-of-Service polices. Although DDS was not intended for embedded systems with limited resources as found in WSN, the authors of this paper propose an approach for a customisable DDS implementation (sDDS) using an model-driven software-development process to tailor the middleware functionality on an individual node level. In this paper, the platform-abstraction layer of sDDS, the platforms supported so far and the effort needed for the porting process are presented. With this approach, it was possible to port sDDS to different platforms, like Contiki, RIOT-OS, FreeRTOS and self-contained ZigBee stack implementations.

The product- and technology-specific nature of embedded software components often results in a high overhead for the setup and automation of testing procedures. This paper presents the adaptation of the generic, modular Eclipse Test and Performance Tools Platform Project (TPTP) framework for managing and running automated tests and creating test reports for firmware releases of a computerized numerical control (CNC) device product. TPTP, built upon the development and user interface context of the Eclipse platform, is extended by product-specific plugins and TPTP module extensions. The actual test execution is performed remotely on CNC devices, controlled and evaluated by the testing framework. CNC firmware releases, configuration setups and the test results become manageable as TPTP resources. To facilitate modelling of test oracles and complex algorithms for CNC track evaluation, dynamic MATLAB code execution in the course of test procedures is supported.

Wireless Sensor Network technologies are maturing into industrial applicability, but the availability of broadly supported architectures and integration of existing standards is still lagging behind. As a proposal to ameliorate this situation, we consider the OMG Data Distribution Service (DDS), a middleware standard for interoperable data-centric publish/subscribe architectures with real-time capabilities, a candidate for standards-based realisations of equally data-centric WSN scenarios. DDS is rooted in larger-scale architectures. Consequently, the RTPS network protocol defined as a sibling standard to DDS assumes Ethernet-sized network frames and substantial communication resources, which are potential obstacles to applying DDS to WSNs. We therefore propose SNPS as an alternative transport protocol for DDS communication. SNPS has been designed as part of sDDS, a model-driven DDS realization architecture supporting highly resource-constrained embedded sensor node platforms. Still, SNPS is generally independent of sDDS; it was demonstrated to be usable even with minimalist solutions lacking a proper DDS run-time system. For SNPS, a modular layered architecture focussing on minimum footprint and scalability was defined. SNPS packets are assembled from an extensible set of submessages that are chosen and filled according to a well-defined, unambiguous protocol state engine. Keeping a stack of context information for a stream of SNPS submessages minimizes explicit state representation in the resulting packet structure. Other aspects of the protocol design are the support for bundling of data to minimize the total number of link layer frames exchanged and the leveraging of multi- and broadcast properties of wireless sensor networks. Most of the features are beneficial for wired sensor networks as well. SNPS has been implemented for several wireless and wired network protocols (ZigBee, 6LoWPAN, and Ethernet/UDP/IP) on diverse embedded sensor node and PC platforms. It has been embedded in the sDDS architecture and integrated into minimalist standalone implementations. It will be the main DDS transport protocol used at sensor integration levels of current projects of the authors' laboratory research group targeting home and industrial automation scenarios.

The development of embedded systems challenges software engineers with timely delivery of optimised code that is both safe and resource-aware. Within this context, we focus on distributed systems with small, specialised node hardware, specifically, wireless sensor network (WSN) systems. Model-driven software development (MDSD) promises to reduce errors and efforts needed for complex software projects by automated code generation from abstract software models. We present an approach for MDSD based on the data-centric OMG middleware standard DDS. In this paper, we argue that the combination of DDS features and MDSD can successfully be applied to WSN systems, and we present the design of an appropriate approach, describing an architecture, meta-models and the design workflow. Finally, we present a prototypical implementation of our approach using a WSN-enabled DDS implementation and a set of modelling and transformation tools from the Eclipse Modeling Framework.