Abstract
Model-Driven Design (MDD) has proven to be a powerful technology to address the development of increasingly complex embedded systems. Beyond complexity itself, challenges come from the need to deal with parallelism and heterogeneity. System design must target different execution platforms with different OSs and HW resources, even bare-metal, support local and distributed systems, and integrate on top of these heterogeneous platforms multiple functional component coming from different sources (developed from scratch, legacy code and third-party code), with different behaviors operating under different models of computation and communication. Additionally, system optimization to improve performance, power consumption, cost, etc. requires analyzing huge lists of possible design solutions. Addressing these challenges require flexible design technologies able to support from a single-source model its architectural mapping to different computing resources, of different kind and in different platforms. Traditional MDD methods and tools typically rely on fixed elements, which makes difficult their integration under this variability. For example, it is unlikely to integrate in the same system legacy code with a third-party component. Usually some re-coding is required to enable such interconnection. This paper proposes a UML/MARTE system modeling methodology able to address the challenges mentioned above by improving flexibility and scalability. This approach is illustrated and demonstrated on a flight management system. The model is flexible enough to be adapted to different architectural solutions with a minimal effort by changing its underlying Model of Computation and Communication (MoCC). Being completely platform independent, from the same model it is possible to explore various solutions on different execution platforms.
Highlights
Embedded Systems (ESs) are the fundamental constituents of the Internet of Things (IoT), the new paradigm leading the development of electronics in the medium-term
This paper proposes the use of Unified Modeling Language (UML)/MARTE (Modeling and Analysis of Real-Time and Embedded systems), the standard proposed by the OMG for the modeling and analysis of realtime and embedded systems
To be reused, a component may require being inte grated in a completely different structure compared with the design where it was originally developed (Fig. 2), services can be requested with different communication semantics, or the functionality has to be run in a completely different hardware platform
Summary
Embedded Systems (ESs) are the fundamental constituents of the Internet of Things (IoT), the new paradigm leading the development of electronics in the medium-term. Increasing number of computing nodes in the execution platform as, in order to be able to provide the required services, the application has to be distributed over a network of nodes of different kind, from simples sensing motes and embedded systems to computing re sources in the fog and the cloud. The challenges posed by the increasing complexity of CPSoSs require of more powerful, generic, reusable, platformindependent modeling methods able to cope with the complete service under design. The use of these DLSs usually creates closed areas, since it is quite difficult to integrate under one DSL the developments done in others, limiting reusability and composability To solve this problem, this paper pro poses, first, the use of a standard modeling language, and to specify the system components in such a way that, they can be adapted for its integration under different environments, communication mecha nism and/or models of computation. The modeling methodology is assessed on a complex use case, a Flight Management System (FMS)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
