Early exploring design alternatives of smart sensor software with Model of Computation implemented with actors

International audiencenetwork of sensors that aims to study the oceans.Such networks present two drawbacks: the large amount of acquired data prevents manual processing, and a lot of irrelevant data degrades the treatment efficiency. To smartly reject irrelevant data, filtering must happen at the source. This requires the design of complex sensors that support concurrent operations to mix acquisition, treatment and dissemination of data. Such sensors are called smart sensors. Simulation using adequate models of concurrency can help in the design activity of smart sensors. In this paper, we present a framework based on Models of Computation (MoCs) to model concurrency in a smart sensor software that are implemented using actors. This framework enables to rapidly test alternatives of architecture in the early stages of the design process. Some actors implement the computation behavior, some others implement the communication behavior. This distinction promotes agility during the exploration phase. Several MoCs are investigated among which, Kahn Process Network (KPN), Communicating Sequential Process (CSP) and Synchronous Data Flow (SDF). The framework implementation relies on Scala at first, while Smalltalk and then the Biniou framework are used to speed up the process, through alleviating the need for a software simulator and compiling to hardware platforms instead