Forschungsthemen
[MA] Supporting parallel role-based programming with actors
Modern software is expected to adapt dynamically to changing circumstances and contexts. This dynamic behaviour can be modelled using structured context (compartments) and roles using the CROM meta-model. However, modern software is also expected to make use of multi-core processors with parallel programming. Existing implementations of role-based programming languages like SCROLL do not natively support parallel programming and thus can not fully utilize modern systems. Originally proposed in the 1970s, the actor model of computation is a concurrency model where objects themselves are the unit of concurrency and can execute in parallel. Objects communicate by asynchronously passing messages. Each actor sequentially processes incoming messages. Since no memory has to be shared between actors, actor systems can scale to multiple cores as well as to different systems by transmitting messages over the network. A well known implementation of the actor model is the Akka framework. The goal of this work is to facilitate parallel role-based programming by mapping roles and compartments to the actor model.
Betreuer: Tim Kluge